Diabetes Mellitus Related Morbidity

102

JOURNAL OF INSURANCE MEDICINECopyright Q 2003 Journal of Insurance MedicineJ Insur Med 2003;35:102–113

REVIEW

Diabetes Mellitus Related Morbidity, Risk ofHospitalization and DisabilityAbdelouahed Naslafkih, MD, MSc; Francois Sestier, MD, PhD, FACC

Objective.—To investigate the rates of complications, hospitaliza-tions and disabilities attributable to type 1 and type 2 diabetes mel-litus (DM) combined, unless otherwise noted.

Methodology.—Risk assessment of DM-related morbidity, hospi-talizations and disabilities using data from the medical literatureand health statistics on the population. Calculation of morbidity,hospitalization, and disability ratios (MbR, HR, DR) will allow com-parison of observed rates in people with DM to those reported inthe nondiabetic population.

Results.—MbRs vary according to the morbid condition studied:;300% at age 45–64 years for ischemic heart disease, 533% for cor-onary heart disease or stroke, 226% to 388% for chronic heart failure,560% for peripheral vascular disease, 380% for neuropathy at age35–74 years, 890% to 2225% for lower limb amputations, 1458% to3287% for end-stage renal disease. For ocular complications: cata-racts, 165% to 232%; glaucoma 140% to 330%; trouble seeing, 180%to 231%; blindness at age $65 years, 517%. Higher values are notedat younger ages. HR: 200% to 409%. DR: 217% to 328%.

Conclusion.—Among diseases, DM is one of the leading andgrowing causes of hospital admission and disability. Precise riskassessment of morbidity is essential for realistic underwriting ofhealth and disability insurance.

Address: C.P. 6128 Succ. Centre-ville, Montreal (QC) Canada H3C3J7; phone: (514) 343-7606; fax:(514) 343-7074; e-mail: [email protected].

Correspondent: Dr. Francois Sestier,University of Montreal, Programmeof Insurance Medicine

Key words: Diabetes mellitus, mor-bidity, disability, hospitalization.

Diabetes mellitus (DM) is one of the mostcommon chronic diseases in the world.

Its prevalence approaches 8% to 10% of theadult population of the United States or Eu-rope, and affects more than 100 million casesworldwide. It is estimated that about 1.5 mil-lion Canadians (5% of the population) havebeen diagnosed with DM and that nearly 3%to 5% of adults have undiagnosed type 2 di-abetes.1 In the United States, DM incidence isgrowing, not only due to aging of the Amer-ican population. Trends show that the nativeminority and the elderly are even more af-

fected by DM.2 Between 1980 and 1996, thenumber of persons with diagnosed DM in theUnited States increased by 2.7 million.2 In1996, about 8.5 million persons in the UnitedStates (3.2% of population) reported that theyhad DM.3 This figure increased to 12 millionAmericans reporting they had DM in 2000.2

Because DM is often asymptomatic in the ear-ly stages, it is frequently undiagnosed formany years. Insulin resistance and hypergly-cemia develop gradually, while pathologicaland functional changes in various target tis-sues may not be severe enough for patients

NASLAFKIH ET AL—DIABETES MELLITUS

103

to notice any of the classic symptoms of thedisease.4

For more precise correlation with the glu-cose tolerance test and glycemia associatedwith microvascular complications, a new cri-terion for the diagnosis of DM was publishedin 1997 by the American Diabetes Association(ADA).5 The diagnostic level for fasting plas-ma glucose (FPG) was set at $7.0 mmol/L tominimize the discrepancy in 1995 WorldHealth Organization (WHO) criteria, bywhich diabetes was diagnosed by either FPG$7.8 mmol/L or 2-hour post load plasmaglucose (2-h PG) $11.1 mmol/L during the75 g oral glucose tolerance test.6 In 1999, theWHO made further recommendations to de-fine DM.7,8

The predominant clinical form of DM is type2 or non-insulin-dependent, which accounts formore than 90% of all cases.9 Non-insulin-de-pendent diabetes mellitus (NIDDM) may bethe most rapidly growing chronic disease inthe world. Its long-term complications, includ-ing retinopathy, nephropathy, neuropathy andaccelerated macrovascular disease, cause majormorbidity and mortality.10 Diabetes is associ-ated with a worldwide socioeconomic burden:DM-associated disabilities reduce the quality oflife and result in massive, rising direct and in-direct medical costs.11,12 In 1992, diabetics con-stituted 4% of the US population but accountedfor 14.6% of total US health care expenditures($105 billion). It was estimated that per capitaannual health care expenditures were morethan 4 times greater than for nondiabetics.13 Al-though these costs derive from a variety of fac-tors, chronic complications account for morethan 50% of the total cost of diabetes.14

Underwriting disability and health insur-ance in people with diabetes could be refinedthrough a review of the medical literatureduring the last decade. As suggested by Sing-er,15 life table methodology could be appliedto determine morbidity ratios (MbR) using anapproach similar to mortality studies. MbR isthe ratio of the ‘‘observed’’ morbid eventsrates in diabetic patients to either the rates innondiabetic patients or to the rates in the gen-eral population as specified (‘‘expected’’ mor-

bid events rate); the ratio is a decimal that ismultiplied by 100 to obtain a percent. Hos-pitalization and disability risks could be es-tablished by calculating hospitalization ratios(HR) and disability ratios (DR). The ‘‘nor-mal’’ or standard risk is 100%.

We summarized representative MbR, HRand DR results calculated from data present-ed in publications surveyed in our literaturereview (large cohorts, large registries) for Ta-bles 1–4. Details of our review of diabetes re-lated morbidity, hospitalization and disabilityare included in the following text.

