ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1148

Lumbar spinal stenosis

Body mass index and the patient's perspective

BJÖRN KNUTSSON

ISSN 1651-6206ISBN 978-91-554-9376-9urn:nbn:se:uu:diva-264589

Dissertation presented at Uppsala University to be publicly examined in Rosénsalen,Akademiska barnsjukhuset, ingång 95-96, 751 85, Uppsala, Friday, 4 December 2015 at13:00 for the degree of Doctor of Philosophy (Faculty of Medicine). The examinationwill be conducted in Swedish. Faculty examiner: Associate professor Paul Gerdhem(Department of Ortopaedics, Karolinska University Hospital and Department of Ortopedicsand Biotechnology, Karolinska Institutet).

AbstractKnutsson, B. 2015. Lumbar spinal stenosis. Body mass index and the patient's perspective.Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine1148. 70 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9376-9.

During recent decades, lumbar spinal stenosis (LSS) has become the most common indicationfor spine surgery, a change that coincides with a higher worldwide prevalence of overweightand obesity. Thus, surgical treatment of LSS in the overweight and obese population is commonand increasing in scope.

The overall aim of this thesis was to investigate whether body mass index (BMI) is relatedto the development of LSS, and whether BMI is linked to outcome after surgery for LSS. Wefurther evaluated whether there are specific experiences of LSS from a patient perspective.

Data were obtained for all patients registered in the Swedish Spine Register who hadundergone surgery for LSS between January 1, 2006 and June 30, 2008. After adjusting fordifferences in baseline characteristics, patients with obesity showed both poorer results aftersurgery and a higher rate of dissatisfaction than patients with normal weight (odds ratio 1.73;95% confidence interval, CI, 1.36-2.19).

Furthermore, patients with obesity in the cohort reported modest weight loss at follow-up (2.0kg; 95% CI, 1.5-2.4), and only 8% reported a clinical important weight loss 2 years after surgery.

Our analysis of 389,132 construction workers, showed that overweight (incidence rate ratio,IRR 1.68; 95% CI, 1.54-1.83) and obesity (IRR 2.18; 95% CI, 1.87-2.53) were associated withan increased future risk in developing LSS when compared with patients with normal weight.

To gain insight into the patients' perspective of LSS, we performed interviews with 18 patientswho were on a waiting list for LSS surgery. The transcripts, analyzed with content analysis,revealed that living with LSS is a physical, mental and social challenge in which resources tocope with the condition are of major importance.

In summary, obesity is associated with poorer results after surgery, and patients with obesityreport modest weight loss during follow-up. In addition, obesity is associated with an increasedrisk to develop LSS. Our findings revealed that being a patient with LSS, naturally involvesconsiderable suffering and pain, but it also implies being a person with his or her own resourceswho is able to cope with these adverse conditions.

Keywords: Antonovsky, back pain, BMI, body mass index, Bygghalsan, cohort study,coping, EQ-5D, excess weight, leg pain, LSS, lumbar spinal stenosis, obesity, ODI Oswestrydisability index, overweight, patient-centered, patient-physician relationship, patient-perspective, patient-related-outcome-measure, PROMs, qualitative study, spinal stenosis,salutogenesis, salutogenic, sciatica, spine surgery, suffering, Swedish spine register, Swespine,VAS, weight loss, weight change

Björn Knutsson, Department of Surgical Sciences, Akademiska sjukhuset, Uppsala University,SE-75185 Uppsala, Sweden.

© Björn Knutsson 2015

ISSN 1651-6206ISBN 978-91-554-9376-9urn:nbn:se:uu:diva-264589 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-264589)

To my family

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Knutsson B, Michaëlsson K, Sandén B. Obesity is associated with inferior results after surgery for lum-bar spinal stenosis: a study of 2633 patients from the Swedish spine register. Spine (Phil Pa 1976) 2013;38:435-41.

II Knutsson B, Michaëlsson K, Sandén B. Obese patients report modest weight loss after surgery for lum-bar spinal stenosis: a study from the Swedish spine register. Spine (Phil Pa 1976) 2014; 39:1725-30.

III Knutsson B, Sandén B, Sjödén G, Järvholm B, Michaëlsson K. Body mass index and risk for clinical lumbar spinal stenosis: a cohort study. Spine (Phil Pa 1976) 2015;40:1451-56.

IV Knutsson B, Jong M, Sjödén G, Augutis M. Waiting for lumbar spinal stenosis surgery: suffering, resources to cope and expectations. Submitted.

Reprints were made with permission from the publisher Wolters Kluwer Health Lippincott Williams & Wilkins.

Contents

Introduction ................................................................................................... 13 Definition and classification of lumbar spinal stenosis (LSS) ................. 13 

History ................................................................................................. 14 Clinical presentation ............................................................................ 14 Radiological criteria and relation to symptoms ................................... 14 Pathophysiology .................................................................................. 15 

Epidemiology ........................................................................................... 15 Prevalence ............................................................................................ 15 Incidence for surgically treated LSS .................................................... 16 Risk factors for LSS............................................................................. 16 

Treatment for LSS .................................................................................... 17 The natural course versus non-surgical treatment ............................... 17 Surgical treatment ................................................................................ 17 Mortality, complications and re-operation rate related to surgery for LSS ...................................................................................................... 18 Outcome predictors after LSS surgery................................................. 18 

Obesity and weight loss ............................................................................ 20 Body Mass Index ................................................................................. 21 

Being a patient with lumbar spinal stenosis ............................................. 21 Development of the salutogenic model ............................................... 22 

Rationale for the studies ........................................................................... 22 

Aims .............................................................................................................. 24 Specific aims ............................................................................................ 24 

Patients and Methods .................................................................................... 25 Data sources ............................................................................................. 25 

Swespine (Study I and II) .................................................................... 25 Bygghälsan cohort (Study III) ............................................................. 25 National Patient Register (Study III) ................................................... 26 

Body Mass Index ...................................................................................... 26 Outcome measures ................................................................................... 26 

EQ-5D, ODI and VAS (Study I and II) ............................................... 26 Self-reported walking ability and use of analgesics (Study I) ............. 27 Patient satisfaction with surgery (Study I) ........................................... 27 Weight change and clinically important weight loss, (Study II) ......... 27 Definition of LSS and surgically treated LSS (Study III) ................... 27 

Population................................................................................................. 28 Study I and II ....................................................................................... 28 Study III ............................................................................................... 28 Study IV ............................................................................................... 29 

Statistical methods (Study I, II and III) .................................................... 30 Study I .................................................................................................. 31 Study II ................................................................................................ 32 Study III ............................................................................................... 32 

Qualitative methods (Study IV) ............................................................... 32 The qualitative inductive approach ...................................................... 32 Interviews as method of data collection .............................................. 33 Content analysis ................................................................................... 33 Theoretical framework ......................................................................... 33 

Ethics ........................................................................................................ 33 

Summary of findings in comparison with other studies ................................ 35 Study I ...................................................................................................... 35 

Obesity as a plausible cause of unsatisfactory results after LSS surgery ................................................................................................. 37 

Study II ..................................................................................................... 38 Study III ................................................................................................... 39 

Obesity as a plausible cause for LSS ................................................... 40 Study IV ................................................................................................... 44 

General Discussion ....................................................................................... 47 Choice of study design ............................................................................. 48 Limitations and strengths of the studies ................................................... 48 

Study I and II ....................................................................................... 48 Study III ............................................................................................... 49 Study IV ............................................................................................... 49 The precision of an estimate ................................................................ 50 Random and systematic errors ............................................................. 50 Selection bias ....................................................................................... 50 Information bias ................................................................................... 51 Confounding ........................................................................................ 51 Differential and non-differential misclassification .............................. 51 

Generalization .......................................................................................... 52 Trustworthiness ........................................................................................ 52 

Preconceptions ..................................................................................... 52 Reflexivity ........................................................................................... 53 Credibility ............................................................................................ 53 Transferability ..................................................................................... 54 

Conclusions ................................................................................................... 55 Clinical implications ................................................................................ 55 Future perspectives ................................................................................... 56 

Acknowledgements ....................................................................................... 57 

Svensk sammanfattning ................................................................................ 59 

References ..................................................................................................... 61 

Acronyms

BMI Body mass index CI Confidence interval CIWL Clinically important weight loss CSA Cross-sectional area CT Computed tomography DAG Directed acyclic graph DS Degenerative spondylolisthesis EQ-5D EuroQol group five dimensions of health GLM Generalized linear model GRR General resources of resistance HRQoL Health-related quality of life ICD International classification of disease ICHOM International consortium for health outcomes measures IRR Incidence rate ratio Kg Kilogram LBP Low back pain LDL Low-density lipoprotein LSS Lumbar spinal stenosis M Meter MLS Multi-level stenosis MRI Magnetic resonance imaging NPR National patient register ODI Oswestry disability index OR Odds ratio PROM Patient related outcome measures QoL Quality of life RCT Randomized controlled trial SD Standard deviation SF-36 The short form (36) health survey SOC Sense of coherence SPORT Spine patient outcomes research trial SRWA Self-reported walking ability VAS Visual analogue scale WHO World Health Organization

13

Introduction

Definition and classification of lumbar spinal stenosis Lumbar spinal stenosis (LSS) is defined as a narrowing of the lumbar spinal canal that is caused by facet joint arthrosis, thickened and shrunken ligamentum flavum, bulging disc, thickened lamina and in some cases spondylolisthesis. These spondylotic changes diminish the space for neural and vascular stuctures.1 Spinal stenosis can further be divided into central, lateral or foraminal.1, 2 All patients included in this study have central spinal stenosis with or without spondylolisthesis as the main diagnosis, although combined types can exist. LSS can also be classified according to etiology into two main categories, namely congenital/developed and acquired.1, 2 The most common type is degenerative acquired LSS with predominance in the lower lumbar levels.1

Figure 1. Lumbar spinal stenosis with images of a cross-sectional (A) and a sagittal plane (B). In the sagittal plane, the minor slip between L4 (lumbar vertebrae 4) and L5 (lumbar vertebrae 5) is a spondylolisthesis. Arrows indicate stenosis. Picture C and D shows a normal spine.

14

History During the 19th and beginning of the 20th century several descriptions of neurogenic claudication as a differential diagnosis to vascular claudication and syphilis were described.3 Thomson measured the diameter of the spinal canal in 1913, the same year Elsberg surgically decompressed a lumbar nerve root.3 During the following decades, methods for surgical treatment of LSS as well as morphological and radiological classification of LSS devel-oped.

In 1949, Verbiest introduced the term spinal stenosis. His case report in-cluded seven patients with clinical and radiological features of spinal steno-sis.4 The development of computed tomography (CT) increased the under-standing of LSS. However, the diagnosis and treatment of LSS have been debated for decades and was not classified in the International classification of disease (ICD) until 1987.

Clinical presentation Clinical degenerative LSS is typically diagnosed in the 6th and 7th decade of life. Even if back pain were common among patients with LSS it is consid-ered a non-specific symptom.5 The typical symptoms include leg pain, espe-cially during walking, associated with numbness and paraesthesia, and some-times loss of motor control and bladder disturbances.1 In addition these symptoms should be induced or exacerbated during standing or walking and relieved by lumbar flexion.1, 6

Clinical findings are often scarce. Even if some kind of motor or sensory deficit is common, the findings are typically mild.6

Differential or coexisting diagnosis is hip osteoarthritis, trochanteric bur-sitis, vascular claudication, restless legs and neuropathy, but also more se-vere conditions like neoplasia and normal pressure hydrocephalus.2, 6-8

Radiological criteria and relation to symptoms Examination with CT, CT myeolography or magnetic resonance imaging (MRI) is traditionally used to diagnose LSS. MRI is the modality of choice today with high sensivity and specificity.9, 10 There are at least sixteen differ-ent radiological criteria for diagnosis of LSS.10 An expert group recom-mended qualitative methods for classification of LSS in clinical practice, and quantitative measures for research. Recommended qualitative measures are the compromise of the central zone and the relation between fluid and the cauda equina.10 Furthermore, the cross-sectional area (CSA) of the dural sac were the recommended quantitative method for clinical outcome studies, with fair to excellent reproducibility.10, 11 The expert group considered the CSA method difficult during the daily working routine. However, today

15

most spine units in Sweden have computerized software for MR images, and measurement of the CSA is implemented in the daily routine. Schönström et al. measured the CSA in 7 cadavers.12 They concluded that once 60-80 mm2

is reached, further compression of the dural sac increases the pressure on the nerve roots. Anthiwiraham et al found an association between CSA<70mm2 and preoperative function among surgically treated patients.13 Whether there is a linear correlation between CSA and clinical symptoms once the critical area is reached is uncertain.11, 13, 14 Ogikubo et al. reported a positive correla-tion between preoperative symptoms and CSA, which is in contrast to Sigmundsson et al who found no relation between symptoms and the CSA.11,

14 Both studies included patients before surgical treatment for LSS, where the majority of the patients demonstrated an CSA≤80 mm2.

Sirvanci et al. found no correlation between CSA and function in the Oswestry Disability Index (ODI).15 Similar to Sigmundsson et al., their ret-rospective study comprised a high proportion of patients with multi-level stenosis (MLS). The number of patients with MLS in Ogikubo's study was not shown.11, 14 MLS and coexisting degenerative features of the spine are common which is a limitation of the one level CSA method.

Radiographic LSS without symptoms is common. For instance, Boden et al performed MRI on 67 asymptomatic persons and found radiological spinal stenosis in 21% of the participants over 60 years old.16 Finally, even if asymptomatic radiological LSS were common among elderly, Ishimoto et al. observed a correlation between severe stenosis and clinical symptoms in a cross-sectional study on a Japanese population.17

Pathophysiology The pathophysiology hypothesis for these LSS symptoms is a combination of direct pressure on neural elements and vascular structures, but also an increased pressure inside the dural sac, a condition that disturbs nutrition to the neural elements. Postural positions, such as extension of the back and standing can potentially increase both the direct and indirect pressure.1

Epidemiology Prevalence Few studies have described the prevalence of LSS in the general population. The prevalence is highly dependent of the age of the studied population and definition of spinal stenosis. Kalichman et al noted a prevalence of 24% (45/191) of CT-evaluated LSS in a subgroup analysis in the Framingham cohort. The mean age was 53 years (range 32-79), the prevalence of LSS showed similar sex distribution.5 The authors used the axial diameter of the

16

spinal canal as the definition of LSS. In a recent cross-sectional study on the prevalence of radiological and clinical LSS, moderate to severe LSS was common (78%, 731/938), although symptoms were uncommon (13%, 94/731).17 A qualitative evaluation based on MRI imaging was used for di-agnosis. The mean age was 67 years (range, 40-93) and the prevalence of radiographic LSS increased from 64% in people 50 to 59 years to 93% in people 80 to 89 years. The prevalence of symptomatic and asymptomatic LSS was similar in male and female persons.

