Journal Logo


NPWT Resource Use Compared With Conventional Wound Treatment in Subcutaneous Abdominal Wounds With Healing Impairment After Surgery

SAWHI Randomized Clinical Trial Results

Seidel, Dörthe MD; Lefering, Rolf PhD

Author Information
doi: 10.1097/SLA.0000000000004960


Wound complications are one of the leading causes of postoperative morbidity worldwide, with mortality affecting 1%–4% of patients after gastrointestinal surgery.1,2 Impaired wound healing in the acute setting is typically caused by surgical site infections (SSIs), wound dehiscence (without infection), and periwound maceration.2–5 Dependent on the extent and infection status of the resulting lesion, adapted wound treatment is necessary. Subcutaneous abdominal wound healing impairment (SAWHI) after surgery requires immediate treatment to prevent fascial dehiscence and progression into deeper tissue layer and underlying organs.

Wound complications not only affect the patient's health status, quality of life, and mortality, but are associated with a negative economic impact. Economic impact is best reported for SSIs. SSIs, regardless of severity, increase hospitalization time 6,7 and incur considerable extra healthcare costs associated with additional hospital days and treatment.8 To reduce hospitalization time, patients are frequently transferred to the home care setting, resulting in a resource use shift.9 A substantial number of resources are used in outpatient facilities and home care.9,10

Negative pressure wound therapy (NPWT) represents a clinical wound care innovation that is not only a potentially effective alternative to conventional wound treatment (CWT),11–15 but can also help reduce resource use.10 Published literature examining the use of NPWT in wounds of various origin demonstrated that NPWT could reduce rehospitalizations, associated surgical procedures, dressing changes, personnel commitments, hospitalization and treatment time, and time until subsequent surgery.16–21 However, there is an increasing demand for quality outcome data to support the economic decision-making process with attention to resource use efficiency and assessment of consequence rather than simplistic cost arguments, particularly in postsurgical wounds.10,22 To best of our knowledge, no prospective study published to date compared the resource use of NPWT and CWT in SAWHI after surgery.

For SAWHI after surgery treated with NPWT, we previously reported significantly shorter wound closure time but an increased number of wound-related adverse events, whereas QoL was not significantly different from CWT.23,24 The aim of this evaluation was to compare inpatient and outpatient direct and indirect resource use of NPWT and CWT in SAWHI after surgery.


Study Design

This resource use evaluation was performed as an add on to the randomized clinical SAWHI trial which was conducted in 34 abdominal surgical departments in Germany, Belgium, and the Netherlands. The study protocol (Supplement) and the informed consent documents were approved by the lead ethical committee of the Witten/Herdecke University. The study was registered with the Identifier: NCT01528033.


Adult patients (age ≥18 years) with spontaneous wound dehiscence, reopened suture, or open wounds that could not be closed by primary intention after abdominal surgery were screened for study participation. A minimum wound size was required for correct application of the NPWT device, which was applicable for both treatment arms due to randomization. Inclusion, randomization, debridement or thorough wound cleansing, and treatment start were to be performed within 48 hours after diagnosis of the SAWHI. A nonclosable defect of the abdominal fascia was an exclusion criterion. Patients expected to be incompliant with the protocol and study-related requirements, or participating in another trial, which was thought to interfere with the study procedures, patient's compliance, wound healing, or targeted end points, were excluded. Patients were also excluded when receiving concomitant therapies or procedures deviating from the standard clinical wound care or had investigational character within 30 days before screening or with the need for concomitant therapies or procedures directly affecting wound healing. Pregnant women and patients with any pre-existing or ongoing organ system failure, which could not be stabilized or solved by appropriate medical treatment, with unremovable necrotic tissue present, with nonenteric and unexplored fistulas or malignancy of the wound were not allowed to participate. The use of any other NPWT device on the study wound within ≤8 days before screening was an exclusion criterion.

The initially defined wound size limitation was deleted because this was not in line with clinical practice and led to difficulties with patient inclusion.

Randomization and Masking

After providing written informed consent, patients were randomly allocated to NPWT or CWT in a 1:1 ratio using a computer-generated list created by the trial statistician located on a centralized web-based tool hosted by a professional information technology service. The randomization list consisted of permuted blocks of variable length, which were randomly arranged. Patients were stratified by study site and wound size (≤60 cm3 and >60 cm3). Each registered investigator received individual access to the tool without knowing the randomization sequence, which ensured allocation concealment. The investigators were responsible for adequately implementing the assigned therapy. Neither study participants, medical staff, nor resource use outcome assessors were blinded to the treatment assignment.


At baseline, patients received an extensive examination of the study wound, actual surgical history, and overall health status. After wound debridement or thorough wound cleansing, study treatment started either in-hospital or outpatient and was to be continued in outpatient care whenever possible.