DIABETES MELLITUS AND MORBIDITY

The risk of morbidity attributable to DMcan be assessed by calculating MbRs for eachmorbid event from observed rates in the di-abetic population compared with rates in thenondiabetic population. MbRs are expressedas percentages, 100% being the standard or‘‘normal risk,’’ 200% corresponding to 2-foldrisk, 300% to 3-fold risk, and so on. We willreview the macrovascular (cardiovascular)and microvascular (ocular, renal and neuro-logic) complications related to DM. Tables 1and 2 summarize the MbRs for different mor-bid events. In different age groups, we cal-culated from data presented in representativestudies surveyed in our literature review.

Cardiovascular Complications

Previously-diagnosed DM, newly-diag-nosed DM and impaired glucose toleranceare important determinants of the risk ofclinical cardiovascular disease.16 There iswidespread agreement that type 2 diabetesincreases the age-related risk of coronary ar-tery disease and peripheral, as well as cere-brovascular disease in most populations.9

Based on the large 1989 US National HealthInterview Survey (NHIS), the prevalence ofself-reported ischemic heart disease (IHD),heart and rhythm disorders, as well as ath-erosclerosis in the United States was higheramong adults with than without self-report-ed diabetes.17 The percentage of American

JOURNAL OF INSURANCE MEDICINE

104

Table 1. Morbidity Ratios (MbR) of Macrovascular Complications by Age Group: Diabetics vs. Non-DiabeticsBased on Representative Studies Referenced in This Review

Macrovascular ComplicationsAge

(Years)Morbidity Ratios

MbR (%)

Ischemic heart disease17

Heart disease or stroke23

18–4445–64$6535–64$65

1,350304172533162

Stroke45

Chronic heart failure19

45–6445–6465–74$75

356500247228

Peripheral vascular disease24

Lower limb amputation

Carotid artery disease40

All agesAll ages37

Adjusted38

45–64

560890

1,258240

adults reporting heart conditions by agegroup and diabetes status in 1989 was: 2.7%(diabetics) vs 0.2% (non-diabetics) at age 18–44 years, 14.3% vs 4.7% at age 45–64 years,and 20.0% vs 11.6% at age $65 years for IHD.The rates for heart and rhythm disorderswere: 1.2% vs 1.9% for age 18–44 years, 7.1%vs 4.0% for age 45–64 years, and 7.0% vs 6.6%for age $65 years. The rates for atheroscle-rosis were 6.8% vs 1.2% for the age group 45–64, and 7.6% vs 3.6% for age $65 years.17

Comparing rates in subjects with DM tothose without DM, the MbR for ischemicheart disease is 1350% in age group 18–44years, 304% in age group 45–64 years, and172% in age $65 years (Table 1). For heartand rhythm disorders, the MbR is 63%, 177%and 106% in these age groups, respectively.The MbR for atherosclerosis is 567% in agegroup 45–64 years and 211% in age group$65 years.17

The Framingham Heart Study18 first dem-onstrated an increased risk of congestiveheart failure (CHF) in patients with diabetesover 20 years ago. A recent investigation19 ofsubjects from the Kaiser Permanente North-west Division Diabetes Registry reported thatCHF was found in 11.8% (n 5 1131) of type2 diabetic subjects and 4.5% (n 5 435) of con-trols with a mean age of 63 years at baseline.

Incident cases of CHF were observed in 7.7%of type 2 diabetic subjects free of CHF atbaseline and in 3.4% of controls over 30months of follow-up. Annual incidence ratesincreased with age in both groups: 10 casesper 1000 persons with type 2 DM vs 2 casesper 1000 among persons without type 2 DMaged 45–54 years, 25 vs 5 per 1000 in agegroup 55–64 years, 37 vs 15 per 1000 in agegroup 65–74 years, and 80 vs 35 per 1000 insubjects over 75 years. From these rates ofthis large cohort, we calculated a MbR of500% for the age group 45–64 years, 247% forthe age 65–74 years, and 228% for the ageover 75 years19 (Table 1). Poor glycemic con-trol may be associated with an increased riskof heart failure.20 The relationship betweenexposure to hyperglycemia over time and therisk of macrovascular or microvascular com-plications in patients with type 2 diabetes hasbeen reported for the 3055 patients followedfor 7.9 years in the UK Prospective DiabetesStudy.21 A comparison of all complications re-lated to diabetes was made by event/person-years in the worst category of patients (HbA1c

$ 10%) vs those in the low-normal category(HbA1c , 6%). The MbR was calculated as260% for all complications related to diabetes,388% for CHF, 152% for fatal or nonfatal myo-


105

Table 2. Morbidity Ratios (MbR) for MicrovascularComplications by Age Groups: Diabetics vs Non-

Diabetics Based on Representative Studies Referencedin This Review

MicrovascularComplications

Age(Years)

MorbidityRatios

MbR (%)

End Stage Renal Disease1973–197558

19962

AdjustedAdjusted

1,4583,287

NeuropathySaint Luis VDS 199550

NHANES II, 198151

20–4445–6465–7435–74

1,000700457266

CataractsUSA, 199559

Canada23

$65All adults.65

232679156

GlaucomaUSA59

Canada23

All adults$65.65

821295140

Blindness62

Germany Adjusted 517

Table 3. Hospitalization* Ratios (HR) by Age Groupin the United States, Canada, Finland and Scotland:Diabetics vs Non-Diabetics Based on Representative

Studies Referenced in This Review

Age(Years)

HospitalizationRatios HR (%)

USA2 All adults18–4445–6465–74$75

305411239168184

Canada23

Finland65

Scotland64

35–64.65All adultsAll adults

216172409200

* At least 1 hospital stay for any cause in the year.

cardial infarction and 187% for fatal or non-fatal stroke.

The Honolulu Heart Program22 prospec-tively followed a cohort of 3006 Japanese-American men age 45–68 years old for 23years. The cohort was divided into 4 catego-ries of glucose tolerance: low normal, highnormal, asymptomatic hyperglycemia andknown diabetes. Coronary heart disease(CHD) incidence was 14.1 per 1000 person-years in known diabetes, 8.8 in asymptomatichyperglycemia, 5.9 in the high-normal cate-gory, and 5.0 in the low-normal category.Comparison of the CHD incidence rate in theknown diabetes category to the low normalcategory (considered as reference) gave aMbR of 282%.

