Journal of Plastic, Reconstructive & Aesthetic Surgery (2014) 67, 629e633

Risk factors for prolonged treatment andhospital readmission in 280 cases ofnegative-pressure wound therapy

Georg Osterhoff*, Pawel Zwolak, Carmen Kruger,Verena Wilzeck, Hans-Peter Simmen, Gerrolt N. Jukema

Division of Trauma Surgery, University of Zurich, Raemistrasse 100,8091 Zurich, Switzerland

Received 26 April 2013; accepted 12 January 2014

KEYWORDSNegative-pressurewound therapy;Vacuum-assistedclosure;Open fractures;Risk factors;Wound healing

* Corresponding author. DepartmentSwitzerland. Tel.: þ41 44 255 1111; f

E-mail addresses: georg.osterhoff@

1748-6815/$-seefrontmatterª2014Brihttp://dx.doi.org/10.1016/j.bjps.2014.0

Summary Background: There is evidence of certain beneficial effects and increasing under-standing of the mechanisms of action of negative-pressure wound therapy (NPWT). However, itis known that prolonged duration of NPWT is associated with increased bacterial growth andefforts should be made to decrease the duration of NPWT. It was the aim of this study to eval-uate potential risk factors for the duration, from first application of NPWT to secondary woundclosure and to identify factors that increase the rate of hospital readmission.Methods: In a retrospective cohort study, 261 patients (46 � 19 years, 70 female) who under-went 280 treatments with NPWT were analysed. Patient-specific and demographic character-istics and the presence of several risk factors were documented. The duration of treatmentfrom first application of NPWT to secondary wound closure, the number of interventions,the duration of hospital stay and the incidence of readmissions due to complications of thewound treated by NPWT were recorded and a risk factor analysis was performed.Results: The median number of NPWT procedures was 2.0 � 2.0, the duration of NPWT was6.0 � 14.7 days and the length of hospital stay was 16.0 � 27.9 days. Presence of an open frac-ture (p Z .002) and increased age (p Z .004) were identified as independent risk factors for aprolonged duration of NPWT. Patients who smoked (p Z .001) or patients with alcohol/drugabuse (p Z .015) were more likely to return to hospital (smoking: 18 out of 58 cases;alcohol/drug abuse: 7 out of 19 cases). No such association was seen for diabetes(p Z .702), peripheral vascular disease (PVD) (p Z .052), immunosuppressive medication(p Z .187), immunodeficiency (p Z .404), trauma (p Z .358), infection (p Z .298) and openfracture (p Z .061).

of Surgery, Division of Trauma Surgery, University Hospital Zurich, Raemistrasse 100, 8091 Zurich,ax: þ41 44 255 4406.usz.ch, g.osterhoff@gmx.de (G. Osterhoff).

tishAssociationofPlastic,ReconstructiveandAestheticSurgeons.PublishedbyElsevierLtd.All rightsreserved.1.010

630 G. Osterhoff et al.

Conclusions: Patient age and presence of an open fracture are independent predictors of aprolonged duration from first application of NPWT to secondary wound closure. These resultsshould be taken into account for the calculation of average costs and anticipated hospital stayassociated with this therapy.ª 2014 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published byElsevier Ltd. All rights reserved.

Negative-pressure wound therapy (NPWT) combines theconcept of protection against intruding microorganisms bya sterile wound dressing with the need for continuousdrainage of exudate and haematoma in larger wounds. Thisis achieved by embedding a polyvinyl alcohol (PVA) or apolyurethane ether (PUE) sponge into the wound cavity,sealing the wound by semi-occlusive drapes and applyingnegative pressure through drape and sponge.1 Besidessterile coverage and wound drainage, NPWT has beenshown to decrease oedema, to promote tissue granulationand to result in an improved blood supply of the perifocaltissue.2e4 In addition, it renders daily changes of dressingsunnecessary, as the NPWT system has to be changed onlyevery 3e6 days and, during this interval, it allows forcontinuous wound contraction.2 When compared to astandard wet gauze therapy, NPWT achieves better woundconditions and an accelerated reduction of wound surfacearea.5 Thus, it has become commonly used in the treatmentof larger soft-tissue wounds or soft-tissue infections andopen fractures or osteomyelitis,1,6 as well as in the man-agement of surgical wounds associated with a high risk ofdeterioration.7

Even though there is strong evidence of certain benefi-cial effects and increasing understanding of the mecha-nisms of action of NPWT,2e4,8 the clinical benefit of NPWTover conventional treatment is still under debate.9,10 Someauthors report an increase of bacterial growth in woundstreated by NPWT with time.5,11 This, next to health eco-nomic considerations, is one reason why efforts should bemade to decrease the duration of NPWT.