In the intervention arm, commercially available CE-marked NPWT systems (V.A.C. Therapy, KCI, now part of 3M Company, San Antonio, TX) and consumables, were used at the discretion of the clinical investigator and according to manufacturer's instructions. Mainly reticulated open cell foam (ROCF, V.A.C. GRANUFOAM, KCI) dressings were used as indicated for dehisced wounds. Polyvinyl alcohol (V.A.C. WHITEFOAM, KCI) dressings were used for superficial and sensitive wounds. ROCF silver (V.A.C. GRANUFOAM SILVER, KCI) dressings were used for wounds with need for a barrier to bacterial penetration. NPWT as interim therapy was discontinued once the condition of a wound was suitable for closing, either by epithelialization or surgically.

The control therapy was CWT, which was any local wound treatment regularly used in the respective study site that did not have an experimental status or was NPWT. CWT was applied according to the hospitals’ local clinical standards and guidelines, based on the individual needs of the wound in the process of healing, and with special attention paid to exudate amount and local infection status.

In both treatment arms, wound-related procedures were performed when considered clinically necessary. Wounds were closed either surgically or by secondary intention. In the NPWT arm, secondary healing was achieved with CWT dressings after NPWT was discontinued.

Study visits were performed weekly until the end of maximum study treatment time at day 42. All study participants were followed until 132 days after randomization.

Quality assurance of the data collection was ensured by 100% monitoring of the study data documented in the case report forms. Clinical research associates visited the study sites regularly with the frequency of visits adapted on the number of patients randomized per study site.


Resource use parameters assessed included time, personnel (physician, nurse, nursing assistance, and relatives), and material. Evaluation of care and treatment periods included inpatient and outpatient treatment length, length of hospital stay, time to first discharge from hospital, and hospital readmissions. Time, personnel, and material required for dressings changes were documented within booklets for inpatient and outpatient care. Wound treatment material was placed in bags belonging solely to the patient treated. Each material entry was documented in the booklets with category, name, size, unit, and amount.

Wound-related procedures were documented with date of performance and time and human resources used. Surgical wound closure and wound closure by secondary intention were documented with number of study participants, location, anesthesia performed, time and human resources used, and associated procedures performed (wound cleansing and lavage, debridement, wound drainage application, and performance of secondary suture, skin graft or flap).

Indirect resource use included the period of working disability, and mobility and leisure activities at baseline and at the end of the active study treatment period after 42 days or during the wound closure confirmation visit.

In addition to the clinical effectiveness analysis, the number of closed wounds within 132 days was determined. The criteria defined in the study protocol for complete, verified wound closure lasting at least 14 days were not applied. Any wound closure documented by the clinical investigator that did not conflict with another documented outcome was considered. Other than a priori planned, direct nonmedical resource use (transport; care and housekeeping services) was not assessed due to problems of the study participants to provide the information.

Statistical Analysis

Assuming a complete wound closure rate of 50% in the CWT arm and a minimum difference of 12.5% between the treatment arms after 42 days, 492 study participants were calculated to be necessary to achieve 80% power (β = 0.2) with α = 0.05.23,25 Due to 1 planned interim analysis after 250 participants completed 42 days, α was adjusted using the O’Brien-Fleming method (α = 0.005 for the interim analysis and α = 0.048 for the final analysis), which led to a marginally increased sample size of 498 participants.24 Interim results did not show the predefined positive effect at P less than 0.005 or a negative effect at a P less than 0.05 level for NPWT. The study was continued without sample size adjustment.

Resource use analysis was primarily based on the per protocol (PP) population representing study participants without treatment changes and with complete documentation, excluding patients violating inclusion and exclusion criteria, with unauthorized treatment changes, deviations from the recommended frequency of NPWT dressing changes, early treatment termination, or without valid documentation until wound closure confirmation or end of maximum treatment time. However, the results of the modified intention-to-treat (ITT) population including all randomized participants with a valid baseline and at least 1 post baseline wound assessment (Fig. 1) were analyzed secondarily, since this study population corresponds to real-life.

Study Participant flow diagram in the subcutaneous abdominal wound healing impairment (SAWHI) randomized clinical trial. Patient flow diagram according to consolidated standards of reporting trials (CONSORT), including reasons for exclusions from the per protocol (PP) population and distribution of study participants across treatment sectors (inpatient and outpatient care). CWT indicates conventional wound treatment; EOMTT, end of maximum treatment time; NPWT, negative pressure wound therapy; WC, wound closure; WCC, wound closure confirmation.

Resource use parameters are presented descriptively with mean and standard deviation, (minimum and maximum) per study participant or per treatment procedure as applicable. Statistical significance was determined using the Chi-squared or Mann-Whitney U test, respectively, with an alpha level of .05. SPSS statistical software, version 23 (IBM Inc, Armonk, NY), was used for all analyses.