In Canada,23 approximately 21% of patientswith DM have heart disease or stroke com-pared to 4% of those without DM; the great-est difference is found in the 35–64 age group,in which the overall prevalence is 16% among

those with DM, and 3% among those with-out, for a MbR of 533%. This MbR for heartdisease or stroke decreases for age $65 years.(Table 1)

Peripheral Vascular Disease (PVD) andLower Limb Amputations

The incidence and prevalence of lower ex-tremity peripheral vascular disease (PVD) in-creases with age in both diabetic and non-diabetic subjects.24 The prevalence of PVD isabout 2%-6% for men and women youngerthan 50 years of age, rising to .7% in thoseolder than 70 years of age. Worse arterial dis-ease and poorer outcome are reported in pa-tients with DM.25

In a Rochester, Minn, population-based co-hort of diabetic patients, the cumulative in-cidence of lower limb PVD measured bypulse deficit was 21.3 per 1000 person-yearsof diabetes for men and 17.6 per 1000 person-years for women.26 The actuarially-estimatedcumulative incidence of lower limb PVD was15% at 10 years after the initial diagnosis ofdiabetes and 45% after 20 years. Using a his-tory of intermittent claudication as the crite-rion for lower extremity artery disease(LEAD), the Framingham study27 identified alower incidence of PVD in diabetic persons inthat population (12.6 per 1000 person-years


106

Table 4. Disability Ratios (DR) in the United States and Canada. Diabetics vs. Non-Diabetics Based onRepresentative Studies Referenced in This Review

DR (%)

$1 Disability Days in2-week Period in the Year

Disability for anEntire 2-week Period

USA:36

Canada:23

age $18 yearsage 35–64 years

217209 320

for men and 8.4 per 1000 person-years forwomen). For individuals without diabetes inthe Framingham population, incidence rateswere 3.3 and 1.1 per 1000 person-years formen and women, respectively. We calculateda MbR of 382% for men and 764% for women.

Based on 285,900 diagnoses listed in thehospital discharge records from the 1989–1991National Hospital Discharge Survey, lowerlimb PVD was found in ;3% of diabetic hos-pitalizations vs 0.6% of non-diabetic hospital-izations.24 The MbR was 560%. (Table 1)

The loss of a limb is a frequent complica-tion of DM, the result of diabetic foot prob-lems, such as ulcers and infection. The risk oflower limb amputation is increased up to 15-fold in people with diabetes.28 More than halfof all nontraumatic lower limb amputationsoccur in patients with diagnosed DM. Be-tween 1980 and 1986, the number of DM-re-lated hospital discharges post-lower limb am-putations increased from 36,000 to 86,000 peryear. Almost 60% of diabetes-related lowerlimb amputations occurred among patientsover 65 years, with an average length of stayof 13.7 days. Lower limb amputations per1000 persons rose with age. In 1996, the ratesamong patients over 75 years were more thantwice those among persons under age 65. Therate was higher among men than women andamong blacks than whites.2 Higher mortalityrates are reported among patients with DMwho have had a lower limb amputation.29,30,31

Half of those who had an amputation willrequire amputation of the remaining limbwithin 5 years.32 This morbidity results inhigh medical and rehabilitation costs. About10% of diabetes-related health care costs areassociated with lower limb amputation.33 The

duration of hospital stay has been identifiedas one of the main determinants of cost as-sociated with lower limb amputation.34 Thereis a large variation in length of hospital stayvarying from 16.0 days in the United States,724.7 days in Australia,28 and 40.8 days in theNetherlands.33

Amputation rate was determined in 1044NIDDM patients aged 45–64 years followedfor 7 years in Finland.35 The rate of amputa-tion was 6.5% in men and 5.3% in women.High fasting plasma glucose (FPG) at base-line and duration of DM were associated withtwice the rate of amputation. Similarly, gly-cemic control measured at baseline by HbA1c

was an important predictor. A dose-responserelationship was found between plasma glu-cose or HbA1c, and the rate of amputation.The effect of hyperglycemia was seen clearly,even after adjustment for other cardiovascu-lar risk factors. Signs of peripheral neuropa-thy, bilateral absence of the Achilles tendonreflex as well as vibration sense were alsopredictors for increased rate of amputation.

In 1989, the NHIS36 determined that therate of amputation was 10 times higheramong diabetic vs non-diabetic individualsaged $18 years (2.8% vs 0.29%). In an anal-ysis from Medicare claims37 for 1996–1997,the yearly rate of major amputations was3.38/1000 individuals with DM vs 0.38/1000without DM. We calculated a MbR of 890%.(Table 1)

In a retrospective study38 involving a pop-ulation of 253,000 inhabitants in eastern Fin-land, among 477 patients (85 diabetic men,127 non-diabetic men, 169 diabetic women,and 96 non-diabetic women) followed for 6years, the annual age-adjusted amputation


107

incidence was 349.1/100,000 for diabetic men,33.9/100,000 for non-diabetic men, 329.4/100,000 for diabetic women, and 17.2/100,000for non-diabetic women. We calculated aMbR of 1030% in men, 1392% in women, and1258% after adjustment, comparing incidencerates observed in diabetic patients with thosein the non-diabetic population. (Table 1)

Based on records from 3 German hospitals,Trautner and colleagues39 reported amputa-tion incidence rates standardized to the entireGerman population of 33.8/100,000 per yearfor all amputations per total population,209.2/100,000 for amputations in diabetic in-dividuals per diabetic population, and 9.4/100,000 for amputations in non-diabetic in-dividuals per non-diabetic population. Com-parison of the incidence rates of amputationin patients with DM to those without DMgave an MbR of 2225%, as opposed to 619%for the general population.