Especially in view of the broad range of indications forNPWT with its very inhomogeneous patient population, itwas the aim of this study to evaluate potential risk factorsfor differences in the duration from first application ofNPWT to secondary wound closure and to identify factorsthat increase the rate of hospital readmission.

Patients and methods

A monocentric retrospective cohort study was conducted ata university level 1 trauma centre through review of charts.

This study was carried out in accordance with the localinstitutional ethics committee’s terms of reference (Kant-onale Ethikkommission Zurich, Switzerland). Patientsincluded were treated by NPWT during the year 2011. Pa-tients with a follow-up interval after end of hospital stayshorter than 1-year were excluded. In total, 261 patientswho underwent 280 treatments with appropriate follow-upwere analysed for this study. At the time of first applicationof NPWT, patient-specific and demographic characteristicsand body region were documented. Special focus was puton the presence of diabetes mellitus, peripheral vascular

disease (PVD), smoking, alcohol/drug abuse and an immu-nodeficiency (human immunodeficiency virus (HIV), tuber-colosis (TBC) and malignancy) or immunosuppressivemedication.

The standard treatment protocol included multiple sur-gical debridements, irrigation and temporary closure withNPWT using a PVA or a PUE foam (Kinetic Concepts Inc.(KCI), San Antonio, TX, USA). All NPWT treatments wereconducted in an inpatient care setting and all changes ofthe dressing/foam were done in the operation room; noneof the patients was discharged with NPWT in situ. We took asecond look within 24e48 h; further changes of the dres-sing/foam e if necessary e were scheduled within intervalsof 3e5 days, depending on the amount of necrosis andcontamination seen in the intervention before. Microbio-logical samples were harvested during the first debride-ment, periodically during the sequential interventions andbefore definitive closure was planned. Infection wasdefined to be present if multiple sequential samplesshowed growth of a microorganism or in case of wounddrainage or presence of an abscess with at least one posi-tive sample. All patients with proven or suspected infectionas well as patients with a high risk for infection (e.g., pa-tients with open fractures) were initially treated with anintravenous broad-spectrum antibiotic (amoxicillin/clav-ulanic acid). Antibiotic therapy was then changed based onthe antibacterial sensitivity tests. Patients withoutincreased risk for infection (e.g., patients after fasciotomydue to compartment syndrome) received an antibioticsingle-shot prophylaxis (cefazolin) 30 min before surgery.

Secondary wound closure was performed when previ-ously harvested microbiological samples were negative andwound inspection revealed abundant granulation tissue. Incase of persisting soft-tissue defects that were too large fordelayed primary closure by sutures, wound closure wasattained by a flap or skin graft.

During the course of therapy, the duration from firstapplication of NPWT to secondary wound closure (durationof NPWT), the number of interventions and the length ofhospital stay were documented for every case.

Frequent routine follow-up included clinical assessmentand blood tests for inflammatory parameters in patients atrisk. The incidence of readmissions due to complications ofthe wound treated by NPWT and the time until return towork were recorded.

Statistical analysis

Statistical analysis was performed by an institutional stat-istician. All data were recorded in an Excel database(Microsoft Corp., Washington, D.C., USA) and exported toSPSS 21.0 (SPSS Inc., Chicago, IL, USA) for statistical

Table 2 Patients’ baseline characteristics (with andwithout infection).

Infection p Total

Yes No

N (cases) 104 (110) 157 (170) 261 (280)Age [y; �SD] 49 � 18 44 � 19 .040a 46 � 19Gender [f:m] 32:72 38:119 .256b 70:191Comorbidity

Smoking 29 (28%) 30 (19%) .130b 59 (23%)PVD 13 (13%) 8 (8%) .038b 21 (8%)Alcohol/drug abuse 14 (13%) 5 (3%) .003b 19 (7%)Diabetes 6 (6%) 8 (5%) .788b 14 (5%)Imm.-deficiency 9 (9%) 3 (2%) .015b 12 (5%)

b

Risk factors of NPWT 631

analysis. Unless otherwise denoted, data were summarisedas median � standard deviation.