Incidence for surgically treated LSS During recent decades, the most common indication for spine surgery in many countries is LSS. In Sweden the incidence of LSS surgery increased from 4.7 to 13.2 per 100 000 inhabitants per year from 1987 to 1999,18 and the annual number of registered surgical procedures for LSS is continuously on the rise.19 We have also observed a similar trend in Europe and the United States.20-22 Increasing age in the population, increased use of MRI and sys-tematic differences in health care are major factors to account for this trend, although the impact of biological factors might also be relevant.20, 23-28

Risk factors for LSS A recent study on 598 Finish male twins described radiographic LSS as a highly genetic condition that is to some degree mediated by disc degenera-tion.29 However, the degree of hereditability also depend on the definition of LSS and lumbar level: the environmental influence increases with decreasing lumbar level (from L1 to S1).29

Older age is a risk factor for degenerative LSS, and the majority of the pa-tients included in studies on degenerative LSS are ≥50 years old. The results from a Swedish incidence study found a low proportion of patients (13%) with LSS who were <49 years of age.18 The causality between age and LSS could be low grade inflammation, immunosenescence and impaired repair pathways both genetically and environmentally.30

Smoking and strenuous work are not established risk factors for LSS, alt-hough they are reported as risk factors for spondylosis and disc degeneration which could develop into clinical LSS.31-33

Research on sex differences has produced diverse and inconsistent results. Several reports on surgically treated LSS showed a higher proportion of female participants, especially LSS that was due to spondylolisthesis.34-36 In contrast, male predominance or equal sex distribution is also presented in the literature.18, 37, 38 To my knowledge, there were no studies that have examined the links between body mass index (BMI) an the incidence rate for LSS.

17

Treatment for LSS The natural course versus non-surgical treatment The natural course of LSS is poorly documented. Johnsson et al. selected 32 patients for non-surgical treatment of LSS. Of these 32 patients, 27 were followed for 4 years. The majority of the patients reported unchanged or decreased symptoms during follow-up (70-85%); severe deterioration was not found in these patients.39

The first choice of treatment is non-surgical including bed rest, bracing, painkillers, epidural injections and physical therapy.40, 41 Although the evi-dence for non-surgical treatment is poor, several studies have reported un-changed or decreased symptoms at follow-up in patients selected for non-surgical treatment.40-42 Simotas et al. followed 40 patients with an active treatment algorithm, including physical and pharmacologic treatment, with a mean follow-up of 33 months (range, 16-55). Of the 40 patients, 77% (n=31) reported unchanged or decreased symptoms. The results also indicated that neurologic deterioration was rare.40 Although epidural injections could have a short-term benefit; epidural injections with steroids have minimal or no treatment effect compared to epidural injections with local anesthetics.43-45

Surgical treatment Several studies have compared surgical with non-surgical treatment, some of which have been randomized controlled trials (RCTs). Although, the results from these studies, have shown conflicting results; surgical treatment is con-sidered superior over non-surgical treatment for strictly selected patients.34,

37, 38, 41, 46-49 The golden standard and most common method for surgery is decompression of the osseous and soft tissue elements that compress the spinal canal with conventional laminectomy.50

The indication for adding a fusion to decompression remains unclear, but is associated to increased risk for complications compared with the decom-pression procedure alone.20, 21, 35, 51

Micro-decompression surgery is a procedure that entails minimal skin in-cision and preserves the supraspinous process and interspinous ligaments. The rationale for this procedure is that preserving the posterior structures enhances rehabilitation and improves the outcomes compared with standard decompression. A Norwegian register study compared micro and standard decompression, and found no clinical important differences in outcomes or complication rates 12 months after surgery.52 Similar results were reported in a recent Cochrane review.53 The authors compared techniques that preserved the posterior structures with conventional laminectomy and found equal results regardless of surgical technique.

18

Because of comorbidities, many patients are not eligible for decompres-sion procedures. Accordingly, a new minimalistic technique with an interspinous device has been introduced. Through distraction between the supraspinous processes, the patients symptoms can decrease owing to indi-rect decompression of the dural sac. Moojen et al. employed a sample of 205 patients in a multicenter RCT.54 The interspinous process device was com-pared with standard decompression in a 1-year follow-up. The methods showed equal improvement in patient-reported outcome measures (PROMs), but the reoperation rate in the interspinous process device group was 29% (21/73) which was more than threefold compared with the standard decom-pression group (8% , 6/72).54

Mortality, complications and re-operation rate related to surgery for LSS Jansson et al. studied the mortality and re-operation rate in Sweden between 1987 and 1999.18, 55 The case fatality rate was 3.5 per 1000 operations within 30 days.18 Age >75 years, fusion procedures and male sex were associated with a higher risk for postoperative mortality.18 The re-operation rate was 5% after 2 years and 11% after 10 years.55

The registration of complication rates is often poor. Surgeons tend to un-derreport, while patients often overreport the complications. Earlier studies on posterior lumbar fusion on obese patients have shown an increased risk for complications compared with normal-weight patients, especially postop-erative infections.56-59

A recent North American register study that included over 49 000 elective surgical procedures of the spine60 showed that obesity, particularly obesity class II and III, was associated with minor and major complications. Howev-er, elevated BMI was not associated with an increased mortality. In a sensi-tivity analysis, cohorts were matched for comorbidities. The results from the analysis indicated that differences between the BMI groups diminished, but BMI was still a risk factor in the obese class III group compared with the normal-weight group. The following complication rates were reported in patients with normal weight: minor complications (2.1%), major complica-tions (3.8%), 30-day return to operation room (2.3%) and 30-day mortality (0.1%).60

Outcome predictors after LSS surgery This section is limited to PROMs, including health related quality of life (HRQoL), function, pain, walking capacity, use of analgesics and satisfac-tion with treatment. A predictor can be associated with inferior (negative) or superior (positive) outcome. It is important to emphasize that all predictors

19

may be outcome- specific, which is a limitation especially in studies using one or few outcome measures.61

A meta-analysis comprised of 74 studies reported a good to excellent out-come in 64% of the patients after LSS surgery.62 Consequently, a substantial proportion of the patients were not satisfied with the outcome after surgery, probably because of residual symptoms, complications associated with sur-gery and unrealistic expectations of the surgical outcome. Aalto et al. per-formed a systematic review on preoperative predictors for postoperative outcome in LSS. Totally, 885 abstracts were assessed and only 21 of ac-ceptable quality were included in the study.61 A summary of these studies and more recent research are presented below.

Radiologic predictors In a study following 109 patients for 2 years, Sigmundsson et al. found that a narrow preoperative CSA was related to greater improvement in back and leg pain.63

Another study divided a cohort (n=47) based on preoperative CSA and, found an association between a CSA <70 mm2 and better postoperative func-tion as measured by the ODI.64 Jönsson et al described similar findings as those in Sigmundsson i.e. a severe encroachment in the sagittal plane (<6mm) was associated with better improvements in leg pain and walking ability.65

Degenerative spondylolisthesis (DS) is discussed as a predictor for out-come after LSS surgery. However, a recent cohort study included 5390 pa-tients with and without DS; no significant differences in outcomes were found during follow-up.35 LLS is also associated with degenerative scoliosis which is reported as a negative predictor for back pain in a group of 90 pa-tients followed for 2 years after surgery.66

Health, comorbidities, walking ability, pain and function Katz et al. analyzed the predictors for postoperative outcome in LSS in 272 participants. He found that self-rated health was moderately correlated to preoperative function and pain. Self-rated health was also a positive predic-tor for better walking ability and satisfaction.67 A sense of coherence (SOC) and ability to cope were associated with self-rated health, as well as with quality of life (QoL), function and outcomes after LSS surgery.68, 69 Somatic and emotional health are positive predictors, and comorbidities especially cardiovascular comorbidities and depression, predict worse outcome.61, 67

Predominant leg pain is associated with better improvement in leg pain and walking ability,63 but long duration of pain is associated with poor out-comes.63, 70, 71

Not surprisingly, good preoperative walking ability predicts better post-operative walking ability. Additionally, in a systematic review, disorders affecting the walking ability were found to be associated with worse postop-

20

erative walking ability and pain.61 Furthermore, low preoperative score in EuroQol group five dimensions of health (EQ-5D), high levels of pain and preoperative use of analgesics are predictors of poor outcome.63, 72 Earlier back surgery is also a negative predictor of outcome probably related to re-sidual symptoms, complications, low function, high levels of pain and long pain duration.73

High income and educational level High income is associated with better walking ability, a higher degree of improved symptoms and satisfaction.61 Correlations also exist between high educational level and improved back pain, leg pain, ODI and the short form (36) health survey (SF-36).74

Expectations Iversen et al. followed 257 patients for 6 months. The authors found that more ambitious expectations of physical function are associated with better function and satisfaction.72 In contrast, a great number of expectations of pain relief proved to be a negative predictor for postoperative pain.72 Anoth-er study that included 184 patients reported affirmative expectations as a positive predictor for ODI and leg pain.74

Age, sex and life style factors The majority of the studies that included age or sex as a predictor show no difference associated with either age or sex.61, 75

In a recent cohort study, Sandén et al. showed that smoking is associated with dissatisfaction and poorer outcomes.76 The cohort study, which includ-ed 4555 surgically treated patients, found inferior results in smokers com-pared to non-smokers in all outcome measures: EQ-5D, SF-36, ODI, back pain, leg pain, use of analgesics, walking ability and return to work. Cobo Soriano et al also demonstrated that smoking is a negative predictor for pain relief.74

The results of earlier studies on BMI as a predictor for outcome after sur-gery are inconsistent. Several studies have reported no significant effect of BMI on outcome.61, 74, 77-80 In contrast, Athiviraham et al. found a correlation between BMI and poorer outcome on the Roland Morris Score. This study, however has several important limitations; the study was not designed to specifically evaluate BMI, it included a small number of patients (n=89) and no BMI data were provided.81

Obesity and weight loss Obesity is a condition of excessive fat accumulation in adipose tissue, and was an important determinant for survival among our prehistoric ancestors.

21

For thousands of years obesity was rare, and even indicating wealth and status in some cultures. Even if there were several historical monographs and case reports on obesity and related morbidity, obesity as a major public health problem was not appreciated until the end of the 1940s.82 During the recent decades, obesity has become a growing epidemic with no cessation in sight. Indeed, a majority of the adult population in many developed and de-veloping countries is overweight or obese.83, 84

Obese individuals have a higher rate of death from cardiovascular diseas-es, as well as a significantly elevated risk for diabetes and many specific cancers. Overweight and obesity are also associated with higher risk for lumbar disc disease, hospitalization for low back pain (LBP), several types of osteoarthritis and pain.25, 33, 83, 85-90

Weight loss in patients with obesity could be an important issue. Further, patients reduced activity secondary to LSS is cited as an important factor for weight gain and inability to lose weight.91 Postoperative weight loss in obese individuals could reduce back pain and leg pain, improve function, and enhance HRQoL.89, 92-95 However, weight loss after spine surgery may not positively correlate with better results.91

Body Mass Index In clinical practice and epidemiological research, BMI is a widely accepted proxy to classify underweight, normal weight, overweight and obesity in adults96 Although BMI is an established measure of overweight and obesity in both a clinical and research setting, the measure has admittedly a major limitation because it cannot differentiate lean mass from adipose tissue.97 In addition, age, sex, self-reported weight and height might bias the estimates.98 Nevertheless, a BMI ≥30 demonstrates high specificity and positive predic-tive value for excessive fat. Moreover, women do tend to increase their BMI by an increase in fat mass and not by higher lean mass.97, 98

Being a patient with lumbar spinal stenosis From a patients perspective LSS can be experienced as a world of suffering and despair, with an impaired QoL comparable to stroke, heart-related dis-eases and diabetes.99 Illness-suffering that is due to pain, impaired function or from being on a waiting list for surgery is obvious, but also "caring suffer-ing" and life-suffering may be important forms of suffering. Caring-suffering refers to suffering that is due to care, e.g. by patients experiencing feelings of not being believed or not being taken seriously. Life-suffering comprises experiences of a threatened life situation such as uncertainty of the world and loneliness.100, 101

22

Individuals attempt to manage suffering and demands created by one or sev-eral stressors, are often referred to as coping strategies. These efforts seek to manage, master, tolerate, reduce or minimize the demands of a stressful en-vironment.102 Furthermore, Antonovsky's salutogenic model that includes but is not restricted to coping abilities, delineates the importance of general resources of resistance (GRRs) for dealing with stressors. Through strong GRRs, life stressors can be comprehensible, manageable and meaningful, all which produce a high SOC. This capacity is relatively stable over time, but increases with age. Depending on context, stressors and available GRRs, SOC can be dynamic, at least in the short term. A strong SOC is associated with life satisfaction, perceived health, mental health, high QoL, and also better function after surgery for LSS.68, 69, 103-106

Development of the salutogenic model The Israeli-American medical sociologist Aaron Antonovsky first described his theory of salutogenesis in his book "Stress, Health and Coping" 1979, and further developed it in the work "Unraveling the mystery of health".103,

104 His theory originated from interviews with female survivors from the Holocaust. A subgroup of the interviewees showed good mental and physical health, despite their experiences of violence and traumatic events. He con-cluded that stress is a natural part of life and raised the question: How can we survive despite all chaos and stress? 107

With support from earlier research he developed, in contrast to disease (pathogenesis), a concept that focuses on health (salutogenesis). The theory has been developed to become an important tool both for quantitative and qualitative research.68, 105-113

Between 2005 and 2007 Lindström and Eriksson published several sys-tematic reviews on the salutogenic model. The authors concluded that SOC was a valid measure and a health-promoting resource.68, 105, 106

Rationale for the studies During recent decades, LSS has become the most common indication for spine surgery, a change that coincides with a higher prevalence of obesity. Thus, surgical treatment of LSS in the overweight and obese population is common and increasing in scope and number.

We hypothesized that an elevated BMI would decrease the success rate after surgery for LSS (study I) and increase the rate of clinical LSS through similar metabolic and biomechanical pathways (study III).

Because the results from previous studies on BMI as a predictor for out-come after LSS were inconsistent and because they included a limited num-ber of patients as well as other limitations we performed study I.

23

The incidence rate in relation to BMI has not been amply described. Whether an increased BMI is a risk factor for the development of a clinical LSS was unknown. Therefore, we performed study III.

Furthermore, weight loss can be important in patients with obesity. Be-cause of the limited evidence on weight change after surgery for LSS, and whether weight loss after spine surgery correlates with better results we per-formed study II. Finally, qualitative studies exclusively for patients with LSS who are waiting for surgery are scarce. Study IV describes the experi-ence of being a person waiting for LSS surgery from a patient's perspective.

24

Aims

The overall aim of this thesis was to investigate whether BMI is related to the development of LSS or with outcome after surgery for LSS. A further aim was to examine weight change after surgery for LSS and if weight loss were associated with the reported outcomes. A final aim was to study wheth-er there are specific experiences of LSS from a patient's perspective.

Specific aims I To determine the association between BMI and outcome of

lumbar spine surgery for spinal stenosis (study I).

II To determine the extent of weight change 1 and 2 years after surgery in obese LSS patients, as well as to define the propor-tion of obese patients who reported a clinically important weight loss (CIWL) and whether this change is related to im-provement in PROMs 2 years after surgery (study II).

III To assess the relation between BMI and later development of

clinically overt LSS (study III).

IV To describe the experience of being a person waiting for LSS surgery, and how everyday life and suffering are managed un-der the influence of the disease (study IV).