Carotid Artery Disease

Carotid artery disease is also higher in pa-tients with DM than in those without DM. Across-sectional study in Finland40 compareddata from 10-year follow-up in a cohort of133 patients aged 45–64 years with NIDDM,and 144 age-matched, non-diabetic controlsrandomly selected from the general registry.The frequency of carotid atherosclerosis wasmarkedly higher in patients with NIDDM(25.0%) than in non-diabetic subjects (10.4%).The MbR was 240% (Table 1). A similar MbRwas calculated from an Italian study41 com-paring the 46% prevalence of carotid athero-sclerosis in 54 diabetic subjects to 18% in 54matched non-diabetic controls. The MbR was255%.

Stroke

Diabetes is a clear risk factor for stroke,particularly ischemic stroke. From prospec-tive studies, relative risk for stroke42 is 2.5(95% CI 1.6 to 4.0), possibly because manywith DM also suffer from hypertension.43

Based on the 1990–92 NHIS, the prevalence

of persons in the US population who reporta medical history of stroke increases withage, from 1.7% of those aged 45–64 years to8.1% of those aged $75 years.42

In the Nurses Health Study44 (116,177women followed for 8 years), the age-adjust-ed risk of stroke for diabetic vs non-diabeticwomen was 4.1 (95% CI 2.8 to 6.1). The Ath-erosclerosis Risk in Communities (ARIC)study,45 included more than 12,000 adultsaged 45–64 years without cardiovascular dis-ease at baseline and followed them for 6–8years for ischemic stroke occurrence. InARIC, the incidence rates per 1000 person-years (adjusted for age, race and sex) were5.38 for diabetic subjects vs 1.51 for non-di-abetics. We calculated an MbR of 356%. (Ta-ble 1)

In Canada, an increased risk of stroke hasbeen reported in a study at Kahnawake (Que-bec), a Mohawk community near Montreal.46

Thirteen percent of people with DM had hadstroke vs 3% of a comparable group withoutDM with an MbR of 433%. This same studyfound that the risk of macrovascular diseasewas 6 times higher among people with dia-betes. This ratio was comparable to that in thenonaboriginal population.

Neuropathy

Neuropathy is a common complication ofdiabetes, affecting 60%-70% of patients. Dis-tal symmetrical polyneuropathy is the mostcommon type.47 Neuropathy is the most com-mon diabetic complication in developedcountries of the world, leading to high mor-bidity and mortality, and resulting in a hugeeconomic burden for diabetes care.48

In the Rochester Diabetic NeuropathyStudy,49 the prevalence was 60% for any neu-ropathy, 47% for distal polyneuropathy, 34%for carpal tunnel syndrome, and 5% for au-tonomic neuropathy. The frequency distribu-tion by type of neuropathy and the severityof distal neuropathy were similar for insulin-dependent diabetes mellitus (IDDM) andNIDDM.

In the San Luis Valley Diabetes Study,50 def-


108

inite neuropathy was found in 28% of His-panic and white patients with NIDDM. Theprevalence of distal symmetrical neuropathyincreased from 17% at 0–4 years of NIDDMduration to 50% at $15 years duration. Neu-ropathy rates were approximately 5 timeshigher in those with NIDDM compared tonon-diabetic subjects. The prevalence of neu-ropathy in subjects with DM vs controls witha normal glucose tolerance test by age groupwas 10% vs 1% at age 20–44 years, 28% vs4% at age 45–64 years, and 32% vs 7% at age65–74 years. Comparison of these rates gaveMbRs of 1000%, 700%, and 457% for theseage groups, respectively.

In the 1989 NHIS,36 the prevalence of sen-sory neuropathy symptoms was 38% inNIDDM subjects vs 10% in those without di-abetes. This prevalence increased similarly inmen and women with the duration of diabe-tes. The MbR was 380% when comparingthese rates.36 The absence of knee and/or an-kle jerks was ascertained in the NationalHealth and Nutrition Examination SurveyII.51 Among all persons with diabetes aged35–74 years, the absence of 1 or both reflexeswas noted in 12.5%, compared to 4.7% insubjects without diabetes. The MbR of 266%when comparing patients with a medical his-tory of DM to those without DM. (Table 2)Intensive treatment of diabetes with near-normalization of glycemia reduced the 5-yearprevalence of neuropathy by 60%-70%, com-pared with the conventional treatmentgroup.52

Nephropathy and Renal Complications

According to the US Renal Data System, 53

in the past 2 decades there has been a contin-ual increase in the incidence of end-stage re-nal disease (ESRD) among diabetics, predom-inantly among those with type 2 diabetes.The proportion of patients with both ESRDand DM rose from 27% to 36% between1982–1992. Both the prevalence and incidenceof ESRD are approximately twice what theywere 10 years ago.54

Statistics from the Canadian Organ Re-

placement Register55 Annual Report 1998 in-dicate an increase in the proportion of pa-tients with newly-diagnosed kidney failurewho also have diabetes, from 16% in 1981 to28% in 1996. A similar trend has occurred inother developed countries as well, makingESRD in patients with type 2 diabetes a med-ical problem of worldwide dimension.56

In the United States,2 DM is the leadingcause of ESRD and kidney failure requiringdialysis or transplantation. It accounts for ap-proximately 40% of all new cases. The num-ber of persons who began treatment forESRD attributable to diabetes (ESRD-DM) in-creased from 6981 in 1984 to 31,647 in 1996.Between 1984 and 1996, the age-adjusted in-cidence of ESRD-DM treatment per 100,000persons with DM rose by 200%, from 117.4to 378.1. Among both blacks and whites, therelative increase was also greater than 200%.From 1982 through 1986, treatment for ESRDrelated to diabetes increased more than 10%each year.57 In 1996 in the United States,2 theage-adjusted rate of initiation of treatment forESRD related to diabetes was 11.5 per 100,000population, but rose to 378.1 per 100,000 indiabetics. Comparison of these rates resultedin an MbR of 3287%. (Table 2)

The Multiple Risk Factor Intervention Tri-al58 screened 332,544 men aged 35–57 yearsfrom 18 US cities in 1973 to 1975. Over anaverage follow-up of 16 years, there were 136cases of ESRD in 5147 diabetic men and 678cases in 327,397 non-diabetic men. The age-adjusted incidence of all-cause ESRD in dia-betic men was 199.8 per 100,000 person-yearscompared with 13.7 per 100,000 person-yearsin their non-diabetic counterparts, with anMbR of 1458%.