Primary outcome was the duration of NPWT. Secondaryoutcome was the readmission rate. Using ManneWhitneytests, a step-wise explorative risk factor analysis was per-formed for the variables age, diabetes mellitus, PVD,smoking, alcohol/drug abuse, immunodeficiency, immuno-suppressive medication, trauma, infection and open frac-ture. For those parameters identified as potential riskfactors for NPWT days a regression analysis was performed.For regression analysis, the variable ‘duration of NPWT’ waslogarithmised to approximate normal distribution.

For the secondary outcome parameter ‘readmission’,crosstabs were associated using Fisher’s Exact tests. Thelevel of significance was defined as p < .05.

Imm.-sup. med. 7 (7%) 3 (2%) .094 10 (4%)

Imm.-deficiency: Immunodeficiency. Imm.-sup. med.: Immu-nosuppressive medication.a ManneWhitney-U-test.b Fisher’s exact test.

Results

During the evaluated interval of 1-year, 261 patients (meanage 46 � 19 years, 70 female) underwent NPWT in 280cases. Nineteen patients had wounds requiring NPWT in twodifferent regions of the body, and therefore, were analysedas two cases each. The lower extremity was the mostaffected localisation and trauma was the most commonreason leading to the need for NPWT, followed by infectionand other causatives (Table 1). Out of the 191 traumacases, 65 (34%) cases had open fractures where NPWT wasused to prepare the wound until definitive fracture fixation.In addition to the 110 cases with primary infection, 65 ofthe 191 (34%) trauma cases developed secondary infectionbefore or during NPWT treatment (as defined in the Mate-rials and Methods section).

The most common comorbidities were smoking (23%),PVD (8%) and alcohol/drug abuse (7%). Patients with PVD,alcohol/drug abuse, an immunodeficiency and older pa-tients were more likely to present with an infection (Table2).

The median number of NPWT procedures was 2.0 � 2.0.The median duration from first application of NPWT tosecondary wound closure was 6.0 � 14.7 days and themedian length of the whole hospitalisation was 16.0 � 27.9days. Patients returned to work after median 31.5 � 81.7days.

Age and open fracture were identified as potential in-dependent risk factors for the duration of NPWT andconfirmed by regression analysis (R2 Z .064, adjustedR2 Z .057, p < .001). NPWT treatment was maintainedlonger in patients with an open fracture (coefficientB Z .161, t Z 3.08, p Z .002; Figure 1) and increased age(coefficient B Z .003, t Z 2.91, p Z .004; Figure 2). Theaverage duration of NPWT in the presence of an openfracture in patients of the same age was 1.45 times longer.

Table 1 Affected localizations treated by NPWT.

Localization

Abdomen Spine/Trunk

Cases 9 (3%) 58 (21%)Duration of NPWT [d] 3.0 � 9.1 5.0 � 16.5

Ten additional years of patient age prolonged the averageduration of NPWT by 8%. In total, patients with an openfracture were treated with NPWT for a median of 9.0 � 15.1days requiring 3.0 � 2.2 surgical debridements, and thosewithout an open fracture for 5.0 � 14.5 days and 2.0 � 1.9debridements.

Forty patients (15%) had to return to the hospital duringthe follow-up due to recurrent infection or disorders ofwound healing. Patients who smoked (p Z .001) or patientssuffering from alcohol/drug abuse (p Z .015) were morelikely to return to the hospital (smoking: 18 out of 58 cases,31%; alcohol/drug abuse: 7 out of 19 cases, 37%). No asso-ciation with the need for readmission was seen for diabetes(pZ .702), PVD (pZ .052), immunosuppressive medication(p Z .187), immunodeficiency (p Z .404), trauma(p Z .358), infection (p Z .298) and open fracture(p Z .061).

Discussion

This is the first study, known to the authors, evaluating riskfactors for the duration of NPWT and for hospital read-mission after this treatment. It was shown that the pres-ence of an open fracture and age were independent riskfactors for the duration from first application of NPWT tosecondary wound closure. No such association was observedfor the variables diabetes, PVD, smoking, alcohol/drugabuse, immunosuppressive medication, immunodeficiency,trauma and infection.

Total

Upper Extremity Lower Extremity

56 (20%) 157 (56%) 280 (100%)4.0 � 4.6 7.0 � 16.4 6.0 � 14.7

Figure 1 Duration of NPWT in patients with and withoutopen fracture. The dotted line represents the median of allpatients.

632 G. Osterhoff et al.

The median duration of NPWT in our population ofmainly trauma patients was 6 days, which is within therange of results from previous studies.5,6,12,13 An associa-tion between age and duration of NPWT is intuitively ex-pected. Wound healing is impaired in old patients. Studiesdescribing longer intervals usually treated older patientswith chronic soft-tissue infects.14,15

In contrast to our data, Stannard et al. reported a meanduration of NPWT of only 3 days for open fractures but donot provide information on patients without open frac-tures.13 One possible explanation of the difference be-tween our results and the data reported by Stannard et al.may be differences in treatment algorithms. In our study eand especially in patients who received internal fracture

Figure 2 Duration of NPWT over patient age splitted for openfracture. The effect of 40 years of age is comparable to theeffect of having an open fracture.

stabilisation e secondary wound closure was only per-formed when previously harvested microbiological sampleswere negative. This might also be the most important fac-tor why open fractures were a risk factor for a prolongedtreatment with NPWT.

Patients who smoked or patients with alcohol/drugabuse had a higher hospital-readmission rate. All NPWTtreatments in this study were conducted as inpatient care.Different treatment algorithms are possible and it is clearthat for patients treated in an outpatient setting, the re-sults of the secondary outcome analysis (hospital read-mission) are less relevant. However, especially for patientstreated in an outpatient setting, it might be of interest toknow which patients are rather prone to failure of NPWT.

The chosen definition of an infection to be present ifmultiple sequential samples showed growth of a microor-ganism might be low-threshold and is, for sure, one reasonfor the rather large proportion (24%) of the traumaticwounds that were labelled ‘infected’. The observed time tosecondary wound closure in our study population did notexceed those intervals described in most of the clinicalstudies on NPWT.5,9,12,14,15 In addition, the rate of read-missions due to wound complications was 15%. This is ratherin the low range of wound complication rates with need forsurgery (from 10% to 58%) described in the literature.6,14e16

Although our study population is one of the largestsamples in comparison to other studies on NPWT, the in-homogeneity of our population with treatments in differentregions of the body is a limitation. This study took place in acentre for orthopaedic trauma and most of the patientsrequired NPWT due to trauma or acute infections. Chronicwounds as diabetic foot ulcers might be underrepresented.Further limitations of our study are the retrospective studydesign and the statistical procedure of an explorative riskfactor analysis. However, our results point towards factorsthat should be further investigated in prospective trials oneach localisation with larger subgroup samples.

It is important to note that even though statisticallysignificant both identified risk factors were rather weak(R2 Z .064) in their ability to predict the duration of NPWTfor an individual patient. There must be some further majorfactors contributing to the total duration of NPWT thatwere not investigated in this study. In the clinical routine,the working pace of the microbiological laboratory, treat-ment algorithms (see above) and availability of operationroom slots will strongly influence the total duration ofNPWT and hospital stay. Therefore, our results are espe-cially interesting from a health economic perspective.There is evidence for the economic benefit of NPWT whencompared with both traditional and advanced wounddressings.17 However, it is known that the main portion ofthe costs generated by NPWT is not due to high materialcosts but due to long-term treatments.18 During the intro-duction of diagnosis-related groups (DRGs) in many coun-tries, these data should be taken into account for thecalculation of average costs and anticipated hospital stay inthe DRG catalogues.

In this context, early discharge from hospital with anongoing NPWT was suggested.19,20 Further studies arerequired to investigate if this cost-effective alternativemight be an option to improve the quality of life in patientswith open fractures, as well.

Risk factors of NPWT 633

Conclusion

Patient age and presence of an open fracture are inde-pendent predictors of a prolonged duration from firstapplication of NPWT to secondary wound closure. No suchassociation was observed for the variables diabetes, PVD,smoking, alcohol/drug abuse, immunosuppressive medica-tion, immunodeficiency, trauma and infection. These re-sults should be taken into account for the calculation ofaverage costs and anticipated hospital stay associated withthis therapy.

Conflict of interest statement

There are no conflicts of interest. There were no externalsources of funding.

References

1. Streubel PN, Stinner DJ, Obremskey WT. Use of negative-pressure wound therapy in orthopaedic trauma. J Am AcadOrthop Surg 2012;20:564e74.

2. Orgill DP, Manders EK, Sumpio BE, et al. The mechanisms ofaction of vacuum assisted closure: more to learn. Surgery2009;146:40e51.

3. Timmers MS, Le Cessie S, Banwell P, Jukema GN. The effects ofvarying degrees of pressure delivered by negative-pressurewound therapy on skin perfusion. Ann Plast Surg 2005;55:665e71.

4. Saxena V, Hwang CW, Huang S, et al. Vacuum-assisted closure:microdeformations of wounds and cell proliferation. PlastReconstr Surg 2004;114:1086e96 [discussion 97e8].

5. Moues CM, van den Bemd GJ, Heule F, Hovius SE. Comparingconventional gauze therapy to vacuum-assisted closure woundtherapy: a prospective randomised trial. J Plast ReconstrAesthet Surg 2007;60:672e81.

6. Timmers MS, Graafland N, Bernards AT, van Dissel JT,Nelissen RGHH, Jukema GN. Negative pressure wound treat-ment with polyvinyl alcohol foam and polyhexamide antisepticsolution instillation in posttraumatic osteomyelitis. WoundRepair Regen 2009;17:278e86.

7. Stannard JP, Gabriel A, Lehner B. Use of negative pressurewound therapy over clean, closed surgical incisions. Int WoundJ 2012;9(Suppl. 1):32e9.

8. Lancerotto L, Bayer LR, Orgill DP. Mechanisms of action ofmicrodeformational wound therapy. Semin Cell Dev Biol 2012;23:987e92.

9. Peinemann F, Sauerland S. Negative-pressure wound therapy:systematic review of randomized controlled trials. Dtsch Arz-tebl Int 2011;108:381e9.

10. Evans D, Land L. Topical negative pressure for treating chronicwounds: a systematic review. Br J Plast Surg 2001;54:238e42.

11. Moues CM, Vos MC, van den Bemd GJ, Stijnen T, Hovius SE.Bacterial load in relation to vacuum-assisted closure woundtherapy: a prospective randomized trial. Wound Repair Regen2004;12:11e7.

12. Gabriel A, Heinrich C, Shores J, et al. Outcomes of vacuum-assisted closure for the treatment of wounds in a paediatricpopulation: case series of 58 patients. J Plast Reconstr AesthetSurg 2009;62:1428e36.

13. Stannard JP, Volgas DA, Stewart R, McGwin Jr G, Alonso JE.Negative pressure wound therapy after severe open fractures:a prospective randomized study. J Orthop Trauma 2009;23:552e7.

14. Blume PA, Walters J, Payne W, Ayala J, Lantis J. Comparison ofnegative pressure wound therapy using vacuum-assistedclosure with advanced moist wound therapy in the treatmentof diabetic foot ulcers: a multicenter randomized controlledtrial. Diabetes Care 2008;31:631e6.

15. Mody GN, Nirmal IA, Duraisamy S, Perakath B. A blinded, pro-spective, randomized controlled trial of topical negativepressure wound closure in India. Ostomy Wound Manage 2008;54:36e46.

16. Chio EG, Agrawal A. A randomized, prospective, controlledstudy of forearm donor site healing when using a vacuumdressing. Otolaryngol Head Neck Surg 2010;142:174e8.

17. Flack S, Apelqvist J, Keith M, Trueman P, Williams D. An eco-nomic evaluation of VAC therapy compared with wounddressings in the treatment of diabetic foot ulcers. J WoundCare 2008;17:71e8.

18. Kolios L, Kolios G, Beyersdorff M, et al. Cost analysis of topicalnegative pressure (TNP) therapy for traumatic acquiredwounds. Ger Med Sci 2010:8 [Doc13].

19. Hiskett G. Clinical and economic consequences of dischargefrom hospital with on-going TNP therapy: a pilot study. J TissueViability 2010;19:16e21.

20. Trueman P. Cost-effectiveness considerations for home healthV.A.C. Therapy in the United States of America and its po-tential international application. Int Wound J 2008;5(Suppl. 2):23e6.