Between August 2, 2011, and January 31, 2018 539 patients were randomized in 34 study sites with the last patient follow-up visit on June 11, 2018. Recruitment was completed. 331 study participants were analyzed in the PP population. The patient flow according to Consolidated Standards of Reporting Trials (CONSORT), including inpatient and outpatient care periods is provided in Figure 1. Reasons for screening failures and exclusion from the ITT and the PP population were previously reported.24

Demographics and Baseline Characteristics

Demographics and baseline characteristics of the PP population relevant for resource use analyses are provided in Table 1 and eTable 1 and 2 in the Supplement, Wound volume was slightly lower in the CWT arm. The mean number of comorbid diagnoses per study participant were the same in both treatment arms. More study participants with CWT had cardiovascular and endocrine/metabolic comorbidities at baseline. Details on the mandatory last and final wound pretreatment procedure performed before initiation of study treatment are provided in eTable 3 in the Supplement,

TABLE 1 - Demographics and Selected Baseline Parameters of the PP Study Population
Randomized Treatment Arms NPWT CWT
Study participants in the PP population, no. 157 174
Age in years, median (IQR) [min-max] 64 (19) [26–88] 66 (18) [26–91]
Sex, male/female, no. (%) 96 (61.1)/61 (38.9) 92 (52.9)/82 (47.1)
Study participants with documented comorbidities, no. 149 160
Diagnoses per study participant, mean (SD) [min-max] 4.2 (2.6) [1–16] 4.2 (2.9) [1–16]
Diagnoses per body system/study participants with diagnosis, no.:
 Cardiovascular 134/88 155/102
 Respiratory 32/25 29/24
 Gastrointestinal 195/107 222/111
 Renal, genitourinary 48/40 46/37
 Metabolic, endocrine 84/56 95/71
 Neurologic 6/6 23/17
 Dermatologic 12/12 8/6
 Hematologic 16/14 17/15
 Musculoskeletal 20/15 22/19
 Head, eyes, ears, nose, throat 10/8 7/5
 Psychiatric 9/9 5/5
 Other 55/40 47/33
Reason for initial hospitalization no.
 Elective surgery 77 87
 Diagnostic procedure 6 4
 Acute setting 38 39
 Emergency 37 44
Hospitalization time before screening (d)
 Mean (SD) [min-max] 11 (7) [0–35] 13 (11) [0–81]
CWT indicates conventional wound treatment; IQR, interquartile range; NPWT, negative pressure wound therapy; PP, per protocol; SD, standard deviation.

Inpatient and Outpatient Care and Treatment Periods

In the PP population, total treatment length within 42 days was significantly shorter in the NPWT arm (mean difference 7.8 days; P < 0.001), with the extended treatment in the CWT arm predominantly conducted during outpatient care (Table 2). Total outpatient treatment time was significantly shorter in the NPWT arm (mean difference 6.2 days; P = 0.017). Hospitalization time was 2.1 days shorter with CWT (P = 0.047) (Table 3). In the CWT arm, 6.8% less study participants were treated exclusively inpatient and 4.2% more participants were discharged from hospital within 42 days, but in case of hospital discharge, time to discharge was only marginally longer in the NPWT arm (mean difference 0.9 days; P = 0.306). Overall, 6.8% more study participants with CWT than with NPWT were treated outpatient. Hospital readmissions were approximately the same in both treatment arms (difference 0.6%; P = 0.875). After day 42, 164 study participants [NPWT 60 (38.2%); CWT 104 (59.8%)] were still in care.

TABLE 2 - Treatment Periods Within 42 d in the PP Population
Randomized Treatment Arms/Statistical Significance NPWT CWT P-value (Test Used)
Study participants in the PP population, no. 157 174 NA
Length of treatment (d), no. 157 174
 Mean (SD) 22.8 (13.4) 30.6 (13.3) <0.001 (U)
 Min-Max 3–42 2–42
Length of NPWT (d), no. 157 NA
 Mean (SD) 14.6 (9.1) NA NA
 Min-Max 1–42 NA
Length of CWT (d), no. 64 174
 Mean (SD) 20.2 (9.3) 30.6 (13.3) NA
 Min-Max 3–41 2–42
Study participants with inpatient NPWT, no. 144 0 NA
Length of inpatient NPWT
 Mean (SD) 10.1 (6.6) NA NA
 Min-Max 1–30 NA
Study participants with inpatient CWT, no. 33 159 NA
Length of inpatient CWT (d)
 Mean (SD) 9.1 (6.9) 11.9 (10.0) NA
 Min-Max 1–25 1–42
Study participants with outpatient NPWT, no. 65 0 NA
Length of outpatient NPWT (d)
 Mean (SD) 12.7 (8.3) NA NA
 Min-Max 1–35 NA
Study participants with outpatient CWT, no. 56 138
Length of outpatient CWT (d)
 Mean (SD) 17.4 (9.7) 24.9 (11.7) NA
 Min-Max 1–39 1–42
CWT indicates conventional wound treatment; NA, not applicable; NPWT, negative pressure wound therapy; PP, per protocol; SD, standard deviation; U, Mann-Whitney U test.