Ocular Complications

Retinopathy, cataracts and glaucoma are 3complications of DM that may lead to blind-ness. An estimated 97% of insulin-taking and80% of non-insulin-taking persons who havehad diabetes for $15 years have retinopa-thy.59 Diabetic retinopathy is a highly specificmicrovascular complication of diabetes; its


109

prevalence is directly related to the durationof diabetes.60 The 2000 ADA position state-ment61 indicates that vision-threatening reti-nopathy does not occur in patients with type1 diabetes during the first 3- to 5-year dura-tion, and that nearly all patients with type 1diabetes have retinopathy after a 20-year du-ration of diabetes. Diabetic retinopathy is es-timated to be the most frequent cause of newcases of blindness in the United States amongadults aged 20–74 years.

Significantly higher rates of vision prob-lems are reported in Canadians over 65 yearswith DM than those of the same age withoutDM.23 Cataracts occur in 21.9% of DM sub-jects and in 14% of the general population.Glaucoma affects 7% of the DM and 5% ofthe general population. Vision problems thatcannot be corrected and total vision loss oc-cur in 9% of patients with DM vs 5% of per-sons without the disease. The MbR is 156%for cataracts, 140% for glaucoma, and 180%for vision problems or vision loss, when com-paring rates in diabetic and non-diabetic pop-ulations.

The 1989 US59 prevalence of self-reportedcataracts in all adults ($18 years) was 22.42%in those with DM vs 3.30% in controls with-out DM. Glaucoma was 6.98% vs 0.85%, andtrouble seeing was 13.72% vs 3.58%. TheMbR was 679% for cataracts, 821% for glau-coma, and 383% for trouble seeing. The ratesfor age $65 years were 38.37% for cataractsin individuals with DM vs 16.56% in thosewithout DM, 11.18% vs 3.79% for glaucoma,and 15.92% vs 6.88% for trouble seeing. TheMbRs for this age category were 232% for cat-aracts, 295% for glaucoma and 231% for trou-ble seeing.

Incidence rates of blindness in the diabeticpopulation compared to the non-diabeticpopulation were collected in the district ofWuttemberg-Hohenzollern, Germany, be-tween 1990 and 1993.62 Standardized to theGerman population, the incidence rates (per100,000 person-years) were 13.5 in the gen-eral population, 60.0 in the diabetic popula-tion, and 11.6 in the non-diabetic population.Comparing incidence rates in the diabetic

population with the non-diabetic population,the MbR was 517%.

RISK OF HOSPITALIZATION

DM is usually diagnosed and treated on anoutpatient basis, and hence, tends to be poor-ly represented in hospital discharge records.23

Furthermore, people with DM are usually ad-mitted to the hospital because of related com-plications.63

Data from the Centers for Disease Controland Prevention2 indicate that in 1996 thenumber of hospital discharges with diabetesas the first-listed diagnosis (all ages) was503,000, with an average length of stay of 6.3days. The age-adjusted rate of emergency de-partment visits was 14.1 per 100 persons withDM. Based on the 1989 NHIS,63 23.8% of alladults with DM vs 7.8% of those without DMreported being hospitalized at least once inthe previous year. We calculated a hospital-ization ratio (HR) of 305% (ie, for all adults,persons with DM are 3 times more likely toreport being hospitalized in the previous yearthan persons without DM). The HR was 411%for age 18–44 years, 239% for age 45–64 years,168% for age 65–74 years, and 184% for age$75 years.

In Canada,23 approximately 34,000 hospitaldischarges were attributable to DM in 1995.This figure represents over 400,000 hospitaldays for DM. The age-standardized rateswere 112 hospital discharges per 100,000population, and 1368 hospital days per100,000 population. In 1996–1997, self-report-ed overnight stay in hospital, nursing homeor convalescent home in the previous year inCanada was: 16.2% in persons with DM aged35–64 years vs 7.5% in persons without DM,and 23.9% vs 13.9% in persons over age 65years. The HR was 216% and 172% in theseage groups, respectively.

In a retrospective cohort study in Tayside,Scotland,64 during 1995, in 366,849 peopleregistered with general practitioners, thelength of stay was highest for patients withDM. This group had approximately doublethe risk of admission, due to a higher risk of


110

neurological, cardiovascular, renal and oph-thalmic hospitalization compared to peoplewithout DM. Approximately 24% of all studysubjects with DM had at least 1 hospital ad-mission, compared to 12% of the non-diabeticpopulation. We calculated an HR of 200%.

In a register-based linkage study65 compar-ing hospital use among diabetic and non-di-abetic populations in Finland, 50.7% of pa-tients with DM had at least 1 hospital stay forany cause in the year, compared to 12.4% ofthe non-diabetic population. The HR was409%. Table 3 summarizes hospitalization ra-tios calculated from these studies.

RISK OF DISABILITY

Disability affects large numbers of personswith DM in the United States with estimatesranging from 20% to 50% of the diabetic pop-ulation.66 The rates of disability are substan-tially higher among persons with DM thanamong subjects without this disease. The con-sequences of disability among persons withDM include increased use of health care ser-vices, unemployment, work absenteeism, anddecreased quality of life.2

Chronic diseases, including diabetes, arestrongly linked with reported activity limi-tation in the NHIS.67 In 1983–1985, DM wasthe 11th most common condition cited as themain cause of activity limitation in the UnitedStates, accounting for 2.7% of all reportedcases. When considered as the contributingcause of activity limitation, it was the 6thmost frequent condition cited, explaining6.5% of all cases.