25

Patients and Methods

Data sources Swespine (Study I and II) The National Swedish Register for Spine Surgery (Swespine) is a prospec-tive registry of patients who have undergone surgery for spinal disorders, including LSS. The definition of LSS in Swespine is based on the surgical record from the surgeon. More than 80% of the total number of surgical pro-cedures for degenerative lumbar spine disorders in Sweden are included.114 The patients complete a preoperative questionnaire and postal follow-up questionnaires at 1, 2, 5 and 10 years after surgery. The surgeon records surgical data, including diagnosis, without having access to the patient ques-tionnaires. Preoperative data completed by the patient include age, sex, smoking habits, consumption of analgesics (regular use, intermittent or no use), weight, height, back pain, leg pain, walking ability, duration of symp-toms, the EQ-5D and the ODI. Walking ability is recorded as one of four categories (<100m, 100-500m, 500-1000m and >1000m). Back and leg pain are measured on a visual analog scale (VAS). Patient self-reported satisfac-tion with surgery is recorded as one of three categories: satisfied, uncertain and dissatisfied.

Bygghälsan cohort (Study III) The Bygghälsan cohort consists of Swedish construction workers who par-ticipated in a nationwide occupational health surveillance program, that was initiated through a trade agreement between employers and unions.115 The workers were invited to participate in health examinations (participation rate was at least 80%). At inclusion, weight was measured with a scale and height with a stadiometer. The workers’ job title (22 categories) and smok-ing habits were also registered. From 1971 the data were computerized. The examinations ended in early 1993 and the computerized register contains 389 132 persons. The proportion of females was 5%, of which 42% were office workers.

26

National Patient Register (Study III) The National Patient Register (NPR) started in 1964, covered 83% of the Swedish population in 1972 and all inpatient care in Sweden since 1987. The completeness of ascertainment and the accuracy of diagnostic classification in the NPR are both high. The completeness for spinal diagnosis and surgical procedures in the NPR was on average 86% in 2001-2012.116 117

Body Mass Index Weight and height were self-reported in study I, II and IV and measured in study III. BMI was calculated as bodyweight in kilograms (kg) divided by the square of height in meters (m). We used BMI both as a continuous and categorical variable.

Because self-reported BMI might bias the estimates, we performed a sen-sitivity analysis in study II using a correction algorithm described by Nyholm et al.118

Table 1. Body Mass Index (BMI) categories defined by the World Health Organiza-tion (WHO)

BMI category BMI range

Underweight <18.5Normal weight 18.5-24.9Overweight 25-29.9Obese Class 1 Obese Class 2 Obese Class 3

≥30-34.9 35-39.9 ≥40

Outcome measures EQ-5D, ODI and VAS (Study I and II) The EQ-5D, ODI and VAS are PROMs with good validity, reliability and responsiveness for spine surgery. All three measures reflect how patients perceive their treatment.119-121

The EQ-5D is a generic instrument of HRQoL, that contains five dimen-sions of health: mobility, self-care, activities of daily life, pain and depres-sion. Each dimension has three levels and the measure includes 243 different health states. Respondents' answers are transformed to a numeric score, (range -0.594-1): lower scores indicating inferior HRQoL.120

The ODI is a disease-specific functional score and HRQoL instrument that contains 10 items related to limitations in daily life activities (Pain In-tensity, Personal Care, Lifting, Walking, Sitting, Standing, Sleeping, Sex

27

Life, Social Life and Traveling). Each question has five points and the sum-mary is transformed to percentage scored from 0 to 100; higher scores indi-cate more severe symptoms.120, 121

VAS consists of a 100 mm line, with 0 at one end indicating no pain at all and 100 at the other end indicating worst possible pain.119

Self-reported walking ability and use of analgesics (Study I) In study I, we used patients' self-reported walking ability (SRWA) and use of analgesics as outcome measures. Walking ability was dichotomized into "improved walking ability" or "not improved walking ability" (unchanged or impaired walking ability). Although there is a correlation between estimated and measured walking distance, overestimation is common and hence could lead to inaccurate estimates.122, 123

Self- reported use of analgesics was dichotomized into "use" (regular or intermittent use) or "no use".

Patient satisfaction with surgery (Study I) In study I, patients' satisfaction with surgery at the 2-year follow-up was dichotomized into "satisfied" or "not satisfied" (uncertain and dissatisfied). This measure of satisfaction is associated with PROMs in Swespine63 In contrast, the definition of satisfaction can differ between studies, not neces-sarily reflecting the treatment effect.124

Weight change and clinically important weight loss, (Study II)

In study II, the weight change in kilograms 1 and 2 years after surgery was our main outcome measure. Further, we analyzed any associations between CIWL and outcomes. A weight loss of ≥10% was considered a CIWL.125 The cohort was divided into three subclasses depending on weight change from time of inclusion to the 2-year follow-up: weight stable (9% weight reduction to 9% increased weight), weight loss (≥10% weight loss) and weight gain (≥10 % weight gain). Additionally, weight loss is associated with improvements in EQ-5D, ODI and pain (VAS).92-95

Definition of LSS and surgically treated LSS (Study III) In study III, primary LSS diagnosis was defined by the ICD-9 codes 724, 724.0 or 724.00 and ICD-10 codes M48.0 or M48.0K. In addition, a sensitiv-ity analysis was performed using as outcome a diagnosis of LSS combined with a surgical procedure of the lumbar spine.

28

Population Study I and II Data were obtained for all patients registered in Swespine who had under-gone surgery for LSS between January 1, 2006 and June 30, 2008. Figure 2 is a flow diagram for inclusion in study I and II.

Table 2. Baseline characteristics of the patients in study I divided by categories of Body mass index. Continuous variables given as means with standard deviations (SD).

Normal weight n=819

Overweight n=1208

Obese n=606

Female sex 63% (n=518) 52% (n=628) 58% (n=352)Current smoker 19% (n=152) 14% (n=169) 11% (n=69)Previous surgery 17% (n=138) 17% (n=204) 17% (n=106)Analgesics 51% (n=416) 51% (n=611) 54% (n=327)Age 70 (9) 69 (8) 67 (8)EQ-5D 0.38 (0.32) 0.39 (0.32) 0.34 (0.31)ODI 42 (16) 43 (16) 46 (14)Back pain (VAS) 52 (28) 52 (28) 57 (26)Leg pain (VAS) 60 (26) 60 (26) 61 (26)

Table 3. Baseline characteristics of the patients in study II, divided by categories of weight change during follow-up. Continuous variables given as means with stand-ard deviation (SD).

Weight stable n=483

Weight loss n=45

Weight gain n=10

Female sex 56% 64% 80%Current smoker 11% 23% 0%Age 66 (8) 66 (9) 63 (12)EQ-5D 0.35 (0.31) 0.34 (0.32) 0.23 (0.32)ODI 46 (15) 47 (13) 51 (10)Back pain (VAS) 56 (27) 65 (19) 55 (30)Leg pain (VAS) 61 (26) 62 (22) 46 (31)

Study III 389 132 constructions workers were registered from 1971 through 1992, while the observation period was from cohort entry until December 31, 2011, death, emigration or the occurrence of first diagnosis that was due to LSS, whichever occurred first. Figure 3 is a flow diagram for inclusion in study III.

29

Table 4. Characteristics of the study group at baseline divided by categories of body mass index.

Underweight n=5795

Normal weight n=235 247

Overweight n=106 730

Obese n=16 695

Female sex 15% (n=846) 5% (n=12 888) 3% (n=3665) 7% (n=1147) Current smoker 46% (n=2654) 39% (n=90 924) 35% (n=37 801) 33% (n=5584) Former smoker 7% (n=407) 13% (n=29 533) 19% (n=20 048) 20%(n=3270)

Age, mean (SD) 25 (9) 31 (12) 39 (13) 42 (13)

Study IV During the study period (October-December, 2013), 18 consecutive patients on the waiting list for LSS surgery at a Swedish county hospital were re-cruited into the study. All patients asked agreed to participate.

Our inclusion criterion were age ≥18 and good knowledge in the Swedish language. Patients with dementia were not eligible for the study. A few days after the patient were scheduled for surgery, a specific research nurse, not involved in the care of the patients, gave information about the study.

Table 5. Characteristics of the 18 interviewees at baseline.

Obese n=6 (33%)

Overweight n=6 (33%)

Female sex n=11 (61%)

Current smoker n=2 (11%)

Age mean (SD) 66 years (11)

Employed n=5 (28%)

Retired n=8 (44%)

Disability pension n=5 (28%)

Partnership n=14 (78%)

One or several comorbidities*, n=15 (83%)

*One patient could have several comorbidities. Number of comorbidities at inclusion hyper-

tension (n=11), heart disease (n=2), Lung disease (n=3), diabetes (n=7), depression (n=1),

rheumatiod arthritis (n=1), muscle rheumatism (n=1), osteoporosis (n=1), prostatism (n=2),

ulcerative colitis (n=1).

30

Figure 2. Flow diagram for inclusion of patients in study I and II.

Figure 3. Flow diagram for inclusion of construction workers in study III.

Statistical methods (Study I, II and III) Proper choice of statistical methods is essential to achieve high internal and external validity. The choice not only depends on the research question, but also on the nature and distribution of the exposure and outcome.

Data, including a number of categories, are referred to as categorical data and further divided into nominal data (not ordered categories) and ordinal (ordered categories) data. Numerical values are defined as numerical data, divided into discrete and continuous data.126 p14

31

The Chi-squared test and Fisher's test are commonly used to determine whether there are statistically significant differences between categorical data.126 p69 The T-test or ANOVA, referred to as parametric tests are often used to test whether there are significant differences between numerical da-ta.126 p 60 However, the validity of these tests are dependent upon certain as-sumptions, i.e. the variable in question should be normally distributed with an equal variance. However, if the sample size were reasonably large para-metric tests are fairly robust to departures from the assumption of normality.126 p 60, 63

When the assumptions for parametric tests are not fulfilled non-parametric tests are a viable option.126 p63 The non-parametric Kruskal-Wallis test is comparable with one-way ANOVA, and the non-parametric Mann-Whitney U-test with the Student's t-test.126 p63

Linear regression models can describe any linear effect of an independent variable (X) on a dependent variable (Y). The residuals should be normally distributed with constant variance. The influence of covariates can be con-sidered by multivariable linear regression.126 p 78, 84

To use a mathematical model that best describes relation between one or several exposures and outcomes we used different regression models. The generalized linear model (GLM) can be expressed in the form.126 p 96

g(Y)= a+b1x1 + b2x2 + bkxk

g(Y) is called the link function and the dependent variable in simple and multiple regression. In logistic regression the link function is the loge of the odds and the loge of the rate in Poisson regression.

Model for logistic regression: ln (odds) = a+b1x1 + b2x2 + bkxk

Model for Poisson regression: ln (rate) = a+b1x1 + b2x2 + bkxk

To gain insight into potential nonlinearity, we modeled restricted cubic spline curves in study I and III.127, 128 We used knots placed at recommended percentiles of 5 (21.3 kg/m2), 35, 65 and 95 (34.4 kg/m2).129

Study I For continuous outcome variables in study I, adjusted means for each BMI category were estimated using the procedure GLM in the SAS package sys-tem. For dichotomous dependent variables, multivariable logistic regression was applied to assess odds ratios (ORs) with 95% confidence intervals (CIs). The models were adjusted for age, sex, smoking, use of analgesics, previous back surgery, duration of symptoms and the baseline values of the variable

32

under investigation. We modeled the nonlinear trend in the risk of dissatis-faction by a restricted cubic-spline logistic regression analysis using knots placed at percentiles of 5 (21.3 kg/m2), 35, 65 and 95 (34.4 kg/m2), with 25.0 kg/m2 as the reference.

Study II In study II, paired t-tests were used to access weight change 1 and 2 years post-surgery. Because of unequal group sizes and because of the non-normal distribution of our outcome variables (PROMs), Kruskal-Wallis tests were conducted. Post-hoc tests were performed with Mann-Whitney U-tests and a Bonferroni-adjusted significance level of 0.017 was calculated (α-level 0.05). Lehr's formula was applied to calculate a necessarily group size of 26-41 patients for a power of 80%, using a minimal detectable change of 0.2 in the EQ-5D, 10 in the ODI and 18 in the VAS.

Study III Poisson regression models were used in study III to estimate age- and multi-variable adjusted incidence rate ratios (IRRs) with 95% CIs. The latter mod-el included age (continuous), sex, smoking status (never, former, moderate current, heavy current, unknown) and occupation (22 categories). Non-linear trends of risk were displayed using restricted cubic spline curves estimated by a multivariable-adjusted Poisson regression model. We used knots placed at the 5th, 35th, 65th and 95th percentiles of the cumulative BMI distribution. The reference level was set to 20 kg/m2. Finally, we performed stratified analysis by sex.

Qualitative methods (Study IV) The qualitative inductive approach Inductive analysis involves discovering categories, themes and patterns from a large amount of information.130 With an inductive approach extensive raw material can be condensed into a summary, as well as establish relevant links between the study participants and the findings. The findings can further develop a model or theory. However, in this study we exam our data using the salutogenic model as a sensitizing concept instead of developing a new theory.130 p541-545

33

Interviews as method of data collection The purpose of qualitative interviewing with open questions is to capture the interviewees' experiences and perceptions.130 p 442 The individual interviews with open questions provide a framework for the patients in study IV to de-scribe their experiences in their own words from an individual perspective.130 The individual face-to-face interviews started with structured questions on socio-demographic and medical information. This phase was followed by open-ended questions designed to stimulate a dialogue on the topic under study. An interview checklist was used to compromise broad domains i.e. care, patient perspective on experiences and feelings from LSS and expecta-tions before surgery. Interviews were digitally recorded and later transcribed verbatim. The interviews lasted from 60-90 minutes and were conducted by the first author (BK).

Content analysis The interview transcripts were analyzed by content analysis, an approach that aims to capture the core of the interviews and explore differences and similarities in the transcribed text.131

All interview transcripts were read by all authors, with first impressions were discussed and documented. The next step was to condense (shorten but still preserving the core concepts) the transcripts into meaningful units, fur-ther aggregated (interpretation on a higher logical level) into codes. These codes were aggregated into categories, which further created themes.131

Theoretical framework To explain the patterns in the data we need a framework. Several descrip-tions of coping mechanisms, GRRs and description of the patients own re-sources were found in the interviews. Therefore, the study is framed within the context of Antonovsky's salutogenic theory. Hence, we chose to present our interpretation of the categories and themes using the salutogenic model as a sensitizing concept.103, 104

Ethics Ethical approval was obtained for all parts of this thesis.

In study IV patients gave informed consent and were granted confidential-ity. The interviewer (BK) was also the forthcoming surgeon. Serving in both roles could be considered an ethical problem, but all patients were on the waiting list for surgery and therefore the interview could not affect the planned surgery or follow-up. Finally, patients were informed that they

34

could terminate their participation in the study at any time without conse-quences. None of the other authors participated in the care of the patients.

To be included in a register or an interview study could arouse dormant negative experiences. In addition, patients might agree to participate because of perceptions that their choice may affect their treatment.

Furthermore, the results from study I, II and III could also be stigmatizing for patients with excess weight.

35

Summary of findings in comparison with other studies

Study I Of the 2633 patients included in the analysis 819 (31%) had normal weight, 1208 (46%) were overweight and 606 (23%) were obese. The adjusted means with 95% CIs are reported in Table 6. For obese patients, the ODI and the EQ-5D indicated a lower HRQoL at baseline and at the 2-year follow-up. When differences in baseline were adjusted, the obese patients still showed inferior results after surgery compared with the normal weight patients. Dif-ferences between the normal weight and overweight groups were modest. Moreover, the obese patients reported more back and leg pain compared to the normal weight group.