TABLE 3 - Care Status Within 42 d in the PP Population
Randomized Treatment Arms/Statistical Significance NPWT CWT P-value (Test Used)
Study participants in the PP population, no. 157 174 NA
Study participants with inpatient care, no. (%) 144 (91.7%) 158 (90.8%) 0.769 (C)
Length of hospital stay (d)
 Mean (SD) 13.9 (11.1) 11.8 (10.8) 0.047 (U)
 Min-Max 0–42 0–42
Study participants exclusively treated inpatient, no. (%) 17 (10.8%) 7 (4.0%) NA
Length of hospital stay for study participants exclusively treated inpatient
 Mean (SD) 32.1 (9.5) 37.6 (7.6) NA
 Min-Max 17–42 25–42
Length of hospital stay excluding study participants exclusively treated inpatient, no. 140 167
 Mean (SD) 11.6 (9.1) 10.7 (9.5) 0.219 (U)
 Min-Max 0–40 0–41
Study participants with inpatient start and discharged until end of study treatment period of 42 d, no. (%) 127 (80.9%) 148 (85.1%) NA
Time until first discharge from hospital (d), no. 127 148
 Mean (SD) 12.9 (8.3) 12.0 (8.0) 0.306 (U)
 Min-Max 2–41 2–42
Study participants with outpatient care, no. (%) 140 (89.2%) 167 (96.0%) 0.017 (C)
Length of outpatient care (d)
 Mean (SD) 19.3 (12.2) 25.5 (12.2) <0.001 (U)
 Min-Max 0–42 0–42
Study participants exclusively treated outpatient, no. (%) 13 (8.3%) 16 (9.2%) NA
Number of study participants with hospital readmission within 42 d, no. (%) 18 (11.5%) 19 (10.9%) 0.875 (C)
C indicates Chi-squared test; CWT, conventional wound treatment; NA, not applicable; NPWT, negative pressure wound therapy; PP, per protocol; SD, standard deviation; U, Mann-Whitney U test.

More information on treatment periods and care status is available in Table 2, Table 3, and Figure 1. In study participants remaining with open wounds after 42 days, treatment was continued during follow up (eTable 4 in the Supplement,

Surgical Wound Closure and Wound Closure by Secondary Intention

Within 42 days, surgical wound closure was performed in 103 of 331 (31.1%) study participants [NPWT 69 of 157 (44.0%); CWT 34 of 174 (19.5%)] (eTable 5 in the Supplement, In 66 of 331 (19.9%) participants [NPWT 31 of 157 (19.8%); CWT 35 of 174 (20.1%)] wounds were closed by secondary intention. Whereas surgical wound closure in the CWT arm was approximately evenly distributed between the inpatient and outpatient settings, the majority of surgical wound closure in the NPWT arm was performed in the hospital. Wound closure by secondary intention was predominantly achieved in the outpatient setting in both treatment arms. Additional information on the surgical wound closure procedure is provided with eTable 6 in the Supplement,

Post Hoc Analysis of Wound Closures Within 132 Days

In the PP population, for a total of 288 of 331 (87.0%) study participants [NPWT 140 of 157 (89.2%); CWT 148 of 174 (85.1%)] a closed wound was documented within 132 days (Chi-squared test; P = 0.266) (eTable 7 in the Supplement, Time to this first wound closure documentation was significantly shorter in the NPWT arm [median 35 (95% confidence interval 29.2–40.8) days] than in the CWT arm [median 57 (95% confidence interval 46.6–67.4) days, P < 0.001].

Time, Human Resources, and Material Required for Dressing Changes and Wound-related Procedures

In the NPWT arm, significantly fewer dressing changes were performed (mean difference 10.1) with a significantly lower time expenditure per study participant (mean difference 81.7 minutes) in total and in both care settings (Table 4). Physician contacts were significantly less frequent with NPWT (mean difference 2.7, P = .011) (eTable 8 in the Supplement, However, their time spent per study participant was nearly the same in the treatment arms (mean difference 3.4 minutes, P = 0.585). Nurse contacts and their total time expenditure were significantly lower with NPWT (mean difference 7.7, P < 0.001 and 69.1 minutes, P = 0.001, respectively). In both treatment arms, more dressing changes were performed outpatient than inpatient. CWT dressing changes following NPWT in case of wound closure by secondary healing, were mainly performed outpatient (eTable 9 in the Supplement,