In 1989,36 the age-standardized percent ofpersons aged $18 years reporting any re-stricted activity days in the previous 2 weekswas 22.4% in individuals with NIDDM,21.3% in individuals with IDDM vs 10.3% insubjects without DM. The disability ratio, cal-culated as the ratio in subjects with NIDDMto those without DM, was 217%. Absenteeismassociated with diabetes was also assessed inthe 1989 NHIS. Measured as the number ofwork-loss days in the previous 2 weeks, itwas higher among both NIDDM and IDDM

respondents, compared with the non-diabeticpopulation: 11.8 and 12.7 vs 6.9 at age 18–44years. The DR was 184% when comparingNIDDM respondents with the non-diabeticpopulation.

In 1994, approximately 42% of individualsaged 18–69 years with diabetes reported be-ing unable to work or being limited in thekind or amount of work activity they coulddo, and approximately 33% were sufficientlydisabled to be completely out of the work-force.68

Work disability was reported by 25.6% ofindividuals with diabetes, compared with7.8% of those without diabetes.69 Work-lossdays per year were 5.9 in men and 5.4 inwomen with DM vs 3.4 in men and 3.8 inwomen without diabetes, according to theNational Medical Expenditures Survey-2.70

From these results, the DR for disability inindividuals with diabetes vs those withoutdiabetes was 328%.

In Canada,23 among people of working age(35–64 years), 23% of those with DM reported1 or more disability days (in bed or with re-stricted activities) in a 2-week period, com-pared to 11% of those without DM. The DRwas 209%. Disability for an entire 2-week pe-riod was reported by 9.3% of people with DMcompared to 2.9% of people without DM. TheDR was 320%. Of those who reported at least1 disability day in the previous 2-week peri-od, those with DM had an average of 9 dis-ability days, compared to an average of 6days among those without DM. A prospec-tive, population-based cohort study71

screened individuals of working age (18–64years) in Manitoba found 25,554 personswithout diabetes and 608 with diabetes. Di-abetic individuals with complications weretwice as likely not to be in the labor forcecompared to non-diabetic individuals (OR2.07, 95% CI 1.49–2.87).

Disability in persons with DM is influencedby a number of demographic and diabetes-related factors.66 Impairments reported bypersons with DM increase with age. At age18–44 years, 45% report activity limitation; atage 45–64 years, 55% report activity limita-


111

tion; and at age $65 years, 60% report activ-ity limitation. Disability is more common inminority groups. Persons with NIDDM, par-ticularly those using insulin, appear to bemore affected than those with IDDM. Thepresence of late complications of DM seemsto be a major determinant of disability. Table4 summarizes the DR calculated from thesestudies.

CONCLUSIONS

Diabetics have higher rates of morbidevents, hospitalization and disability daysthan non-diabetic persons.

Morbidity ratios (MbR) and hospitalizationratios (HR) should be carefully reviewed ac-cording to age when underwriting privatehealth insurance.

Disability ratios (DR), MbRs and HRsshould also be carefully reviewed when un-derwriting disability insurance for diabetics.

ACKNOWLEDGEMENTS

The authors acknowledge the editorialwork of Mr. Ovid M. Da Silva, Redacteur/Reviseur of the Research Support Office, Cen-tre de recherche, Centre hospitalier del’Universite de Montreal 2 Hotel-Dieu.

REFERENCES

1. Leiter LA, Barr A, Belanger A, et al. DiabetesScreening in Canada (DIASCAN) Study. Preva-lence of undiagnosed diabetes and glucose intol-erance in family physician offices. Diabetes Care.2001;24:1038–1043.

2. CDC’s Diabetes Program Statistics. Diabetes Sur-veillance 1999. The public health burden of diabetesmellitus in the United States. CDC, Atlanta, 1999:chap1.

3. Carter JS, Pugh JA, Montosa A. Non-insulin-de-pendent diabetes mellitus in minorities in theUnited States. Ann Intern Med. 1996;125:221–232.

4. Expert Committee on the Diagnosis and Classifi-cation of Diabetes Mellitus. Report of the ExpertCommittee on the Diagnosis and Classification ofDiabetes Mellitus. Diabetes Care. 2000;23(suppl.1):S4-S19.

5. Expert Committee on the Diagnosis and Classifi-cation of Diabetes Mellitus. Report of the Expert

Committee on the Diagnosis and Classification ofDiabetes Mellitus. Diabetes Care. 1997;20:1183–1197.

6. World Health Organization. Diabetes Mellitus: Re-port of a WHO Study Group. Geneva: 1985, Tech.Rep. Ser. no. 727.

7. World Health Organization. Definition, diagnosisand classification of diabetes mellitus and its com-plications. Report of a WHO Consultation. Part 1.Diagnosis and classification of diabetes mellitus.Geneva:1999.

8. Gabir M, Hanson R, Dabelea D, et al. The 1997American Diabetes Association and 1999 WorldHealth Organization Criteria for Hyperglycemia inthe Diagnosis and Prediction of Diabetes. DiabetesCare. 2000;23: 1108–1112.

9. Nathan D, Meigs J, Singer D. The epidemiology ofcardiovascular disease in type 2 diabetes mellitus:how sweet it is. . .or is it. Lancet. 1997;350(suppl I):4–9.

10. Nathan D. Prevention of long-term complicationsof non-insulin-dependent diabetes mellitus. ClinInvest Med. 1995;18(4):332–339.

11. American Diabetes Association. Direct and indi-rect costs of diabetes in the United States in 1992.Alexandria, VA: American Diabetes Association,1993.

12. Meldrum N, Mecklenburg R, Lovell K. Outcomesand costs in diabetes care: trends over three years(abstract). Diabetes Care. 2001;24:1038–1043.