We found that 39% of the obese (p<0.0001), 31% of the overweight (p=0.002), and 26% of the normal weight patients (reference group) were still using analgesics 2 years after surgery. No correlation was found be-tween BMI and improvement in walking ability: 50% of the normal weight, 53% of the overweight and 48% of the obese patients had improved walking ability 2 years after surgery.

On average, 67% of the normal weight, 64% of the overweight, and 57% of the obese patients were satisfied with the surgical results. Furthermore, in the obese patients, we observed higher odds of dissatisfaction with higher BMI (Figure 4). In comparison with the normal weight patients, the average adjusted OR for dissatisfaction in the obese group was 1.73 (95% CI, 1.36–2.19).

Table 6. Adjusted means with 95% confidence intervals at the 2-year follow up. Models adjusted for age, smoking, earlier back surgery, duration of back pain, use of analgesics and baseline differences for the variable studied.

BMI<25 kg/m2 BMI 25-29.9 kg/m2 BMI≥30 kg/m2

EQ-5D 0.64 (0.62-0.66) 0.63 (0.61-0.65) 0.56 (0.54-0.59) ODI 25 (23-26) 27 (26-28) 33 (31-34)Back pain (VAS) 31 (29-33) 33 (31-34) 39 (37-42)Leg pain (VAS) 31 (29-33) 32 (30-34) 40 (37-43)

 

36

Figure 4. The association between body mass index (BMI) and adjusted odds ratio for dissatisfaction displayed as a restricted cubic-spline curve based on multivariable logistic regression analysis. A BMI of 25 kg/m2 was used as the reference. The dashed lines represent a 95% confidence interval.

Most of the patients regardless of the BMI group were satisfied with the treatment though obesity was a negative predictor of patient satisfaction. Differences between normal and overweight individuals were small and could not be regarded as clinically important.

Our results contrast with those reported in a systematic review on pre-operative and postoperative outcomes in LSS. From this material, 21 studies were included in the analysis.61 Only one examined the relation between outcomes after surgical treatment: no differences in outcome were found between obese and non-obese patients.132 The number of patients (n=170), the definition of obesity, and the outcome measures differ from our study, making that study and our difficult to compare.

Antiviraham et al. found an association between BMI and worse outcome after surgery. The analysis included 89 patients and analyzed 16 covariates including BMI. The outcome measure differed from our study and details about the BMI data were not given.81

The findings from a recent Norwegian register study are similar to our re-sults.75 Giannadakis et al. evaluated the association between obesity (BMI≥30) and outcomes 1 year after surgery for LSS. The patients with a BMI≥30 (n=353) reported poorer outcomes (ODI, back pain, leg pain) com-pared with patients with a BMI<30 (n=1120).

In contrast, two recent analyses based on the Spine Patient Outcomes Re-search Trial (SPORT) reported similar results for obese (BMI≥30) and non-obese (BMI<30) patients, with equal or even greater treatment effects among

37

patients with a BMI≥30, which was largely due to poor outcome in the con-servative treatment arm.133, 134 Patients with degenerative spondylolisthesis were analyzed separately. Only 19% (118/634) reported a BMI<25, and the study was not powered for subgroup analysis according to BMI. The "as-treated" analysis was due to a great crossover between the treatment arms, which also introduces a risk for selection bias.

Several studies have used a small number of patients with LSS, making it difficult to draw any conclusions or to make comparisons with our study.77-79 We believe our study is the first to use BMI as an continuous variable and demonstrates an increasingly higher odds of dissatisfaction with higher BMI values.

Obesity as a plausible cause of unsatisfactory results after LSS surgery Mechanical, structural, metabolic and behavioral changes have previously been discussed.87 Obesity, especially central obesity, contributes to several metabolic disorders. For instance, obesity is correlated to neuropathic disor-ders in both diabetic and nondiabetic individuals.87, 135 Miscio et al demon-strated that even nondiabetic obese individuals could develop impairment in the function of sensory fibers. This impairment is linked to hyperinsulinemia and insulin sensitivity.135 Central obesity is associated with diabetes and atherosclerosis. Increased blood cholesterol is associated with the develop-ment of atherosclerosis, where atherosclerosis of the abdominal aorta and lumbar arteries can lead to disc degeneration and LBP.136 Leino-Arjas et al found baseline serum lipids to be associated with LBP and sciatica,137, 138 and Hangai et al described an association between increased low-density lipopro-tein (LDL) and disc degeneration.139

Toda et al observed a correlation between body composition and LBP in women aged 45 to 69 years. In their study, 203 patients had LBP and the control group consisted of 127 participants. The two groups were matched for age, height, and weight. An increased percentage of body fat and in-creased waist to hip ratio were correlated with LBP.140 Vismara et al noted differences in thoracic spinal mobility and anterior pelvic tilt between obese and normal weight individuals. These mechanical and structural differences in obese individuals could contribute to the development of LBP and disc degeneration.141 Obesity can also be associated with socioeconomic factors and lifestyle choices that may affect surgical results.87

38

Study II Of the 538 patients included in the analysis in study II, the mean weight at baseline was 94.6 kg (SD, 11.9) and the average weight loss was 1.9 kg (95% CI, 1.5-2.3) 1 year after surgery and 2.0 kg (95% CI, 1.5-2.4) 2-years after surgery. Eight percent of the patients who completed the follow-up reported a CIWL (n=45). Table 7 presents the mean improvement in PROMs at the 2-year follow-up. We found no significant differences in EQ-5D, ODI, back pain (VAS) or leg pain (VAS) between weight-stable and weight-loss patients. Ten patients reported a weight gain after surgery. These patients reported lower numerical improvements in EQ-5D, ODI, back pain and leg pain compared to weight-stable patients. However, except for leg pain the differences between the weight-loss and weight-stable patients were not statistically significant.

Table 7. Mean improvement (standard deviation, SD) in patient-reported outcome measures (PROMs)between inclusion and the 2-year follow up. The Mann-Whitney U test was used with the weight-stable group as reference (ref). A Bonferroni-adjusted significance level of 0.017 was calculated (α level, 0.05).

Means (SD) Weight stable N=483

Weight loss N=45

Weight gain N=10

EQ-5D

0.22 (0.36) ref

0.23 (0.35) p=0.802

0.08 (0.36) p=0.217

ODI

14 (18) ref

15 (19) p=0.644

8 (15) p=0.248

Back pain (VAS)

18 (33) ref

27 (29) p=0.095

−6 (40) p=0.113

Leg pain (VAS)

23 (26) ref

31 (36) p=0.147

−14 (34) p=0.011

To our knowledge, this is the largest patient group used to study weight and weight changes after surgery for LSS. It is also the first large prospective study to analyze associations between improvement in PROMs and weight loss after surgery for LSS.

Our results are in contrast to those of Garcia et al. who found a 2.48 kg mean weight gain after surgery for LSS in overweight and obese patients.91 That study included 63 patients with a mean age of 53 years (range 35 to 65). In another study Vaidya et al found a minor weight gain among 58 obese and morbidly obese patients after spine surgery (mean age 53 years).142 Both of these retrospective studies included younger patients than ours, which might affect the generalizability to older patients. The follow-up period was 14-37 months in Vaidya et al. and 24-51 months in Garcia et al.

The majority of patients in our study reported no change in weight, even in those with improved functional outcome after surgery. Weight loss after surgery was not associated with a significant greater improvement in EQ-5D, ODI, back pain or leg pain. We consider our results consistent with those of

39

Garcia et al who found no correlation between weight loss and the outcome results, even though we used different outcome measures than Garcia et al.91

Sensitivity analysis Using the same cut-off level for obesity and correcting for age, sex and self-reported BMI, the number of obese patients increased (n=726), but the pro-portion of patients with reported CIWL did not change (8%). Further, the reported weight changes 1 and 2 years after surgery were not significantly different from the weight changes reported based on unadjusted data (data not shown). Using corrected BMI data, the weight-loss group reported a mean improvement in leg pain of 36 (SD, 37), which can be compared with a mean of 22 (SD, 36) in the weight-stable group (p=0.008). Weight loss was not associated with better improvement in EQ-5D, ODI or back pain when compared with weight-stable patients.

Study III Two-thirds (65%, n=235 247) of the participants had normal weight, 29% (n=106 730) were overweight, 5% (n=16 695) were obese and 2% (n=5795) were underweight. A vast majority of the construction workers were men (95%, n=345 921). The average age at baseline was 34 years (SD, 13). As expected, there was a positive association between age and BMI.

During an average follow-up of 30.7 years (11 190 944 person-years), 2381 participants were diagnosed with LSS. More than half of these cases were overweight or obese at baseline. We observed an almost linear positive association between BMI and LSS (Figure 5). Assuming a linear relation, the incidence of LSS showed an increase of 10% (95% CI, 9%-11%) per BMI unit, i.e. the lowest rates of LSS were found in lean workers and the highest rates in obese workers.

Accordingly, the IRRs of LSS varied across BMI categories, with the highest values seen in obese individuals (Table 8). Compared with normal weight, obesity was associated with a multivariable-adjusted IRR of 2.18 (95% CI, 1.87-2.53) for LSS and overweight with an IRR of 1.68 (95% CI, 1.54-1.83). Underweight workers halved their future risk of LSS (IRR 0.52, 95% CI 0.30-0.90).

As expected, the number of outcomes were lower (n=1816) in patients di-agnosed with LSS and who underwent surgery. Nevertheless, when we used surgically treated LSS as an outcome the results remained comparable with the original analysis (Table 9). Restricted cubic spline curves stratified by sex indicated a higher IRR with increasing BMI, irrespective of sex (Figure 6).

The findings are consistent with those from a cross-sectional subgroup analysis of the Framingham cohort in which an association between osteoar-

40

thritis of the lumbar spine and a tendency for radiological LSS associated with increased BMI were shown.25 The analysis included 187 participants, of whom 13 had radiological findings of LSS.

Another recent cross-sectional study (n=938), in which 78% of the partic-ipants were considered to have more than moderate radiographic central spinal stenosis found a positive association between radiological LSS and BMI in men.17

Obesity as a plausible cause for LSS An increase in body mass leads to an abnormal and altered load on the spine. Particularly noteworthy is that obese individuals have, in addition, a lower relative muscle mass than normal weight individuals, which further imposes strain on the lumbar spine.89, 143, 144 Furthermore, besides a direct biomechan-ical effect on cartilage and skeleton, indirect effects by changes in body mass can be mediated by mechanoreceptors, cytokines and growth factors. These factors have the potential to alter the properties of bone matrix, ligamentum flavum, synovium and cartilage, all of which could promote the development of osteoarthritis, hypertrophy of the ligamentum flavum and disc degenera-tion.23, 25, 26, 86, 145-147

Decreased muscle mass is also associated with insulin resistance, which further weakens the skeletal muscles and promotes systemic inflammation.143, 148 Adiponectin and leptin, hormones secreted by adipo-cytes, regulate low-grade inflammation caused by obesity and an increase in the levels of CRP, interleukins and tumor necrosis factors is related to the progression of spondylosis.89, 147, 149, 150 In addition, a high serum concentra-tion of free fatty acids is known to increase systemic inflammation and the development of osteoarthritis.148, 151, 152 Moreover, hyperlipidemia-induced atherosclerosis is proposed as a cause of disc degeneration and ischemic pain.137, 138 Finally, obesity is related to reduced walking capacity and fear of movement (kinesiophobia), events known to also increase muscle loss and pain.153, 154

41

Figure 5. The association between body mass index (BMI) and Incidence rate ratios (IRRs) of lumbar spinal stenosis displayed as a restricted cubic-spline curve based on multivariable Poisson regression analysis. A BMI of 20 kg/m2 was used as the reference. The dashed lines represent a 95% confidence interval. The model was adjusted for age (continuous), sex , occupation (22 categories) and smoking status (5 categories).

Figure 6. Restricted cubic spline curves stratified by sex indicated a higher inci-dence rate ratio (IRR) with increasing body mass index (BMI), irrespective of sex displayed as a restricted cubic-spline curve based on multivariable Poisson regres-sion analysis. A BMI of 20 kg/m2 was used as the reference. The dashed lines repre-sent a 95% confidence interval. The model was adjusted for age (continuous), occu-pation (22 categories) and smoking status (5 categories).

42

Tab

le 8

. Inc

iden

ce r

ate

ratio

s (I

RR

s) w

ith 9

5% c

onfid

ence

inte

rval

s (C

Is)

for

lum

bar

spin

al s

teno

sis

from

Poi

sson

reg

ress

ion

mod

els

by c

ate-

gori

es o

f bod

y m

ass

inde

x (B

MI)

.

BM

I ca

tego

ry

Und

erw

eigh

t

<18

.5

Nor

mal

wei

ght

18.5

-24.

9

Ove

rwei

ght

25-2

9.9

Obe

se

≥30

Per

one

BM

I un

it

N (

%),

tota

l 364

,467

5,

795

(1.6

) 23

5,24

7 (6

4.6)

10

6,73

0 (2

9.3)

16

,695

(4.

6)

N c

ases

, tot

al 2

381

13

1144

10

19

205

Pers

on-y

ears

, tot

al 1

1,19

0,94

4 18

4,53

0 7,

363,

115

3,18

1,58

3 46

1,71

6

Cru

de r

ate

(n/1

0 00

0 pe

rson

-yea

rs; 9

5% C

I)

0.7

(0.4

-1.2

) 1.

6 (1

.5-1

.7)

3.2

(3.0

-3.4

) 4.

4 (3

.9-5

.1)

IRR

, age

-adj

uste

d m

odel

(95

% C

I)

0.55

(0.

32-0

.94)

1.

00 (

refe

renc

e)

1.67

(1.

53-1

.82)

2.

18 (

1.87

-2.5

4)

1.10

(1.

09-1

.11)

IRR

, mul

tivar

iabl

e m

odel

* (9

5% C

I)

0.52

(0.

30-0

.90)

1.

00 (

refe

renc

e)

1.68

(1.

54-1

.83)

2.

18 (

1.87

-2.5

3)

1.10

(1.

09-1

.11)

*Adj

uste

d fo

r ag

e (c

ontin

uous

), s

ex, o

ccup

atio

n (2

2 ca

tego

ries

) an

d sm

okin

g st

atus

(5

cate

gori

es)

43

Tab

le 9

. Inc

iden

ce r

ate

ratio

s (I

RR

s) w

ith 9

5% c

onfid

ence

inte

rval

s (C

Is)

for

surg

ical

ly tr

eate

d lu

mba

r sp

inal

ste

nosi

s fr

om P

oiss

on r

egre

s-si

on m

odel

s by

cat

egor

ies

of b

ody

mas

s in

dex

(BM

I).

BM

I ca

tego

ry

Und

erw

eigh

t <

18.5

N

orm

al w

eigh

t 18

.5-2

4.9

Ove

rwei

ght

25-2

9.9

Obe

se

≥30

P

er o

ne B

MI

unit

N c

ases

, tot

al 1

816

10

855

793

158

Cru

de r

ate

(n/1

0 00

0 pe

rson

-yea

rs; 9

5% C

I)

0.8

(0.4

-1.4

) 1.

7 (1

.6-1

.8)

3.8

(3.6

-4.1

) 5.

2 (4

.5-6

.1)

IRR

, age

-adj

uste

d m

odel

(95

% C

I)

0.59

(0.

31-1

.10)

1.