TABLE 4 - Time Required for Dressing Changes Within 42 d in the PP Population
Randomized Treatment Arms/Statistical Significance NPWT CWT P-value (Test Used)
Study participants in the PP population, no. 157 174 NA
Number of dressing changes per study participant, no. 133 152
 Mean (SD) 10.7 (10.4) 20.8 (13.0) <0.001 (U)
 Min-Max 1–80 1–72
Time required for dressing changes per study participant (min), no. 133 152
 Mean (SD) 195.8 (221.1) 277.5 (208.2) <0.001 (U)
 Min-Max 5–2098 10–1315
Inpatient dressing changes per study participant, no. 119 130
 Mean (SD) 5.5 (6.1) 10.6 (10.3) <0.001 (U)
 Min-Max 1.0–48.0 1.0–71.0
Time required for inpatient dressing changes per study participant (min), no. 118 130
 Mean (SD) 109.8 (122.1) 128.8 (128.4) 0.234 (U)
 Min-Max 10.0–951.0 5.0–765.0
Outpatient dressing changes per study participant, no. 82 118
 Mean (SD) 9.4 (10.8) 15.0 (11.3) <0.001 (U)
 Min-Max 1.0–78.0 1.0–47.0
Time required for outpatient dressing changes per study participant (min), no. 81 118
 Mean (SD) 161.5 (243.7) 215.6 (186.0) 0.010 (U)
 Min-Max 5.0–2053.0 3.0–888.0
CWT indicates conventional wound treatment; NPWT, negative pressure wound therapy; PP, per protocol; SD, standard deviation; U, Mann-Whitney U test.

Dressing material and wound dressings used during inpatient and outpatient care are listed in eTable 10 in the Supplement,

Significantly fewer wound-related procedures per study participant (mean difference 16.1, P < 0.001) with a correspondingly lower time expenditure (mean difference 99 minutes, P < 0.001) were performed in the NPWT arm (Table 5). The number of procedures involving nurses, assistant physicians, and specialist physicians was higher with CWT than with NPWT (eTable 11 in the Supplement, Time for nurses per study participant was significantly lower in the NPWT arm (mean difference 49 minutes, P = 0.038). Time expenditure for assistant and specialist physicians was not significantly different between the treatment arms.

TABLE 5 - Wound Related Procedures Performed Within 42 d in the PP Population
Randomized Treatment Arms/Statistical Significance NPWT CWT P-value (Test Used)
Number of study participants in the PP population 157 174 NA
Number of study participants with wound related procedures 156 173 NA
Wound related procedures per study participant, no. 157 174
 Mean (SD) 20.5 (17.5) 36.6 (28.0) <0.001 (U)
 Min-Max 0–124 0–174
Total time expenditure per study participant (min)
 Mean (SD) 167 (195) 266 (313) <0.001 (U)
 Min-Max 0–1060 0–2019
Study participants with inpatient wound related procedures, no. 137 150 NA
Inpatient wound related procedures per study participant, no. 157 174
 Mean (SD) 9.7 (13.6) 16.2 (21.8) 0.013 (U)
 Min-Max 0–124 0–119
Time expenditure per study participant (min)
 Mean (SD) 85 (119) 107 (178) 0.365 (U)
 Min-Max 0–920 0–1294
Study participants with outpatient wound related procedures, no. 117 151 NA
Outpatient wound related procedures per study participant, no. 157 174
 Mean (SD) 10.5 (12.6) 19.8 (19.8) 0.001 (U)
 Min-Max 0–57 0–104
Time expenditure per study participant (min)
 Mean (SD) 80 (143) 154 (270) 0.001 (U)
 Min-Max 0–990 0–1952
CWT indicates conventional wound treatment; NA, not applicable; NPWT, negative pressure wound therapy; PP, per protocol; SD, standard deviation; U, Mann-Whitney U test.

More details on dressing changes and wound-related procedures with separate consideration of the post NPWT CWT period, and information on dressing material and wound dressings used during inpatient and outpatient care are provided in Table 4, Table 5, and in eTables 11–14 in the Supplement,

Indirect Resource Use: Working Disability and Activities of Daily Living

A total of 64 study participants with employment status available at the end of the active study treatment time were gainfully employed [NPWT 26/118 (22.0%); CWT 38/137 (27.7%)]. The duration of work disability was similar in both treatment arms (mean (standard deviation) [Min-Max] NPWT N = 24, 40.4 (5.6) [25–45]; CWT N = 31, 38.3 (10.5) [4–45]); P = 0.592; Mann-Whitney U-test).

Study participants were less mobile after 42 days than at baseline, and outdoor leisure activities and sports decreased (Table 6).