13. Rubin R, Altman W, Mendelson D. Health care ex-penditures for people with diabetes mellitus, 1992.J Clin Endocrinol Metab. 1994;78(4):809A–809F.

14. Gagliardino J, Williams R, Clark C. Using hospi-talization rates to track the economic costs andbenefits of improved diabetes care in the Ameri-cas. A proposal for health policy makers. DiabetesCare. 2000;23:1844–1845.

15. Singer R. Comparative morbidity—what are theprospects? J Ins Med. 1988;20(3):47–50.

16. Kuller L, Velentgas P, Barzilay L. Diabetes mellitus:subclinical cardiovascular disease and risk of in-cident cardiovascular disease and all-cause mor-tality. Atheroscler Thromb Vasc Biol. 2000;20:823–829.

17. Wingard D, Barrett-Connor E. Heart disease anddiabetes. In: Harris M, ed. Diabetes in America. 2nded. Bethesda, MD: National Institutes of Health;1995:429–448.

18. Kannel W, McGee D. Diabetes and cardiovasculardisease: the Framingham study. JAMA. 1997;241:2035–2038.

19. Nichols G, Hillier T, Erbey J, Brown J. Congestiveheart failure in type 2 diabetes: prevalence, inci-dence, and risk factors. Diabetes Care. 2001;24:1614–1619.

20. Iribarren C, Karter A, Go A, et al. Glycemic control


112

and heart failure among adult patients with dia-betes. Circulation. 2001;103:2668–2673.

21. The UK Prospective Diabetes Study Group. As-sociation of glycemia with macrovascular and mi-crovascular complications of type 2 diabetes(UKPDS 35): prospective observational study. BMJ.2000;321:405–412.

22. Rodriguez B, Lau N, Burchfiel C, et al. Glucoseintolerance and 23-year risk of coronary heart dis-ease and total mortality: the Honolulu Heart Pro-gram. Diabetes Care. 1999;22:1262–1265.

23. Diabetes Division, Bureau of Cardio-RespiratoryDiseases and Diabetes, Laboratory for DiseaseControl, Health Protection Branch, Health Canada.Diabetes in Canada: national statistics and oppor-tunities for improved surveillance, 1999.

24. Palumbo P, Melton L. Peripheral vascular diseaseand diabetes. In: Harris M, ed. Diabetes in America.2nd ed. Bethesda, MD: National Institutes ofHealth; 1995:401–408.

25. Jude E, Oyibo SO, Chalmers N, et al. Peripheralarterial disease in diabetic and nondiabetic pa-tients: a comparison of severity and outcome. Di-abetes Care. 2001;24:1433–1437.

26. Melton L, Macken K, Palumbo P, et al. Incidenceand prevalence of clinical peripheral vascular dis-ease in a population-based cohort of diabetic pa-tients. Diabetes Care. 1980;13:650–654.

27. Kannel W, Skinner J Jr, Schwartz M, et al. Inter-mittent claudication incidence in the Framinghamstudy. Circulation. 1970;41:875–883.

28. Payne G. Diabetes-related lower limb amputationin Australia. MJA. 2000;173:352–354.

29. Faglia E, Favales F, Morabito A. New ulceration,new major amputation, and survival rates in dia-betic subjects hospitalized for foot ulceration from1990 to 1993. Diabetes Care. 2001;24:78–83.

30. Nelson R, Gohdes D, Everhart J, et al. Lower limbamputation in NIDDM:12-year follow-up study inPima Indians. Diabetes Care. 1988;11:8–16.

31. Ebscov L. Relative mortality in lower limb ampu-tees with diabetes mellitus. Prosthet Orthot Int.1996;20:147–152.

32. Ebscov L. Diabetic amputation and long-term sur-vival. Int J Rehab Res. 1998;21:403–408.

33. Van Houtum W, Lavery L. Outcome associatedwith diabetes with diabetes-related amputations inthe Netherlands and the state of California, USA.J Intern Med. 1996;240(4):227–231.

34. Solomon C, VanRij A, Barnett R, Packer S, Lewis-Barned N. Amputation in the surgical budget. NZMed J. 1994;107:78–80.

35. Lehto S, Ronnemaa T, Pyorala K, Laakso M. Riskfactors predicting lower extremity amputation inpatients with NDDIM. Diabetes Care. 1996;19(6):607–612.

36. National Center for Health Statistics. Current es-

timates from the national Health Interview Survey,1989. Vital and Health Statistics Series 10, no.176,1990.

37. Wrobel J, Mayfield J, Reiber G. Geographic varia-tion of lower extremity major amputation in per-sons with and without diabetes in the Medicarepopulation. Diabetes Care. 2001;24:860–864.

38. Siitonen O, Niskanen L, Laakso M, Siitonen J,Pyorala K. Lower-extremity amputation in diabeticand nondiabetic patients. A population-basedstudy in eastern Finland. Diabetes Care. 1993;16(1):16–20.

39. Trautner C, Haastert B, Giani G, Berger M. Inci-dence of lower limb amputation and diabetes. Di-abetes Care. 1996;19(9):1006–1009.

40. Niskanen L, Rauramaa R, Miettinen H, Haffner S,Mercuri M, Uusitupa M. Carotid artery intima-media thickness in elderly patients with NIDDMand in nondiabetic subjects. Stroke. 1996;27:1986–1992.

41. Pujia A, Gnasso A, Irace C, Colonna A, MattioliPL. Common carotid arterial wall thickness inNIDDM subjects. Diabetes Care. 1994;17(11):1330–1336.

42. Kuller L. Stroke and Diabetes. In: Harris M, ed.Diabetes in America. 2nd ed. Bethesda, MD: Nation-al Institutes of Health; 1995:449–456.

43. Macaulay A, Paradis G, Potvin L, et al. The Kahn-awake Schools Diabetes Prevention Project: Inter-vention, evaluation and baseline results of a dia-betes primary prevention program with a nativecommunity in Canada. Prev Med. 1997;26:779–790.