00 (

refe

renc

e)

1.69

(1.

53-1

.87)

2.

23 (

1.88

-2.6

5)

1.10

(1.

09-1

.12)

IRR

, mul

tiva

riab

le m

odel

* (9

5% C

I)

0.56

(0.

30-1

.04)

1.

00 (

refe

renc

e)

1.71

(1.

55-1

.89)

2.

23 (

1.87

-2.6

5)

1.10

(1.

09-1

.12)

*A

djus

ted

for

age

(con

tinuo

us),

sex

, occ

upat

ion

(22

cate

gori

es)

and

smok

ing

stat

us (

5 ca

tego

ries

)

44

Study IV The themes identified in the content analysis are summarized in Table 10. A content validation with an external experienced qualitative researcher was performed to support the inductive category and theme application.131

Table 10. Themes identified in the content analysis

Themes

Experiencing an impaired physical, social and emotional lifeA good physician-patient relationship can alleviate the burden of long waiting timesStruggling for being believed and taken seriouslyWays to manage pain and disabilityAmbiguous expectations and hope for recoveryWays to handle concerns before surgery

The six themes are summarized below. Full text and quotes are presented in paper IV.

Experiencing an impaired physical, social and emotional life The pre-symptomatic experience of patients provided a context for their current suffering. Patients described having had an active physical life before the LSS became severe. The pain and limited function were also associated with sleeping disturbances. In the interviews the participants expressed pro-found feelings of shame in situations when they had problems with body functioning. The undesired sedentary lifestyle could also be perceived as a factor for weight gain and the inability to lose weight because of impaired function and walking ability.

The interviewees described isolation and how it affected relationships with family and friends. Because of LSS symptoms partnerships were de-scribed as being altered (e.g., using separate bedrooms and restrictive sexual activity).

A good physician-patient relationship can alleviate the burden of long waiting times The patients' situation of dependence and vulnerability in the relationship with the physician and long waiting times was a concern in many interviews. A good physician-patient relationship was perceived as a resource that could alleviate the negative experience of being on a waiting list. In contrast, mis-information and appointments with different unknown physicians increased unpleasant effects of long waiting periods.

Struggling for being believed and taken seriously Some patients described struggling with the healthcare system and feelings of not being believed. Moreover some patients even felt discriminated

45

against because of their age, weight or former alcohol abuse. Continuity in care and personal information and dialogue with the surgeon were depicted as particularly important. The patients felt that the MRI was very important for confirmation that their suffering and pain were real as a means to being properly diagnosed. A diagnosis also excluded other more serious conditions such as cancer.

Ways to manage pain and disability The patients described several ways to manage the disease. One strategy was to make comparisons between themselves and someone else they knew who had worse problems. Another way of handling the situation was to focus on things that were experienced as positive. Several patients described their ability to use their own resources in the sense of a “fighting spirit”. Further-more, physical activities, performing duties for other people or animals, stay-ing at work and visiting church were expressed as important events that helped them manage to carry on with life.

Ambiguous expectations and hope for recovery Concrete and common expectation were improved leg function and reduced leg pain. However, expectations could also be more generic, i.e. some pa-tients described a combination of modest expectations and expectations of full recovery after surgery.

Ways to handle concerns before surgery Some patients expressed fear of complications such as infection, anesthesia awareness or to be confined to a wheelchair after surgery. In contrast, sever-al patients describe no worries at all related to their approaching surgery.

Because of severe suffering, some patients surgery was described as the only viable option. Furthermore, earlier positive experiences from the healthcare and reports from friends successfully operated on for LSS served as important resources to reduce the stress and worry before surgery.

The results from study IV are interpreted from the salutogenic framework, with focus on resources and strengths on different levels: on a legislative level the Swedish healthcare system is equitable and patient-centered care is statutory. According to Antonovsky these are important community re-sources to achieve a high SOC.103 Furthermore, on a community level, mu-nicipalities can provide an individual who suffers from a disability, with community companions, walking aids, home adaptations and other disability tools to make life more convenient and efficient. On the personal level, so-cial support from family and friends is important and although Sweden is a secular society, church was seen as important for social interaction. In addi-tion, other people's experiences of surgery and health care could ease patient concerns before the surgical procedure. Although only a few of the inter-

46

viewees were still working, staying employed has been seen as a way to cope with chronic LBP and sciatica.155

Similar to earlier studies, our results stress a good physician-patient rela-tionship based on patient-centered care, including continuity, proper treat-ment, investigations, examination and information/dialogue, all of which produce more satisfied patients.156-160 In agreement with previous studies, an established diagnosis and a known cause for the pain are important for the patient to understand and accept the symptoms. In contrast, delegitimization (feelings of not being believed or denied) is associated with dissatisfaction and thereby caring-suffering.100, 101, 160-162 A diagnosis is also important to exclude severe spine pathology especially in the elderly who have a high frequency of comorbidities. Furthermore, information is an important and complex issue. Timely and knowledge-based information should be provided from the treating physician and operating surgeon to enhance comprehensi-bility and trust.103, 160

Expectations of recovery from surgery may be seen as the strongest source for the interviewees' to experience life-satisfaction and meaningful-ness in everyday life. Even if improved leg function and decreased leg pain were common goals, some of the expectations were similar to how life used to be perhaps 5-10 years earlier, or how life is for healthy individuals of a comparable age. Living with LSS is a physical, mental and social challenge where resources to cope with the condition may be more important than a full recovery.160, 163

47

General Discussion

Treatment and prevention of LSS are a major challenge on a personal level, for healthcare system and society. On average, two thirds of the patients are satisfied with the outcome after surgery for LSS, with lower rates of satisfac-tion in obese compared with normal weight patients.62, 75, 164, 165 To my knowledge, study I was the first to use BMI as a continuous variable. The study demonstrates increasingly higher odds of dissatisfaction with higher BMI values.165 Furthermore, a trend towards greater surgery increases both costs and surgically related complications without achieving better results, especially in obese patients.21, 56-60, 164

The indication for LSS surgery is seldom immediate and the LSS symp-toms are often unchanged or improved over time and severe deterioration is rare.37-39 Consequently, we could consider other treatment options than sur-gery, especially in patients with unsatisfactory results after surgical treat-ment. However, patients with obesity should not be denied surgery for LSS. In study I, obese patients achieved significant pain reduction, better walking ability and improved HRQoL after surgery. For a person with severe symp-toms from spinal stenosis, surgery is perhaps the best option because of the limited effect of conservative treatment.48, 133, 134

In study II, we found that even after successful surgery patients with ex-cess weight did not achieve a clinically important weight reduction on their own accord.166 If weight loss is considered important, it should be carefully monitored by health professionals with focus on physical activity and diet to help prevent muscle catabolism and impaired health. Furthermore, weight loss after surgery was not associated with significant greater improvement in EQ-5D, ODI, back pain or leg pain.

Similar to obesity, prevention is perhaps the best treatment for LSS. Con-sequently identification of risk factors associated with the disease is im-portant. Study III was the first cohort study on the relation between BMI and clinical LSS. Our major finding was that higher BMI increased the risk of clinical LSS.167

We suggest that future studies attempt to identify causality for this associ-ation, which has the potential to enhance the prevention for LSS.

Finally, many studies have relied on the pathogenetic paradigm, i.e. med-ical-disease model. At the center of the disease model is a dichotomous clas-sification of individuals as diseased or healthy. The salutogenic patient-centered approach described briefly in study IV, is a model that guides

48

health promotion by shifting attention towards the individual's resources and strengths.168

Choice of study design Although estimates from RCT´s are often considered as evidence of the highest grade, the results from many observational and RCT studies are similar.169, 170 Two recent studies on LSS and surgical treatment reported comparable findings between a RCT and a cohort study on Swespine data.35,

51 We consider a cohort observational study as the most appropriate design to answer our research questions in study I, II and III. Nevertheless, the cohort study design has limitations and there is the risk of systematic error.

The choice of a qualitative content analysis in study IV aimed to capture the experience of being a person with LSS and how life, pain and suffering are managed with respect to the disease. The collection of experiences may create a deeper understanding for patients with LSS. Furthermore, the presentation of the context and findings enables alternative interpretations. However, whether the findings can be applied to other settings is ultimately the reader's decision.

Limitations and strengths of the studies Study I and II The major strengths of study I and II are the prospective design and large sample size. The data were collected by patient self-reports from postal questionnaires unrelated to hospital visits or contact with a surgeon. Further, the registration rate in Swespine is high.114 A possible limitation is patient self-reported weight and height.118, 171

Another limitation is the response rate, i.e. only 57% and 67% of the pa-tients in study I and II, respectively, fulfilled the 2-year follow-up. However, in study I the response rate was similar for the three BMI groups and in study II no differences occurred in baseline characteristics between the re-sponders and non-responders. Furthermore, the response rate does not neces-sarily affect the generalization of the results.172 A study on a Scandinavian register for lumbar surgery found no significant differences, including func-tional outcomes, between responders and non-responders at a 2-year follow-up.173

In study II, we lack information on the cause of weight loss, which can bias the results between weight change and outcomes. Weight change in an older population is associated with severe comorbidities and mortality that could affect the estimates.174

49

Study III The main strengths of study III are the prospective design, the large sample size and the high validity of both exposure and outcome. Weight and height were measured at inclusion in the study. Date of diagnosis, date of surgery, occurrence of death and emigration data were collected through national registers known to have high accuracy.116 Complete linkage between the registers is rendered by the individual personal identification number provid-ed to all Swedish residents.175 Furthermore, the sensitivity analysis, using only cases with codes for diagnosis connected to a surgical procedure, re-vealed similar results as the original analysis. The population included both blue- and white-collar workers and the models were adjusted for type of occupation.

Several potential limitations of the study need to be discussed. Although BMI is an established measure of overweight and obesity in both a clinical and research setting, the measure has admittedly a major limitation in its inability to differentiate lean mass from fat mass.

A further conceivable limitation is that weight and height were assessed only once and we had, on average, a rather long follow-up. Further research may elucidate whether time-varying BMI predicts LSS differently from fixed measures. However, obesity as estimated by BMI in mid-life is a pre-dictor for health in later life.176

Another possible limitation is the predominance of male participants, but the increased risk of LSS with higher BMI levels was observed in both males and females.

Study IV The primary strength of study IV are the defined context and method of sampling. All eligible patients, representing various life experiences, accept-ed the offer to participate.131 BK served not only as the interviewer but also as the forthcoming surgeon. This dual role could have resulted in observer bias and is therefore a limitation of the study. Still, to minimize this potential bias, the interviews were analyzed by all authors and the results from the content analysis were validated by an external observer. Furthermore, the quotations translated by an external assistant who had Bachelor of education degrees in both Swedish and English increase the credibility.

Although SOC is a validated quantitative measure and has been applied as a theoretical framework for qualitative research, the concept has some limi-tations: the direction of the causality between SOC and health is uncertain, and SOC could be a measure of physical and mental health instead of a pre-dictor of health; and little is known about whether interventions to increase SOC on a personal and community level have any long-term effects.

50

A social, mental and physical stressor such as living with LSS may be handled differently by men and women.177 Because the majority of the pa-tients in the present study were females, our results could have been affected by the skewed distribution of participants.

The precision of an estimate Variance or standard deviation (SD) is often used to describe the variability of the observations used to calculate the arithmetic mean.126 p 24-25

Variance: s2= ∑ (xi- ̅ )2 /n-1

Standard deviation: s=√

Furthermore, the precision of an estimate can be described with a standard error, often incorporated into an interval estimate (CI).126 p34-35

Standard error of the mean, SEM= s/√

Sample mean with 95% CI, mean ± t0.05 x SEM

A narrow CI indicates a precise estimate, and as shown above the width of the CI is strongly dependent on the sample size. Consequently, a mean from a large population is in general a more precise estimate than an estimate from a small population.126 p34

Random and systematic errors The opposite of precision is random error. Rothman defines random error as "variability in the data we cannot readily explain" 178 p123 A random error could also be defined as sampling variability. Random error can be mini-mized through increased sample size, improved sampling procedures, reduc-tion of measurement variability and proper choice of statistical methods.

Although large populations can reduce random errors, systematic errors also known as bias, could remain. Systematic error/bias could be defined as the difference between the estimate and the true value. Bias can be classified into three main categories: selection bias, information bias and confound-ing.178 p96

Selection bias The population included in a register is not randomized, and patients who select to participate in a register, can differ from non-participants. Selection bias occurs when the association between exposure and outcome differs be-

51

tween participants and non-participants. Selection bias could have been in-troduced into study I if the association between obesity and inferior out-comes was different between participants and non-participants. The associa-tion between obesity and outcomes after LSS in non-participants is unknown and must therefore be inferred.178 p 96-98

Information bias Information about the exposure and outcome collected in a register can be misclassified leading to information bias. For example, an obese person can be classified as overweight because of erroneous weight data, which could result in a misclassification in the exposure. Another example is the classifi-cation of LSS in Swespine and the NPR. As we know it the diagnosis LSS has not been validated either in Swespine or in NPR. However, the website for Swedish Society of Spine Surgeons recommends a radiological criterion of a CSA≤75 mm2 together with typical LSS symptoms for diagnosis.19 Fur-thermore, the diagnosis, indications for surgery and predictors for surgery are discussed on a regular basis at annual meetings. These regular discus-sions may serve to increase the validity of the diagnosis of LSS in the regis-ters, but does not exclude erroneous registration.

Confounding A confounder could be defined as a factor that is mixing the effects. A con-founder must be a cause or proxy for the disease and associated with the exposure. Furthermore, a confounder must be imbalanced between the expo-sure groups.178 p 108 An example is increased BMI as a risk factor for LSS. BMI is a proxy for biomechanical and metabolic causal mediators, and also associated with increased age. Age is a proxy for the disease and a cause for LSS regardless of BMI category. Finally, age is not a mediator between BMI and LSS. Therefore, age could be considered a confounder.128 s129-134 Com-mon methods to deal with confounding is to identify covariates and adjust for them by stratification or in a regression model. However, recent studies have shown that these criteria are insufficient and thus questioned the tradi-tional approach to confounding. A process, referred to as directed acyclic graphs (DAGs), is a novel model to achieve unbiased estimates.179, 180

Differential and non-differential misclassification Misclassification can be divided into differential and non-differential. Dif-ferential misclassification occurs if there were an association between the misclassification in the exposure and outcome. For example, an association between a misclassified BMI category and outcome after surgery. However, we consider this potential misclassification in our cohort studies as non-

52

differential, which in general makes the effect estimates converge towards one another.

Generalization Generalization of our findings in study I and III, should be regarded as an elaboration of the theory that an increased BMI through causal pathways is a cause for LSS, as well as a cause for inferior results after surgery for LSS.178 The biologic theory is based on earlier animal experiments and research on humans.

Statistical representativeness is considered by many authors as a criterion for external validity and generalization of the findings from a study. Howev-er, generalization is a complex process. Thus it is important to bring knowledge from several branches of science, rather than consider generaliza-tion as a statistical process.128, 178 Furthermore, opposite to survey sampling, statistical representativeness is not necessarily a major concern and should perhaps even be avoided.172, 178

Animal studies are often based on a strictly selected population of rats, rabbits or some other animal that rather increase the possibilities for biologic representativeness. Although these strictly selected populations are not sta-tistically representative of a general population of humans, they help us to understand human diseases.178 Another parallel can be drawn to smoking as a cause of premature death, which was first established in male British phy-sicians,172, 181, 182 which is a highly selected subgroup of the population.