TABLE 6 - Mobility and Leisure Activities at Baseline and After End of Study Treatment in the PP Population
Randomized Treatment Arms NPWT CWT
Study Participants in the PP Population, no. 157 174
Survey date Baseline End of Study Treatment Time After 42 d or at Wound Closure Confirmation Baseline End of Study Treatment Time After 42 d or at Wound Closure Confirmation
 Study participants with information available, no. 157 115 174 132
 Bedridden or mobilized in a wheel chair, no. 5 6 8 7
 Able to move freely within home, no. 18 8 15 13
 Able to leave home, no. 134 102 152 114
Leisure activities
 Study participants with information available, no. 156 113 171 130
 Indoor activities, no. 29 31 44 37
 Outdoor activities, no. 106 74 95 88
 Recreational sports, no. 31 9 51 12
 Competitive sports, no. 1 1 1 0
Multiple answers possible.CWT indicates conventional wound treatment; NPWT, negative pressure wound therapy; PP, per protocol.

Evaluating “Real-life” With the ITT Population

In the ITT population, due to treatment changes NPWT was also performed in some study participants in the CWT arm. Length of hospital stay was shorter in the CWT arm but, other than in the PP population, the difference between the treatment arms was not significant. All other care and treatment results show no relevant deviations from those in the PP population. Time and human resources used for dressing changes and wound-related procedures were similar to those the PP population. Further details on the results of the ITT population are provided in eTables 15–29 in the Supplement,


In the SAWHI study, treatment length within 42 days was significantly shorter with NPWT than with CWT. This corresponds to the previously reported finding that NPWT was superior to conventional dressings in achieving complete closure.24 Moreover, in the post hoc analysis on wound closure throughout the observation period, we showed that although the wound closure rate became more similar between the treatment arms within 132 days, the time to first wound closure documentation was significantly shorter in the NPWT arm.

Intersectoral wound treatment was realized with the majority of study participants starting treatment in hospital and being discharged within 42 days. The assumed substantial number of resources used in outpatient facilities and home care,9 was shown also to be applicable for patients with SAWHI after surgery. Length of hospital stay was 2 days shorter with CWT. Findings of studies with other indications which showed NPWT to shorten the hospitalization period, could not be confirmed for the specific patient group included in the SAWHI-study.17–21 Significantly more study participants with CWT were treated as outpatients. Participants with NPWT were less often discharged and more often exclusively treated in hospital. If study participants were discharged from hospital, time to first discharge was similar with NPWT and CWT. Hospital readmissions were also equally represented in the treatment arms. More wounds were closed surgically than by secondary intention. Surgical wound closure was more common in the NPWT arm and was mainly performed during inpatient care, which could explain the longer hospital stay while the total treatment time was longer in the CWT arm. However, neither study participants, medical staff, nor resource utilization assessors were blinded to treatment assignment, which may have introduced bias in objective assessment, particularly in the clinical decision to discharge from the hospital with NPWT. Because the overall frequency of comorbidities was the same in both treatment arms and even more cardiovascular and metabolic/endocrine diagnoses occurred in the CWT arm, comorbidities at baseline can be excluded as a cause of prolonged hospitalization time with NPWT.

In both care settings, the number of dressing changes and the associated time consumption were significantly lower with NPWT. CWT dressing changes after NPWT to achieve wound closure by secondary intention, which by their nature are more frequent and require less effort than NPWT changes, made a relevant contribution to outpatient care.

Significantly more wound related procedures per study participant with a correspondingly higher time expenditure were necessary with CWT. Dressing changes and wound-related procedures required more personnel in the CWT arm with nurses being the major human resource used. The majority of resources were used in the outpatient sector, but a simple resource use-shift rather than decreasing the overall time consumption, because NPWT is thought to demand less frequent but more complex and time-consuming dressing changes than CWT,10 could not be confirmed in this study.


As with the analysis of effectiveness and safety of the SAWHI study, the treatment and observation period of 42 days is also a limitation for the resource use analysis. We previously reported that more than 50% of the wounds were still open after 42 days. The small amount of data available before the start of the study severely limited the possibility of an adequate assessment of the required treatment and observation period to close the majority of the wounds. To address this, an additional post hoc analysis of wound closures within 132 days was performed that captured all wound closures regardless of whether they met the study protocol criteria for a complete, sustained, and verified wound closure or not. Nonetheless, the evaluation of resource utilization remains limited to 42 days and may require adjustment in further cost analyses.

Material use was not sufficiently documented in the SAWHI trial. The planned analysis was not possible to be performed, but is limited to a simple listing of dressings used per treatment arm. Information on optimal material use for subsequent cost analyses may be derived from other resources.

Another limitation of this study is the lack of evaluation of intensive care unit days.

No formal cost analysis was performed which, if you think globally, is more of an opportunity than a limitation. The economic status of different countries is a major problem when analyzing disease state cost.10 Unlike previous studies, which typically present country-specific unit costs for personnel, inpatient accommodations, and supplies based on the prevailing billing system, our study presents resource use results that can be used as the basis for any economic analysis worldwide.