44. Manson J, Colditz G, Stampfer M, et al. A pro-spective study of maturity-onset diabetes mellitusand risk of coronary heart disease and stroke inwomen. Arch Intern Med. 1991;151:1141–1147.

45. Folsom A, Rasmussen M, Chambless L, et al. Pro-spective association of fasting insulin, body fat dis-tribution, and diabetes with risk of ischemicstroke. Diabetes Care. 1999;22:1077–1083.

46. Macaulay A, Montour L, Adelson N. Prevalence ofdiabetes and atherosclerotic complications amongMohawk Indians of Kahnawake. Can Med Assoc J.1988;139:22–24.

47. Estman R. Neuropathy in diabetes. In: Harris M,ed. Diabetes in America. 2nd ed. Bethesda, MD: Na-tional Institutes of Health; 1995: 339–348.

48. Vinik A, Park T, Stansberry K, Pittenger G. Dia-betic neuropathies. Diabetologia. 2000;43:957–973.

49. Dyck P, Kratz K, Lehman K, et al. The RochesterDiabetic Neuropathy Study. Neurology. 1991;41:799–807.

50. Franklin G, Kahn L, Baxter J, Marshall J, HammanR. Sensory neuropathy in non-insulin-dependentdiabetes mellitus: the San Luis Valley DiabetesStudy. Am J Epidemiol. 1990;131:633–643.

51. National Center for Health Statistics. Plan and op-


113

eration of the Second National Health and Nutri-tion Examination Survey. Vital and Health Statis-tics, Series 1, no. 15. Washington, DC: US Govern-ment. Printing Office, 1981. DHHS publication no.PHS 81–1317.

52. The Diabetes Control and Complications Trial Re-search Group. The effect of intensive diabetestreatment on the development and progression oflong-term complication in insulin-dependent dia-betes mellitus: The Diabetes Control and Compli-cations Trial. N Engl J Med. 1993;329:977–986.

53. U.S. Renal Data System. 1998 Annual Data Report.Bethesda, MD: National Institute of Diabetes andDigestive and Kidney Diseases, April 1998. NIHPublication no. 98–3176.

54. U.S. Renal Data System. 2001 Annual Data Report:Atlas of end-stage renal disease in the UnitedStates, National Institute of Diabetes and Digestiveand Kidney Diseases, Bethesda, MD: 2001.

55. Canadian Institute for Health Information. Cana-dian Organ Replacement Register Report 1998. Di-alysis and renal transplantation. Ottawa, 1998;4–6.

56. Ritz E, Rychlik I, Locatelli F, Halimi S. End-stagerenal failure in type 2 diabetes: a medical catastro-phe of worldwide dimension. Am J Kidney Dis.1999;34(5):795–808.

57. Wetterhall S, Olson D, DeStefano F, Stevenson JM,Ford ES, German RR, et al, Trends in diabetes anddiabetic complications, 1980–1987. Diabetes Care.1992;15(8):960–967.

58. Brancati F, Whelton P, Randall B, Neaton J, StamlerJ, Klag MJ. Risk of end-stage renal disease in dia-betes mellitus: a prospective cohort study of menscreened for MRFIT. Multiple Risk Factor Interven-tion Trial. JAMA. 1997;278(23):2069–2074.

59. Klein R, Klein B. Vision disorders in diabetes. In:Harris M, ed. Diabetes in America. 2nd ed. NationalInstitutes of Health, Bethesda, MD: 1995;293–336.

60. Malone J, Morrison A, Pavan P, Cuthbertson D. Di-abetes Control and Complications Trial, Prevalenceand significance of retinopathy in subjects with di-abetes of less than 5 year’s duration screened for

the Diabetes Control and Complication Trial. Di-abetes Care. 2001;24(3):522–526.

61. American Diabetes Association. Clinical PracticeRecommendations 2000. Diabetes Care. 2000;23(suppl I):S730-S776.

62. Trautner C, Icks A, Haastert B, et al. Incidence ofblindness in relation to diabetes. A population-based study. Diabetes Care. 1997;20(7):1147–1153.

63. Aubert R, Geiss L, Ballard D, et al. Diabetes-relat-ed hospitalization and hospital utilization. In: Har-ris M, ed. Diabetes in America. 2nd ed. Bethesda,MD: National Institutes of Health; 1995:553–569.

64. Donnan P, Leese G, Morris A. Hospitalization forpeople with type 1 and type 2 diabetes comparedwith the nondiabetic population of Tayside, Scot-land: a retrospective cohort study of resource use.Diabetes Care. 2000;23:1774–1779.

65. Aro S, Kangas T, Reunanen A, Salinto M, KoivistoV. Hospital use among diabetic patients and thegeneral population. Diabetes Care. 1994;17(11)1320–1329.

66. Songer T. Disability in diabetes. In: Harris M, ed.Diabetes in America. 2nd ed. Bethesda, MD: Nation-al Institutes of Health; 1995:259–282.

67. Institute of Medicine. Disability in America. To-ward a National Agenda for Prevention. In: PopeA, Taylor A, eds. Washington, DC: National Acad-emy Press; 1991.

68. Centers for Disease Control and Prevention. Divi-sion of Diabetes Translation: Diabetes Surveil-lance, 1997, Atlanta, GA: Department of Healthand Human Services; 1997.

69. Mayfield J, Deb P, Whitecotton L. Work disabilityand diabetes. Diabetes Care. 1999;22:1105–1109.

70. Dwarfs W, Berlin M. Questionnaires and data col-lection method for the household survey and thesurvey of American Indians and Alaska Natives.Method 2: National Medical Expenditures Survey2. Rockville, MD: National Center for Health Ser-vices Research. Department of Health and HumanServices; 1989.

71. Kraut A, Walld R, Tate R, Mustard C. Impact ofdiabetes on employment and income in Manitoba,Canada. Diabetes Care. 2001;24(1):64–68.

Documents

Diabetes Mellitus Related Morbidity