Trustworthiness Although there are similarities between qualitative and quantitative research, the concepts and methods used to achieve trustworthiness differ between qualitative and quantitative analysis131, 183 Preconception, reflexivity, credi-bility and transferability are important concepts to achieve trustworthiness in qualitative studies. Some examples from study IV are given in relation to each concept.

Preconceptions Preconceptions could be summarized as the researcher's back pack. Previous personal and professional experiences and prestudy beliefs may affect how researchers judge and interpret information and the course of action they take.183 Several preconceptions in study IV need to be discussed. The inter-views and surgery was performed by the same person (BK). Therefore, the patients are placed in a position of dependence that may affect how they

53

express themselves. This thesis focuses on obesity and overweight and the present author (BK) thought that weight loss should be a common expecta-tion among patients with obesity. Furthermore, BK's preoperative belief about this patient group in general was positive and these patients are often perceived as modest and grateful. BK is a 44-year-old male Caucasian with a high level of education, which may affect the interview and how the infor-mation is interpreted.

Reflexivity Reflexivity involves a process of deeper reflection that should be addressed to increase the trustworthiness. A researcher´s interaction with participants is a subjective process. To accept the subjective nature of qualitative research is not enough, however. How the researcher, participants and findings are presented for the reader is important to avoid bias.183 Our main reflections were not to "force" the patients into the study, and we sought how to de-crease dependency and create a more neutral environment.

BK as a single interviewer would be a conceivable limitation. Should we include several interviewers? One purpose of this study was to educate and develop BK as a qualitative researcher. In addition, the preoperative inter-view could also raise questions about the treatment. Therefore the research team agreed on using BK as the sole interviewer.

Our interpretation of the findings reflect our preconceptions which in-clude models for suffering, coping and the salutogenic model. How should we make this transparent for the reader?

Credibility Credibility summarizes the context, method of data collection and selection of patients. Credibility also includes the selection of the most meaningful units and whether the categories and themes actually cover the data.131 To increase credibility, representative quotations, were translated by an external assistant who had a Bachelor of education degrees in both Swedish and Eng-lish. Furthermore, a specific research nurse gave information about the study once the patients were on the waiting list for surgery and therefore guaran-teed that the interview could not affect the planned surgery or follow-up. The interviews were performed at the patient hostel as this provides a more neu-tral environment than the orthopedic reception or ward. The interviewer (BK) was dressed in casual clothing rather than the traditional white coat.

Credibility was further increased by seeking agreement on the categories and themes among co-researches and an external expert.131

We used the salutogenic model as a sensitizing concept, which creates a possibility for alternative interpretations.130

54

Transferability Although findings from a qualitative study cannot be generalized, the expe-riences and needs described in this study are not necessarily restricted to a specific context, hospital or country.161, 162 Transferability reflects whether the findings can be applied to other settings.131, 183 Even if the author gives suggestions about transferability, it is the reader's choice and interpretation whether the findings are transferable to another setting or population.

55

Conclusions

1. Obesity is associated with inferior PROMs (ODI, EQ-5D, back pain and leg pain) and a higher rate of dissatisfaction compared with normal weight patients 2 years after surgery for LSS.

2. Weight reduction 1 and 2 years after surgery is modest in obese patients,

with only 8% reporting a CIWL. In comparison with weight-stable pa-tients, weight loss in obese patients after surgery for LSS is not associat-ed with greater improvement in PROMs: ODI, EQ-5D, back pain and leg pain.

3. Obesity and overweight are associated with an increased risk to develop

LSS. Our findings indicate that obesity is one plausible explanation for the increased number of patients with clinical LSS.

4. Being a patient with LSS inevitably involves a certain amount of pain

and suffering. Yet, the patient with LSS is also a person with his or her own resources who is able to find strategies to cope with the suffering, of the disease or who has support structures for doing so. Both physi-cians and patients need to work towards a salutogenic perspective, focus-ing on resources to improve care, making it more comprehensible, man-ageable and meaningful.

Clinical implications Our studies expand the knowledge about excess weight and LSS. This in-creased knowledge is potentially important for patients, physicians and health care providers.

The results from this thesis: Provide patients, physicians and health care providers accurate infor-

mation about different outcomes after surgery that are related to BMI. Show that the weight change in general is modest, even after successful

surgery.

56

Inform citizens, patients, physicians and health care providers that high BMI could be a risk factor for the development of clinical LSS, even if future studies are important to confirm our finding.

Increase not only the knowledge and understanding about being a patient with LSS but also suggest a shift towards delineating what resources and strengths are needed to cope with the disease.

Future perspectives A major limitation in studies on LSS is that the definition of LSS differ from one study to another. Thus a national and international agreement on a clear and unambiguous definition of LSS is warranted. A definition of LSS could also be implemented in registers on spine surgery to increase internal validi-ty. An international consensus on outcomes for LBP has been attained through the International Consortium for Health Outcomes Measures, (ICHOM).184 With a similar national and international definition, covariates and outcome measures, results from different studies can be more readily compared.

This thesis has emphasized several aspects of obesity and LLS. Future studies could focus on the following questions.

Can a weight loss program improve the symptoms and results after surgery for LSS in obese persons? Can a weight loss program reduce the progression of LSS? Can the causality between obesity and LSS be further described through studies on leptin concentration in humans and/or Mendelian randomization studies?185 Can a salutogenic in- and outpatient routine increase the patients' sat-isfaction with care, as well as improve the outcome after surgery?

57

Acknowledgements

First I want to thank Karl Michaëlsson and Bengt Sandén my head and co-supervisors for inviting me to the science of epidemiology and statistics. I would also like to express my gratitude for sharing your knowledge and for always being available for discussions an support, as well as making your-selves available in connections with the sometimes intensive work with the thesis. Special thanks to Göran Sjödén my co-supervisor for taking research to the Orthopedic Department in Sundsvall, and for his continuous support. Thank you Bengt Järvholm and the construction workers from the Bygghälsan cohort for making study III possible. Marika Augutis and Mats Jong thank you for teaching me how and why to perform qualitative studies. Warm thanks to my colleagues in Sundsvall for your support during my doctoral studies. I would especially like to thank; Thomas Wikström, my mentor and friend. You taught me spine surgery, shared his vast clinical knowledge and invited me to the Swedish Society of Spinal Surgeons. You also introduced me to research through studies on vertebroplasty and LSS. Anna Grauers my colleague in spine surgery, mak-ing my work twice as rewarding and gratifying. Thanks for the support and sharing the workload during the work with the thesis. Affectionate thanks to Arkan Sayed-Noor for inspiring me to be become a PhD student and for his continuous support. You are always a source of knowledge, inspiration and discussions. Sebastian Mukka You are supposed to be under my supervision, but it is often the case that I learn more than I teach. Deep thanks for your support and knowledge base.

Anders Holtz Thank you for sharing your immense knowledge in spine sur-gery, especially cervical spine and neurosurgery. You have always encour-aged me in research and working career.

58

Michel Guez Your skilful surgical technique and unfailing enthusiasm for spine surgery is a source for friendship and a productive regional coopera-tion. I am happy to express my thanks to the team at the new outpatient depart-ment for patients with spinal disorders, Thomas Torstensson, Ulrika Söderholm, Anders Christensen, Susanne Svensk-Lindfors, Anna Grauers, Urszula Lind, Anne-Charlotte Söderlind including our employers Unni Ivefors and Sigmund Norberg. Together, we have improved the care for patients with spinal disorders.186 Special thanks to: Thomas Torstensson for your good advices and support with the thesis summary. Susanne Svensk-Lindfors for your excellent work with the transcripts in study IV. Thank you Helene Sörell and Elieann Broman for your careful an excellent work with our patients in study IV.

Warm thanks to The Swedish Society of Spinal Surgeons, for the research grant and allowing me to use the Swespine Register for research. I would like to extend my deep appreciations to all the patients included in the thesis for their participation and for making this thesis possible and im-portant. I would like to acknowledge the Department of Research and Development in Västernorrland for their research grant, as well as for providing a produc-tive and energetic environment for research. I would not least like to express enormous thanks to my father, Anders, for being my inspiration to become a physician and researcher, but also for shar-ing your vast knowledge in epidemiology and statistics. My mother, Ingrid deserves my heartiest thanks for her intellectual and emotional support over the years She taught me that writing an article is more about relentless determination and inspiration rather than writing ca-pacity and skills. My beautiful and intelligent wife Ann who inspired me to become a PhD student, and wonderful children Julia and Lukas. I thank you with my heart for your support, patience and ...well everything.

59

Svensk sammanfattning

Lumbal spinal stenos (LSS) är den vanligaste indikationen för ländryggski-rurgi i Sverige, med en ökad förekomst både nationellt och internationellt. Denna ökning är multifaktoriell, om övervikt är en riskfaktor för att utveckla LSS är inte tidigare känt.

LSS är en förträngning av spinalkanalen i ländryggen orsakad av ben och mjukdelsförändringar. Symptomen debuterar ofta i 50-60 års ålder. Typiska symtom är förutom ryggsmärta en läges- och belastningskorrelerad bensmär-ta, s.k. pseudoclaudicatio.

Även om majoriteten av de patienter som opereras för spinal stenos blir nöjda, är en alltför stor del av patienterna missnöjda. Att identifiera faktorer som påverkar resultatet efter en operation är därför viktigt. Hos patienter med högt BMI, kan det finnas en förväntan att operationen ska bidra till en viktnedgång genom en förhöjd fysisk prestationsförmåga efter operationen.

Patienternas perspektiv är viktigt. Tidigare studier har beskrivit upplevel-ser och erfarenheter av rygg och bensmärta, men ett fåtal har inkluderat spe-cifikt patienter med LSS. Studie I: En kohortstudie där 2633 patienter från svenska ryggregistret in-kluderades. Det patientrapporterade resultatet (livskvalitet, funktion, smärta, patientnöjdhet) 2 år efter operation var sämre hos patienter med BMI≥30 (fetma) jämfört med BMI<25 (normalvikt).

Studie II: Från samma kohort som studie I inkluderades 538 patienter med BMI≥30. Under uppföljningstiden på två år reducerade patienterna sin me-delvikt med 2 kg, och enbart åtta procent (45/538) rapporterade en kliniskt relevant viktreduktion. Vi fann ingen korrelation mellan viktreduktion och operationsresultat.

Studie III: Genom samkörning med det nationella Patientregistret och Bygghälsan som inkluderar 389 132 personer, fann vi att övervikt och fetma är en riskfaktor för att utveckla LSS.

Studie IV: Individuella intervjuer med 18 patienter. Analysen utmynnade i sex stycken teman, som beskrev patienternas lidande, men också resurser och styrkor för att klara av sina besvär.

60

Avhandlingens resultat kan bidra till: Utifrån resultatet i studie I, kan vi ge patienter och vårdgivare bättre

information om det förväntade operationsresultatet korrelerat till BMI. Informera om att kroppsvikten vanligen förändras marginellt efter en

operation. Om målet är en substantiell bestående viktreduktion krävs tro-ligen, vilket andra visat, ett aktivt deltagande i viktreduceringsprogram. Interventionsstudier med målet att reducera vikten efter kirurgi för LSS bör dock först göras för att påvisa effektiviteten av ett sådant viktreduk-tionsprogram för just denna patientgrupp

Informera patienter och vårdgivare att ett ökat BMI kan vara en riskfak-tor för att utveckla en operationskrävande spinal stenos, även om våra resultat behöver bekräftas av ytterligare studier.

Öka kunskapen om upplevelsen av att vara patient med spinal stenos, och hur man kan klara av sina besvär. Även om resultaten som redovisas i den kvalitativa studien inte är generaliserbara så tar den upp flera vikti-ga aspekter av att vara patient med spinal stenos och vikten av att foku-sera på sjukvårdens värdegrund och patientens resurser/styrkor. Huruvi-da sådana interventioner kan förbättra patienternas upplevelse av vården och resultatet efter ryggkirurgi, får framtida studier visa.

61

References

1. Genevay S and Atlas SJ. Lumbar spinal stenosis. Best Pract Res Clin Rheumatol 2010;24:253-65.

2. Arbit E and Pannullo S. Lumbar stenosis: a clinical review. Clin Orthop Relat Res 2001:137-43.

3. Nixon JE. Spinal stenosis. First ed. London: Edward Arnold, 1991. 4. Verbiest H. A radicular syndrome from developmental narrowing of the lum-

bar vertebral canal. J Bone Joint Surg Br 1954;36-B:230-7. 5. Kalichman L, Cole R, Kim DH, et al. Spinal stenosis prevalence and associa-

tion with symptoms: the Framingham Study. Spine J 2009;9:545-50. 6. Suri P, Rainville J, Kalichman L, et al. Does this older adult with lower ex-

tremity pain have the clinical syndrome of lumbar spinal stenosis? JAMA 2010;304:2628-36.

7. Szpalski M and Gunzburg R. Lumbar spinal stenosis in the elderly: an over-view. Eur Spine J 2003;12 Suppl 2:S170-5.

8. Katz JN and Harris MB. Clinical practice. Lumbar spinal stenosis. N Engl J Med 2008;358:818-25.

9. Fritz JM, Delitto A, Welch WC, et al. Lumbar spinal stenosis: a review of current concepts in evaluation, management, and outcome measurements. Arch Phys Med Rehabil 1998;79:700-8.

10. Andreisek G, Deyo RA, Jarvik JG, et al. Consensus conference on core radio-logical parameters to describe lumbar stenosis - an initiative for structured re-porting. Eur Radiol 2014;24:3224-32.

11. Sigmundsson FG, Kang XP, Jonsson B, et al. Correlation between disability and MRI findings in lumbar spinal stenosis: a prospective study of 109 patients operated on by decompression. Acta Orthop 2011;82:204-10.

12. Schonstrom N, Bolender NF, Spengler DM, et al. Pressure changes within the cauda equina following constriction of the dural sac. An in vitro experimental study. Spine (Phila Pa 1976) 1984;9:604-7.

13. Athiviraham A, Yen D, Scott C, et al. Clinical correlation of radiological spi-nal stenosis after standardization for vertebral body size. Clin Radiol 2007;62:776-80.

14. Ogikubo O, Forsberg L and Hansson T. The relationship between the cross-sectional area of the cauda equina and the preoperative symptoms in central lumbar spinal stenosis. Spine (Phila Pa 1976) 2007;32:1423-8.

15. Sirvanci M, Bhatia M, Ganiyusufoglu KA, et al. Degenerative lumbar spinal stenosis: correlation with Oswestry Disability Index and MR imaging. Eur Spine J 2008;17:679-85.

16. Boden SD, Davis DO, Dina TS, et al. Abnormal magnetic-resonance scans of the lumbar spine in asymptomatic subjects. A prospective investigation. J Bone Joint Surg Am 1990;72:403-8.

62

17. Ishimoto Y, Yoshimura N, Muraki S, et al. Associations between radiographic lumbar spinal stenosis and clinical symptoms in the general population: the Wakayama Spine Study. Osteoarthritis Cartilage 2013;21:783-8.

18. Jansson KA, Blomqvist P, Granath F, et al. Spinal stenosis surgery in Sweden 1987-1999. Eur Spine J 2003;12:535-41.

19. Webpage for the Society of Swedish Spine Surgeons. Available at: http://www.4s.nu/. Accessed September 9, 2015.

20. Ciol MA, Deyo RA, Howell E, et al. An assessment of surgery for spinal ste-nosis: time trends, geographic variations, complications, and reoperations. J Am Geriatr Soc 1996;44:285-90.

21. Deyo RA, Mirza SK, Martin BI, et al. Trends, major medical complications, and charges associated with surgery for lumbar spinal stenosis in older adults. JAMA 2010;303:1259-65.

22. Du Bois M, Szpalski M and Donceel P. A decade's experience in lumbar spine surgery in Belgium: sickness fund beneficiaries, 2000-2009. Eur Spine J 2012;21:2693-703.

23. Sairyo K, Biyani A, Goel V, et al. Pathomechanism of ligamentum flavum hypertrophy: a multidisciplinary investigation based on clinical, biomechani-cal, histologic, and biologic assessments. Spine (Phila Pa 1976) 2005;30:2649-56.

24. Perruccio AV, Power JD and Badley EM. Revisiting arthritis prevalence pro-jections--it's more than just the aging of the population. J Rheumatol 2006;33:1856-62.

25. Kalichman L, Guermazi A, Li L, et al. Association between age, sex, BMI and CT-evaluated spinal degeneration features. J Back Musculoskelet Rehabil 2009;22:189-95.

26. Guilak F. Biomechanical factors in osteoarthritis. Best Pract Res Clin Rheumatol 2011;25:815-23.

27. Birkmeyer JD, Reames BN, McCulloch P, et al. Understanding of regional variation in the use of surgery. Lancet 2013;382:1121-9.

28. Gandhi R, Woo KM, Zywiel MG, et al. Metabolic syndrome increases the prevalence of spine osteoarthritis. Orthop Surg 2014;6:23-7.

29. Battie MC, Ortega-Alonso A, Niemelainen R, et al. Lumbar spinal stenosis is a highly genetic condition partly mediated by disc degeneration. Arthritis Rheumatol 2014;66:3505-10.

30. Burkle A, Caselli G, Franceschi C, et al. Pathophysiology of ageing, longevity and age related diseases. Immun Ageing 2007;4:4.

31. Battie MC, Videman T, Gill K, et al. 1991 Volvo Award in clinical sciences. Smoking and lumbar intervertebral disc degeneration: an MRI study of identi-cal twins. Spine (Phila Pa 1976) 1991;16:1015-21.

32. Videman T and Battie MC. The influence of occupation on lumbar degenera-tion. Spine (Phila Pa 1976) 1999;24:1164-8.

33. Wahlstrom J, Burstrom L, Nilsson T, et al. Risk factors for hospitalization due to lumbar disc disease. Spine (Phila Pa 1976) 2012;37:1334-9.

34. Malmivaara A, Slatis P, Heliovaara M, et al. Surgical or nonoperative treat-ment for lumbar spinal stenosis? A randomized controlled trial. Spine (Phila Pa 1976) 2007;32:1-8.

35. Forsth P, Michaelsson K and Sanden B. Does fusion improve the outcome after decompressive surgery for lumbar spinal stenosis?: A two-year follow-up study involving 5390 patients. Bone Joint J 2013;95-B:960-5.

36. Stromqvist B, Fritzell P, Hagg O, et al. Swespine: the Swedish spine register : the 2012 report. Eur Spine J 2013;22:953-74.

63

37. Amundsen T, Weber H, Nordal HJ, et al. Lumbar spinal stenosis: conservative or surgical management?: A prospective 10-year study. Spine (Phila Pa 1976) 2000;25:1424-35; discussion 35-6.

38. Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical versus nonsurgical thera-py for lumbar spinal stenosis. N Engl J Med 2008;358:794-810.

39. Johnsson KE, Rosen I and Uden A. The natural course of lumbar spinal steno-sis. Acta Orthop Scand Suppl 1993;251:67-8.

40. Simotas AC, Dorey FJ, Hansraj KK, et al. Nonoperative treatment for lumbar spinal stenosis. Clinical and outcome results and a 3-year survivorship analy-sis. Spine (Phila Pa 1976) 2000;25:197-203; discussions -4.

41. May S and Comer C. Is surgery more effective than non-surgical treatment for spinal stenosis, and which non-surgical treatment is more effective? A system-atic review. Physiotherapy 2013;99:12-20.

42. Tadokoro K, Miyamoto H, Sumi M, et al. The prognosis of conservative treatments for lumbar spinal stenosis: analysis of patients over 70 years of age. Spine (Phila Pa 1976) 2005;30:2458-63.

43. Friedly JL, Comstock BA, Turner JA, et al. A randomized trial of epidural glucocorticoid injections for spinal stenosis. N Engl J Med 2014;371:11-21.

44. Chou R, Hashimoto R, Friedly J, et al. Epidural Corticosteroid Injections for Radiculopathy and Spinal Stenosis: A Systematic Review and Meta-analysis. Ann Intern Med 2015;163:373-81.

45. Meng H, Fei Q, Wang B, et al. Epidural injections with or without steroids in managing chronic low back pain secondary to lumbar spinal stenosis: a meta-analysis of 13 randomized controlled trials. Drug Des Devel Ther 2015;9:4657-67.

46. Atlas SJ, Keller RB, Robson D, et al. Surgical and nonsurgical management of lumbar spinal stenosis: four-year outcomes from the maine lumbar spine study. Spine (Phila Pa 1976) 2000;25:556-62.

47. Atlas SJ, Keller RB, Wu YA, et al. Long-term outcomes of surgical and non-surgical management of lumbar spinal stenosis: 8 to 10 year results from the maine lumbar spine study. Spine (Phila Pa 1976) 2005;30:936-43.

48. Kovacs FM, Urrutia G and Alarcon JD. Surgery versus conservative treatment for symptomatic lumbar spinal stenosis: a systematic review of randomized controlled trials. Spine (Phila Pa 1976) 2011;36:E1335-51.

49. Delitto A, Piva SR, Moore CG, et al. Surgery versus nonsurgical treatment of lumbar spinal stenosis: a randomized trial. Ann Intern Med 2015;162:465-73.

50. Benz RJ and Garfin SR. Current techniques of decompression of the lumbar spine. Clin Orthop Relat Res 2001:75-81.

51. Försth P, Carlsson T, Michaëlsson K, et al. No benefit from fusion in decompressive surgery for lumbar spinal stenosis. 2 year-results from the

Swedish spinal stenosis study, a mulitcenter RCT of 229 patients. Eurospine annual meeting, presentation #68. Lyon,France, 2014.

52. Nerland US, Jakola AS, Solheim O, et al. Minimally invasive decompression versus open laminectomy for central stenosis of the lumbar spine: pragmatic comparative effectiveness study. BMJ 2015;350:h1603.

53. Overdevest GM, Jacobs W, Vleggeert-Lankamp C, et al. Effectiveness of posterior decompression techniques compared with conventional laminectomy for lumbar stenosis. Cochrane Database Syst Rev 2015;3:CD010036.

54. Moojen WA, Arts MP, Jacobs WC, et al. Interspinous process device versus standard conventional surgical decompression for lumbar spinal stenosis: ran-domized controlled trial. BMJ 2013;347:f6415.

64

55. Jansson KA, Nemeth G, Granath F, et al. Spinal stenosis re-operation rate in Sweden is 11% at 10 years--a national analysis of 9,664 operations. Eur Spine J 2005;14:659-63.

56. Wimmer C, Gluch H, Franzreb M, et al. Predisposing factors for infection in spine surgery: a survey of 850 spinal procedures. J Spinal Disord 1998;11:124-8.

57. Patel N, Bagan B, Vadera S, et al. Obesity and spine surgery: relation to peri-operative complications. J Neurosurg Spine 2007;6:291-7.

58. Djurasovic M, Bratcher KR, Glassman SD, et al. The effect of obesity on clin-ical outcomes after lumbar fusion. Spine (Phila Pa 1976) 2008;33:1789-92.

59. Koutsoumbelis S, Hughes AP, Girardi FP, et al. Risk factors for postoperative infection following posterior lumbar instrumented arthrodesis. J Bone Joint Surg Am 2011;93:1627-33.

60. Seicean A, Alan N, Seicean S, et al. Impact of increased body mass index on outcomes of elective spinal surgery. Spine (Phila Pa 1976) 2014;39:1520-30.

61. Aalto TJ, Malmivaara A, Kovacs F, et al. Preoperative predictors for postoper-ative clinical outcome in lumbar spinal stenosis: systematic review. Spine (Phi-la Pa 1976) 2006;31:E648-63.

62. Turner JA, Ersek M, Herron L, et al. Surgery for lumbar spinal stenosis. At-tempted meta-analysis of the literature. Spine (Phila Pa 1976) 1992;17:1-8.

63. Sigmundsson FG, Kang XP, Jonsson B, et al. Prognostic factors in lumbar spinal stenosis surgery. Acta Orthop 2012;83:536-42.

64. Yukawa Y, Lenke LG, Tenhula J, et al. A comprehensive study of patients with surgically treated lumbar spinal stenosis with neurogenic claudication. J Bone Joint Surg Am 2002;84-A:1954-9.

65. Jonsson B, Annertz M, Sjoberg C, et al. A prospective and consecutive study of surgically treated lumbar spinal stenosis. Part II: Five-year follow-up by an independent observer. Spine (Phila Pa 1976) 1997;22:2938-44.

66. Frazier DD, Lipson SJ, Fossel AH, et al. Associations between spinal deformi-ty and outcomes after decompression for spinal stenosis. Spine (Phila Pa 1976) 1997;22:2025-9.

67. Katz JN, Stucki G, Lipson SJ, et al. Predictors of surgical outcome in degener-ative lumbar spinal stenosis. Spine (Phila Pa 1976) 1999;24:2229-33.

68. Eriksson M and Lindstrom B. Antonovsky's sense of coherence scale and its relation with quality of life: a systematic review. J Epidemiol Community Health 2007;61:938-44.

69. Sinikallio S, Lehto SM, Aalto T, et al. Low sense of coherence interferes with lumbar spinal stenosis patients' postoperative recovery: a prospective one-year follow-up study. J Health Psychol 2011;16:783-93.

70. Ng LC, Tafazal S and Sell P. The effect of duration of symptoms on standard outcome measures in the surgical treatment of spinal stenosis. Eur Spine J 2007;16:199-206.

71. Radcliff KE, Rihn J, Hilibrand A, et al. Does the duration of symptoms in patients with spinal stenosis and degenerative spondylolisthesis affect out-comes?: analysis of the Spine Outcomes Research Trial. Spine (Phila Pa 1976) 2011;36:2197-210.

72. Iversen MD, Daltroy LH, Fossel AH, et al. The prognostic importance of pa-tient pre-operative expectations of surgery for lumbar spinal stenosis. Patient Educ Couns 1998;34:169-78.

73. Herno A, Airaksinen O, Saari T, et al. Surgical results of lumbar spinal steno-sis. A comparison of patients with or without previous back surgery. Spine (Phila Pa 1976) 1995;20:964-9.

65

74. Cobo Soriano J SRM, Fabregate Fuente M, et al. Predictors of outcome after decompressive lumbar surgery and instrumented posterolateral fusion. Eur Spine J 2010;19:1841-8.

75. Giannadakis C, Nerland US, Solheim O, et al. Does Obesity Affect Outcomes After Decompressive Surgery for Lumbar Spinal Stenosis? A Multicenter, Ob-servational, Registry-Based Study. World Neurosurg 2015. Epub ahead of print.

76. Sanden B, Forsth P and Michaelsson K. Smokers show less improvement than nonsmokers two years after surgery for lumbar spinal stenosis: a study of 4555 patients from the Swedish spine register. Spine (Phila Pa 1976) 2011;36:1059-64.

77. Lehto MU and Honkanen P. Factors influencing the outcome of operative treatment for lumbar spinal stenosis. Acta Neurochir (Wien) 1995;137:25-8.

78. Andreshak TG, An HS, Hall J, et al. Lumbar spine surgery in the obese patient. J Spinal Disord 1997;10:376-9.

79. Gepstein R, Shabat S, Arinzon ZH, et al. Does obesity affect the results of lumbar decompressive spinal surgery in the elderly? Clin Orthop Relat Res 2004:138-44.

80. Ain MC, Chang TL and Schkrohowsky JG. Laminectomies and achondroplasia: does body mass index influence surgical outcomes? Am J Med Genet A 2007;143A:1032-7.

81. Athiviraham A, Wali ZA and Yen D. Predictive factors influencing clinical outcome with operative management of lumbar spinal stenosis. Spine J 2011;11:613-7.

82. Haslam D. Obesity: a medical history. Obes Rev 2007;8 Suppl 1:31-6. 83. Finucane MM, Stevens GA, Cowan MJ, et al. National, regional, and global

trends in body-mass index since 1980: systematic analysis of health examina-tion surveys and epidemiological studies with 960 country-years and 9.1 mil-lion participants. Lancet 2011;377:557-67.

84. Swinburn BA, Sacks G, Hall KD, et al. The global obesity pandemic: shaped by global drivers and local environments. Lancet 2011;378:804-14.

85. Kaaria S, Kaila-Kangas L, Kirjonen J, et al. Low back pain, work absenteeism, chronic back disorders, and clinical findings in the low back as predictors of hospitalization due to low back disorders: a 28-year follow-up of industrial employees. Spine (Phila Pa 1976) 2005;30:1211-8.

86. Liuke M, Solovieva S, Lamminen A, et al. Disc degeneration of the lumbar spine in relation to overweight. Int J Obes (Lond) 2005;29:903-8.

87. Janke EA, Collins A and Kozak AT. Overview of the relationship between pain and obesity: What do we know? Where do we go next? J Rehabil Res Dev 2007;44:245-62.

88. Stone AA and Broderick JE. Obesity and pain are associated in the United States. Obesity (Silver Spring) 2012;20:1491-5.

89. Vincent HK, Heywood K, Connelly J, et al. Obesity and weight loss in the treatment and prevention of osteoarthritis. PM R 2012;4:S59-67.

90. Berenbaum F, Eymard F and Houard X. Osteoarthritis, inflammation and obe-sity. Curr Opin Rheumatol 2013;25:114-8.

91. Garcia RM, Messerschmitt PJ, Furey CG, et al. Weight loss in overweight and obese patients following successful lumbar decompression. J Bone Joint Surg Am 2008;90:742-7.

92. Khoueir P, Black MH, Crookes PF, et al. Prospective assessment of axial back pain symptoms before and after bariatric weight reduction surgery. Spine J 2009;9:454-63.

66

93. Roffey DM, Ashdown LC, Dornan HD, et al. Pilot evaluation of a multidisci-plinary, medically supervised, nonsurgical weight loss program on the severity of low back pain in obese adults. Spine J 2011;11:197-204.

94. Lidar Z, Behrbalk E, Regev GJ, et al. Intervertebral disc height changes after weight reduction in morbidly obese patients and its effect on quality of life and radicular and low back pain. Spine (Phila Pa 1976) 2012;37:1947-52.

95. Rothberg AE, McEwen LN, Kraftson AT, et al. The impact of weight loss on health-related quality-of-life: implications for cost-effectiveness analyses. Qual Life Res 2014;23:1371-6.

96. Willett WC, Dietz WH and Colditz GA. Guidelines for healthy weight. N Engl J Med 1999;341:427-34.

97. Romero-Corral A, Somers VK, Sierra-Johnson J, et al. Accuracy of body mass index in diagnosing obesity in the adult general population. Int J Obes (Lond) 2008;32:959-66.

98. Gallagher D, Visser M, Sepulveda D, et al. How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am J Epidemiol 1996;143:228-39.

99. Battie MC, Jones CA, Schopflocher DP, et al. Health-related quality of life and comorbidities associated with lumbar spinal stenosis. Spine J 2012;12:189-95.

100. Lindholm L and Eriksson K. To understand and alleviate suffering in a caring culture. J Adv Nurs 1993;18:1354-61.

101. Eriksson K. Understanding the world of the patient, the suffering human being: the new clinical paradigm from nursing to caring. Adv Pract Nurs Q 1997;3:8-13.

102. Lazarus RS. Coping theory and research: past, present, and future. Psychosom Med 1993;55:234-47.

103. Antonovsky A. Health, Stress and Coping. First ed. San Francisco; London: Jossey-Bass Publishers, 1979.

104. Antonovsky A. Unraveling the mystery of health. First ed. San Francisco: Jossey-Bass Publishers, 1987.

105. Eriksson M and Lindstrom B. Validity of Antonovsky's sense of coherence scale: a systematic review. J Epidemiol Community Health 2005;59:460-6.

106. Eriksson M and Lindstrom B. Antonovsky's sense of coherence scale and the relation with health: a systematic review. J Epidemiol Community Health 2006;60:376-81.

107. Lindstrom B and Eriksson M. Contextualizing salutogenesis and Antonovsky in public health development. Health Promot Int 2006;21:238-44.

108. Malterud K and Hollnagel H. Talking with women about personal health re-sources in general practice. Key questions about salutogenesis. Scand J Prim Health Care 1998;16:66-71.

109. Malterud K and Hollnagel H. Encouraging the strengths of women patients. A case study from general practice on empowering dialogues. Scand J Public Health 1999;27:254-9.

110. Hollnagel H, Malterud K and Witt K. Men's self-assessed personal health resources: approaching patients' strong points in general practice. Fam Pract 2000;17:529-34.

111. Persson L, Larsson G, Ohlsson O, et al. Acute leukaemia or highly malignant lymphoma patients' quality of life over two years: a pilot study. Eur J Cancer Care (Engl) 2001;10:36-47.

112. Strang S and Strang P. Spiritual thoughts, coping and 'sense of coherence' in brain tumour patients and their spouses. Palliat Med 2001;15:127-34.

67

113. Nyman P, Palenius K, Panula H, et al. Patients' experiences of shoulder prob-lems prior to and following intervention Physiother Theory Pract 2012;28:221-31.

114. Stromqvist B, Fritzell P, Hagg O, et al. One-year report from the Swedish National Spine Register. Swedish Society of Spinal Surgeons. Acta Orthop Suppl 2005;76:1-24.

115. Bygghälsokohorten. Available at: http://www.phmed.umu.se/enheter/envmed/ forskning/forskningsomraden/bygghalsokohorten/. Accessed may 08, 2015.

116. Ludvigsson JF, Andersson E, Ekbom A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health 2011;11:450.

117. Strömqvist B FP, Hägg O, Knutsson B, Sandén B. The swedish spine register, 2014 report. 08/12/2014, 2014. Available at: http://www.4s.nu/pdf/ Re-port_2014_Swespine_Engl_ver_141204.pdf, 2014.

118. Nyholm M, Gullberg B, Merlo J, et al. The validity of obesity based on self-reported weight and height: Implications for population studies. Obesity (Silver Spring) 2007;15:197-208.

119. Zanoli G, Stromqvist B and Jonsson B. Visual analog scales for interpretation of back and leg pain intensity in patients operated for degenerative lumbar spine disorders. Spine (Phila Pa 1976) 2001;26:2375-80.

120. Solberg TK, Olsen JA, Ingebrigtsen T, et al. Health-related quality of life as-sessment by the EuroQol-5D can provide cost-utility data in the field of low-back surgery. Eur Spine J 2005;14:1000-7.

121. Chapman JR, Norvell DC, Hermsmeyer JT, et al. Evaluating common out-comes for measuring treatment success for chronic low back pain. Spine (Phila Pa 1976) 2011;36:S54-68.

122. Giantomaso T, Makowsky L, Ashworth NL, et al. The validity of patient and physician estimates of walking distance. Clin Rehabil 2003;17:394-401.

123. Okoro T, Qureshi A, Sell B, et al. The accuracy of assessment of walking distance in the elective spinal outpatients setting. Eur Spine J 2010;19:279-82.

124. Godil SS, Parker SL, Zuckerman SL, et al. Determining the quality and effec-tiveness of surgical spine care: patient satisfaction is not a valid proxy. Spine J 2013;13:1006-12.

125. Douketis JD, Macie C, Thabane L, et al. Systematic review of long-term weight loss studies in obese adults: clinical significance and applicability to clinical practice. Int J Obes (Lond) 2005;29:1153-67.

126. Petrie A and Sabin C. Medical statistics at a Glance. third ed. Oxford: Wiley-Blackwell, 2012.

127. Heinzl H and Kaider A. Gaining more flexibility in Cox proportional hazards regression models with cubic spline functions. Comput Methods Programs Bi-omed 1997;54:201-8.

128. Rothman KJ, Sander G and Timothy LL. Modern Epidemiology. Third ed. Philadelphia: Wolters Kluwer Health Lippincott Williams & Wilkins, 2008.

129. Harrell Jr F. General aspects of fitting regression models. Regression modeling strategies: Springer Science+Business Media, Inc., 2001:23.

130. Patton M. Qualitative research and evaluation methods. 4th ed. United States: SAGE, 2014.

131. Graneheim UH and Lundman B. Qualitative content analysis in nursing re-search: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today 2004;24:105-12.

132. Javid MJ and Hadar EJ. Long-term follow-up review of patients who under-went laminectomy for lumbar stenosis: a prospective study. J Neurosurg 1998;89:1-7.

68

133. Rihn JA, Radcliff K, Hilibrand AS, et al. Does obesity affect outcomes of treatment for lumbar stenosis and degenerative spondylolisthesis? Analysis of the Spine Patient Outcomes Research Trial (SPORT). Spine (Phila Pa 1976) 2012;37:1933-46.

134. McGuire KJ, Khaleel MA, Rihn JA, et al. The effect of high obesity on out-comes of treatment for lumbar spinal conditions: subgroup analysis of the spine patient outcomes research trial. Spine (Phila Pa 1976) 2014;39:1975-80.

135. Miscio G, Guastamacchia G, Brunani A, et al. Obesity and peripheral neuropa-thy risk: a dangerous liaison. J Peripher Nerv Syst 2005;10:354-8.

136. Kauppila LI. Atherosclerosis and disc degeneration/low-back pain--a systemat-ic review. Eur J Vasc Endovasc Surg 2009;37:661-70.

137. Leino-Arjas P, Kaila-Kangas L, Solovieva S, et al. Serum lipids and low back pain: an association? A follow-up study of a working population sample. Spine (Phila Pa 1976) 2006;31:1032-7.

138. Leino-Arjas P, Kauppila L, Kaila-Kangas L, et al. Serum lipids in relation to sciatica among Finns. Atherosclerosis 2008;197:43-9.

139. Hangai M, Kaneoka K, Kuno S, et al. Factors associated with lumbar interver-tebral disc degeneration in the elderly. Spine J 2008;8:732-40.

140. Toda Y, Segal N, Toda T, et al. Lean body mass and body fat distribution in participants with chronic low back pain. Arch Intern Med 2000;160:3265-9.

141. Vismara L, Menegoni F, Zaina F, et al. Effect of obesity and low back pain on spinal mobility: a cross sectional study in women. J Neuroeng Rehabil 2010;7:3.

142. Vaidya R, Carp J, Bartol S, et al. Lumbar spine fusion in obese and morbidly obese patients. Spine (Phila Pa 1976) 2009;34:495-500.

143. Vincent HK, Raiser SN and Vincent KR. The aging musculoskeletal system and obesity-related considerations with exercise. Ageing Res Rev 2012;11:361-73.

144. Han KS, Rohlmann A, Zander T, et al. Lumbar spinal loads vary with body height and weight. Med Eng Phys 2013;35:969-77.

145. Piscoya JL, Fermor B, Kraus VB, et al. The influence of mechanical compres-sion on the induction of osteoarthritis-related biomarkers in articular cartilage explants. Osteoarthritis Cartilage 2005;13:1092-9.

146. Kosaka H, Sairyo K, Biyani A, et al. Pathomechanism of loss of elasticity and hypertrophy of lumbar ligamentum flavum in elderly patients with lumbar spi-nal canal stenosis. Spine (Phila Pa 1976) 2007;32:2805-11.

147. Berenbaum F and Sellam J. Obesity and osteoarthritis: what are the links? Joint Bone Spine 2008;75:667-8.

148. Dandona P, Aljada A, Chaudhuri A, et al. Metabolic syndrome: a comprehen-sive perspective based on interactions between obesity, diabetes, and inflam-mation. Circulation 2005;111:1448-54.

149. Dumond H, Presle N, Terlain B, et al. Evidence for a key role of leptin in oste-oarthritis. Arthritis Rheum 2003;48:3118-29.

150. Griffin TM, Huebner JL, Kraus VB, et al. Extreme obesity due to impaired leptin signaling in mice does not cause knee osteoarthritis. Arthritis Rheum 2009;60:2935-44.

151. Griffin TM, Huebner JL, Kraus VB, et al. Induction of osteoarthritis and meta-bolic inflammation by a very high-fat diet in mice: effects of short-term exer-cise. Arthritis Rheum 2012;64:443-53.

152. Brunner AM, Henn CM, Drewniak EI, et al. High dietary fat and the develop-ment of osteoarthritis in a rabbit model. Osteoarthritis Cartilage 2012;20:584-92.

69

153. Hemmingsson E and Ekelund U. Is the association between physical activity and body mass index obesity dependent? Int J Obes (Lond) 2007;31:663-8.

154. Vincent HK, Omli MR, Day T, et al. Fear of movement, quality of life, and self-reported disability in obese patients with chronic lumbar pain. Pain Med 2011;12:154-64.

155. de Vries HJ, Brouwer S, Groothoff JW, et al. Staying at work with chronic nonspecific musculoskeletal pain: a qualitative study of workers' experiences. BMC Musculoskelet Disord 2011;12:126.

156. Hjortdahl P and Laerum E. Continuity of care in general practice: effect on patient satisfaction. BMJ 1992;304:1287-90.

157. Glass RM. The patient-physician relationship. JAMA focuses on the center of medicine. JAMA 1996;275:147-8.

158. Laine C and Davidoff F. Patient-centered medicine. A professional evolution. JAMA 1996;275:152-6.

159. Laerum E, Indahl A and Skouen JS. What is "the good back-consultation"? A combined qualitative and quantitative study of chronic low back pain patients' interaction with and perceptions of consultations with specialists. J Rehabil Med 2006;38:255-62.

160. Hopayian K and Notley C. A systematic review of low back pain and sciatica patients' expectations and experiences of health care. Spine J 2014;14:1769-80.

161. Ong BN, Konstantinou K, Corbett M, et al. Patients' own accounts of sciatica: a qualitative study. Spine (Phila Pa 1976) 2011;36:1251-6.

162. Davis RE, Vincent C, Henley A, et al. Exploring the care experience of pa-tients undergoing spinal surgery: a qualitative study. J Eval Clin Pract 2013;19:132-8.

163. Huber M, Knottnerus JA, Green L, et al. How should we define health? BMJ 2011;343:d4163.

164. Niggemeyer O, Strauss JM and Schulitz KP. Comparison of surgical proce-dures for degenerative lumbar spinal stenosis: a meta-analysis of the literature from 1975 to 1995. Eur Spine J 1997;6:423-9.

165. Knutsson B, Michaelsson K and Sanden B. Obesity is associated with inferior results after surgery for lumbar spinal stenosis: a study of 2633 patients from the Swedish spine register. Spine (Phila Pa 1976) 2013;38:435-41.

166. Knutsson B, Michaelsson K and Sanden B. Obese patients report modest weight loss after surgery for lumbar spinal stenosis: a study from the Swedish spine register. Spine (Phila Pa 1976) 2014;39:1725-30.

167. Knutsson B, Sanden B, Sjoden G, et al. Body Mass Index and Risk for Clinical Lumbar Spinal Stenosis: A Cohort Study. Spine (Phila Pa 1976) 2015;40:1451-6.

168. Knutsson B, Jong M, Sjödén G, et al. Waiting for lumbar spinal stenosis sur-gery: suffering, resources to cope and expectations. Submitted 2015.

169. Benson K and Hartz AJ. A comparison of observational studies and random-ized, controlled trials. N Engl J Med 2000;342:1878-86.

170. Concato J, Shah N and Horwitz RI. Randomized, controlled trials, observa-tional studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-92.

171. Gorber SC, Tremblay M, Moher D, et al. A comparison of direct vs. self-report measures for assessing height, weight and body mass index: a systematic re-view. Obes Rev 2007;8:307-26.

172. Rothman KJ, Gallacher JE and Hatch EE. Why representativeness should be avoided. Int J Epidemiol 2013;42:1012-4.

70

173. Solberg TK, Sorlie A, Sjaavik K, et al. Would loss to follow-up bias the out-come evaluation of patients operated for degenerative disorders of the lumbar spine? Acta Orthop 2011;82:56-63.

174. Seidell JC and Visscher TL. Body weight and weight change and their health implications for the elderly. Eur J Clin Nutr 2000;54 Suppl 3:S33-9.

175. Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, et al. The Swedish per-sonal identity number: possibilities and pitfalls in healthcare and medical re-search. Eur J Epidemiol 2009;24:659-67.

176. Sun Q, Townsend MK, Okereke OI, et al. Adiposity and weight change in mid-life in relation to healthy survival after age 70 in women: prospective co-hort study. BMJ 2009;339:b3796.

177. Carmel S, Anson O, Levenson A, et al. Life events, sense of coherence and health: gender differences on the kibbutz. Soc Sci Med 1991;32:1089-96.

178. Rothman KJ. Epidemiology : an introduction. First ed. New York: Oxford University Press, 2002.

179. Shrier I and Platt RW. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008;8:70.

180. VanderWeele TJ, Hernan MA and Robins JM. Causal directed acyclic graphs and the direction of unmeasured confounding bias. Epidemiology 2008;19:720-8.

181. Doll R and Hill AB. The mortality of doctors in relation to their smoking hab-its; a preliminary report. Br Med J 1954;1:1451-5.

182. Doll R, Peto R, Boreham J, et al. Mortality in relation to smoking: 50 years' observations on male British doctors. BMJ 2004;328:1519.

183. Malterud K. Qualitative research: standards, challenges, and guidelines. Lancet 2001;358:483-8.

184. Clement RC, Welander A, Stowell C, et al. A proposed set of metrics for standardized outcome reporting in the management of low back pain. Acta Orthop 2015;86:523-33.

185. Smith GD and Ebrahim S. 'Mendelian randomization': can genetic epidemiolo-gy contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:1-22.

186. Knutsson B and Torstensson T. Nya rutiner på ortopeden gav effektivare vård och nöjdare patienter - Sjukgymnast och team gör första bedömning vid nybe-sök till ryggkirurg. Lakartidningen 2015;112.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1148

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-264589

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015