Country-specific cost analyses from the perspective of the relevant actors in the respective health care system, incorporating other relevant data sources, are the next step in using the study data presented here.


To our knowledge, the SAWHI trial is the first RCT to report that NPWT reduces treatment time and results in less resource use for dressing changes and wound-related procedures for SAWHI after surgery. NPWT maybe an efficient treatment alternative to CWT, which needs to be demonstrated in a subsequent cost analysis.


The authors thank all clinical investigators and research nurses in the participating hospitals who enrolled at least 1 patient in the SAWHI study. These hospitals are listed with the number of patients randomized: Medizinische Fakultät Mannheim Universität Heidelberg (22); Unfallkrankenhaus Berlin (45); Klinikum Ernst von Bergmann gGmbH Potsdam (27); Chirurgische Klinik und Poliklinik, Klinikum rechts der Isar München (51); Abteilung für Allgemein- und Viszeralchirurgie, Universitätsmedizin Göttingen (37); GRN-Klinik Sinsheim (41); Klinik für Allgemeine Chirurgie, Viszeral-, Gefäß- und Kinderchirurgie, Universitätsklinikum des Saarlandes Homburg/Saar (42); Sana Klinikum Berlin-Lichtenberg, Allgemein-u. Visceralchirurgie (35); St. Josef-Hospital gGmbH, Universitätsklinikum der Ruhr-Universität Bochum (26); Klinik für Allgemein-, Viszeral- und Thoraxchirurgie. Diakoniekrankenhaus Rotenburg (Wümme) gGmbH (19); ZU Brussel (5); Department of Abdominal Surgery, Leuven, Belgium (11); Thüringen-Kliniken “Georgius Agricola” GmbH Saalfeld, Gefäßchirurgie (17); Brandenburg Medical School -University Hospital Neuruppin, Department of General and Visceral Surgery (16); Universitätsmedizin Mainz Allgemein- Viszeral–und Transplantationschirurgie (5); Universitätsmedizin Greifswald, Klinik und Poliklinik für Chirurgie (19); University Hospital Maastricht (3); Universitätsklinikum Rostock Chirurgische Klinik und Poliklinik (6); Hanse Klinikum Stralsund (1); Universitätsklinikum Magdeburg (4); Charité – Universitätsmedizin Berlin (7); Universitätsklinikum Düsseldorf (22); Klinikum Nürnberg Nord (4); Katholischen Kliniken Oberhausen gem. GmbH St. Marien Hospital (6); “Klinikum Offenbach GmbH Chirurgische Klinik (3); Krankenhaus der Barmherzigen Brüder, Abteilung für Allgemein-, Viszeral- und Allgemeinchirurgie Trier (10); Universitätsklinikum Freiburg Allgemein-u. Visceralchirurgie (5); Evangelisches Krankenhaus Oberhausen (28); Klinikum Kempten-Oberallgäu gGmbH (1); Universitätsklinikum Carl Gustav Carus Dresden (9); Viszeral-, Thorax- und Gefäßchirurgie, Klinik für Allgemein/-Viszeralchirurgie, Ketteler Krankenhaus Offenbach am Main (6); Herz-Jesu-Krankenhaus Dernbach (2); Berlin Spandau Vivantes (3); Universitätsklinikum Erlangen, Chirurgie (1). Furthermore, the authors thank all Institut für Forschung in der Operativen Medizin (IFOM) staff members who were involved in the study and all partners. The authors thank the following persons for their support: E. A. M. Neugebauer, Prof, supported the study in his role as former director of the IFOM. Peter Krüger, PD Dr rer nat (IFOM), was responsible for the data management of the study, performed the data transfer and assisted in the statistical analysis. All were employees of the University of Witten/Herdecke and received salary payments from the study grant. We thank all clinical trial associates and research nurses for their work in the study. The following German health services supported the outpatient care of the study participants: Sanitätshaus Kniesche GmbH in Potsdam, Medicops GmbH & Co. KG in Wiesloch, N:AIP Netzwerk Ruhr GmbH in Bochum, Noma-med GmbH in Harsum, and REHA aktiv 2000 GmbH in Saalfeld. We thank the members of the data monitoring committee for monitoring overall study performance and safety.

Christine Bongards, working for 3M, supported the resource analysis with valuable information from the manufacturer's perspective. The authors appreciate her well-structured, mindful, and helpful support without influencing the content of the report.


1. Pearse RM, Moreno RP, Bauer P, et al. Mortality after surgery in Europe: a 7 day cohort study. Lancet 2012; 380:1059–1065.
2. GlobalSurg C. Determining the worldwide epidemiology of surgical site infections after gastrointestinal resection surgery: protocol for a multicentre, international, prospective cohort study (GlobalSurg 2). BMJ Open 2017; 7:e012150.
3. Sandy-Hodgetts K, et al. International best practice recommendations for the early identification and prevention of surgical wound complications. Wounds International. 2020 Available from:
4. World Union of Wound Healing Societies (WUWHS) Consensus Document. Surgical wound dehiscence: improving prevention and outcomes. Wounds International, 2018. Available from:
5. Woodfield JC, Jamil W, Sagar PM. Incidence and significance of postoperative complications occurring between discharge and 30 days: a prospective cohort study. J Surg Res 2016; 206:77–82.
6. Coello R, Charlett A, Wilson J, et al. Adverse impact of surgical site infections in English hospitals. J Hosp Infect 2005; 60:93–103.
7. Weber WP, Zwahlen M, Reck S, et al. Economic burden of surgical site infections at a European university hospital. Infect Control Hosp Epidemiol 2008; 29:623–629.
8. Jenks PJ, Laurent M, McQuarry S, et al. Clinical and economic burden of surgical site infection (SSI) and predicted financial consequences of elimination of SSI from an English hospital. J Hosp Infect 2014; 86:24–33.
9. Probst S, Seppänen S, Gerber V, et al. EWMA document: home care-wound care: overview, challenges and perspectives. J Wound Care 2014; 23: (Suppl 5a): S1–S41.
10. Apelqvist J, Willy C, Fagerdahl AM, et al. EWMA document: negative pressure wound therapy. J Wound Care 2017; 26: (Suppl 3): S1–S154.
11. Ubbink DT, Westerbos SJ, Nelson EA, et al. A systematic review of topical negative pressure therapy for acute and chronic wounds. Br J Surg 2008; 95:685–692.
12. Dumville JC, Owens GL, Crosbie EJ, et al. Negative pressure wound therapy for treating surgical wounds healing by secondary intention. Cochrane Database Syst Rev 2015; 6:CD011278.
13. Webster J, Liu Z, Norman G, et al. Negative pressure wound therapy for surgical wounds healing by primary closure. Cochrane Database Syst Rev 2019; 3:CD009261.
14. Zens Y, Barth M, Bucher HC, et al. Negative pressure wound therapy in patients with wounds healing by secondary intention: a systematic review and meta-analysis of randomised controlled trials. Syst Rev 2020; 9:238.
15. Janssen AH, Mommers EHH, Notter J, et al. Negative pressure wound therapy versus standard wound care on quality of life: a systematic review. J Wound Care 2016; 25:154156–159.
16. Page JC, Newswander B, Schwenke DC, et al. Retrospective analysis of negative pressure wound therapy in open foot wounds with significant soft tissue defects. Adv Skin Wound Care 2004; 17:354–364.
17. Niezgoda JA. The economic value of negative pressure wound therapy. Ostomy Wound Manage 2005; 51: (2A Suppl): 44S–47S.
18. Flack S, Apelqvist J, Keith M, et al. An economic evaluation of VAC therapy compared with wound dressings in the treatment of diabetic foot ulcers. J Wound Care 2008; 17:71–78.
19. Apelqvist J, Armstrong DG, Lavery LA, et al. Resource utilization and economic costs of care based on a randomized trial of vacuum-assisted closure therapy in the treatment of diabetic foot wounds. Am J Surg 2008; 195:782–788.
20. Moues CM, van den Bemd GJCM, Meerding WJ, et al. An economic evaluation of the use of TNP on full-thickness wounds. J Wound Care 2005; 14:224–227.
21. Kaplan M, Daly D, Stemkowski S. Early intervention of negative pressure wound therapy using Vacuum-Assisted Closure in trauma patients: impact on hospital length of stay and cost. Adv Skin Wound Care 2009; 22:128–132.
22. Gottrup F, Apelqvist J, Price P. Outcomes in controlled and comparative studies on non-healing wounds: recommendations to improve the quality of evidence in wound management. J Wound Care 2010; 19:237–268.
23. Seidel D, Lefering R, Neugebauer EA. Treatment of subcutaneous abdominal wound healing impairment after surgery without fascial dehiscence by vacuum assisted closureTM (SAWHI-V.A.C.(R)-study) versus standard conventional wound therapy: study protocol for a randomized controlled trial. Trials 2013; 14:394.
24. Seidel D, Diedrich S, Herrle F, et al. Negative pressure wound therapy vs conventional wound treatment in subcutaneous abdominal wound healing impairment: the SAWHI randomized clinical trial. JAMA Surg 2020; 155:469–478.
25. Dupont WD, Plummer WD Jr. Power and sample size calculations. A review and computer program. Control Clin Trials 1990; 11:116–128.

abdominal wound; conventional wound treatment; negative pressure wound therapy; resource use; surgical wound; wound healing; wound treatment

Supplemental Digital Content

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc.