Pooled Study-level Analysis of Randomized Controlled Trials Analyzing the Effect of Negative Pressure Wound Therapy With Irrigation vs Traditional Negative Pressure Wound Therapy on Diabetic Foot Outcomes

DM is a growing epidemic that is increasing in both prevalence and incidence, with an estimated 8.5% of adults worldwide living with DM as of 2014.¹ DM is associated with multiple comorbidities and was directly responsible for more than 1.5 million deaths worldwide in 2019, with nearly 50% of these deaths occurring prematurely (before age 70 years).¹

A severe sequela of DM is the development of DFU, which is attributed to a multitude of factors resulting from complications of diabetes, including neuropathy, peripheral vascular disease, compromised immune function, and muscle glycosylation leading to biomechanical deformities. Diabetes has historically been the greatest contributor to nontraumatic lower extremity amputations, and foot ulcers remain the primary reason for hospitalization.^2,3 A 2017 study estimated the global prevalence of DFUs to be 6.5%, with prevalence in the United States approximately twice that, at 13%.⁴ DFUs lead to increased amputations, reduced quality of life, and enormous financial burden to both the individual and the community.⁵ A major cause of this morbidity and mortality is the progression of DFU to DFI.^6,7 It has been estimated that over 80% of DFUs progress to DFI, posing a 56 to 155 times increased risk for lower extremity amputation.^5,8 SOC therapy for both DFU and DFI has traditionally consisted of a combination of sharp debridement, bacterial burden control, and offloading. NPWT has become a pivotal tool for improving health outcomes, demonstrating substantial benefit compared with SOC in BWC, time to wound closure, and reduction in amputations.^2,9

The next generation of NPWT has included simultaneous irrigation with solution (NPWT-I) at predetermined settings. NPWT-I continues to show improved outcomes compared with SOC, with the added benefit of decreasing QBCs.^5,8,10-13 Despite the reduction in bacterial burden, debate persists regarding the clinical effect of NPWT-I on diabetic foot outcomes compared with NPWT-T. The authors of the current study investigated the differences in outcomes in patients treated with NPWT-T versus NPWT-I, specifically, BWC, time to wound closure, LOS, and AEs.

PubMed, Medline/Embase, the Cochrane Library, and Web of Science databases were searched for randomized, controlled clinical trials published between January 1, 2002, and March 1, 2022. Keywords used in the search criteria included “NPWT,” “negative pressure wound therapy,” “irrigation,” “instillation,” and “diabetes.” Title and abstract screening for eligibility and selection for full-text review were independently conducted according to study eligibility criteria by 2 authors (A.T. and A.G.). All conflicts were resolved by consensus with a third author (L.L.). This study was conducted in accordance with the PRISMA reporting guidelines (Supplemental eFigure 1), and this systematic
review and meta-analysis is registered with PROSPERO (ID CRD42022347947) and Research Registry (reviewregistry1406).

Inclusion criteria were adults age 18 years or older; English-language articles published between January 1, 2002, and March 1, 2022; a minimum of 75 subjects reported on; subject intervention of either NPWT-T or NPWT-I; comparative, prospective, randomized trial design; and reporting BWC, defined as the absence of soft tissue infection and adequate soft tissue for delayed primary wound closure, local rotational flap, split-thickness skin graft, or composite bioengineered tissue, as an outcome (Supplemental eFigure 2).

Citations were also included based on availability of extractable data from a 2 × 2 confusion matrix. Study populations were categorized as NPWT-T, based on the use of NPWT without simultaneous irrigation, or NPWT-I, based on the use of NPWT with simultaneous irrigation. AEs were extracted based on data reported in each study and were inclusive of rehospitalizations and amputations. Analysis was also performed for the LOS (defined as total number of days hospitalized), time to wound closure, and BWC. LOS and time to wound closure were measured in days. True positives were patients with NPWT-I administration and the presence of any binary outcome reported previously (AE and BWC). False positives were patients with NPWT-I administration and the absence of any binary outcome. True negatives were patients with NPWT-T administration and the absence of any binary outcome. False negatives were patients with NPWT-T administration and the presence of any binary outcome.

Exclusion criteria included case studies, systematic reviews, retrospective cohort studies, nonrandomized trials, and trials that did not include use of both NPWT-T and NPWT-I. Studies in which less than 50% of patients had diabetes and studies in which less than 90% of patients had lower extremity ulcers were also excluded.

Data pooled from each study were used for analysis. All flow diagrams were created in PowerPoint (Microsoft). The software program Comprehensive Meta-Analysis (2006; Biostat Inc) was used to calculate effect sizes, ORs, and SMD with their associated CIs. Statistical heterogeneity was evaluated using the Cochran Q test and I² statistic. A P value less than or equal to .05 was considered significant. Forest plots were generated with the aforementioned meta-analysis software. Funnel plots and Egger’s test of linearity were used to analyze publication bias (Supplemental eFigures 3, 4). All other data extraction and analyses were completed using SAS statistical software (version 9.3; SAS Institute Inc).

A total of 404 citations were generated based on implementation of keyword search criteria across the 4 databases. After duplication removal, title screening, and full-text review, 3 articles were identified and included in this systematic review.^14-16

The 3 studies included in the current study were RCTs published in 2020. NPWT-T was considered the control treatment for each study group, with continuous pressure maintained at 125 mm Hg. The NPWT-I intervention varied between each study regarding the solution used and dwell time. Kim et al¹⁶ used 0.1% polyhexanide plus 0.1% betaine solution as an adjunct to NPWT-T, with a 20-minute dwell time followed by 2 hours of NPWT. Lavery et al¹⁴ used 0.1% polyhexanide-betaine irrigation (NPWT-I) at 30 mL per hour, and Davis et al¹⁵ used 0.9% normal saline infusion at 15 mL per hour. Dressing changes were similar across studies, occurring every 48 to 72 hours. Treatment visits were once weekly for 12 weeks,¹⁵ once weekly for 16 weeks,¹⁴ or every 3 days for 8 weeks.¹⁶

Vascular status inclusion and exclusion criteria were similar across studies using ankle-brachial index, pedal pulse assessment, or toe pressures. In addition, 2 studies reported skin perfusion of the dorsal and plantar surfaces of the foot.^14,15 Wound types included ischemic, neuropathic, decubitus, surgical, venous, traumatic, and other (Table 1). All wound types that were unidentifiable based on context of citation were categorized as “other.”

Conclusions from the 3 included studies indicate a consensus of nonsignificant differences between the use of NPWT-T and NPWT-I for BWC, time to wound closure, LOS, and AEs. Kim et al¹⁶ observed a significant reduction in bacterial load when NPWT-I was used, as well as improved outcomes in subgroups with surgically dehisced wounds.

Pooled data indicated a total of 421 subjects across all studies, with 223 receiving NPWT-T and 198 receiving NPWT-I. There were 177 patients with diabetes in the NPWT-T group, compared with 159 in the NPWT-I group. Participant demographics were similar across all studies (Table 2). No significant differences were identified in the number of patients with diabetes between experimental groups of each study population, with a P value range of .2 to .99. Glycated hemoglobin was reported in Lavery et al¹⁴ and Davis et al,¹⁵ with no difference across groups in Lavery et al.¹⁴ However, Davis et al¹⁵ observed lower glycosylated hemoglobin (10.9% vs 9.6%) for the NPWT-I group. The most common wound location was the foot (N = 308; transmetatarsal amputation = 7, foot = 301), with 2 studies^14,15 examining only the lower extremity; diabetic wounds were the most common etiology. Other anatomic regions included the upper extremity, abdomen, and buttocks, comprising less than 5% of the subjects (20 of 421). Wound size varied across studies. Kim et al¹⁶ reported mean wound area and mean volume of 75.0 cm² and 183.8 cm³, respectively, for NPWT-I and 72.9 cm² and 209.1 cm³, respectively, for NPWT-T. Davis et al¹⁵ reported mean wound area and mean volume of 13.9 cm² and 13.7 cm³, respectively, for NPWT-I and 16.8 cm² and 12.9 cm³, respectively, for NPWT-T. Lavery et al¹⁴ reported a mean wound area of 13.4 cm² ± 11.1 cm² for NPWT-I versus 18.5 cm² ± 19.0 cm² for NPWT-T.

Pooled data analysis of standard difference in means for LOS (SMD, 0.065; 95% CI, −0.128-0.259; P =.508) and time to wound closure (SMD, −0.039; 95% CI, −0.233-0.154; P =.691), and of ORs for BWC (OR, 1.049; 95% CI, 0.709-1.552; P =.810) and AEs (OR, 1.092; 95% CI, 0.714-1.670; P =.69) demonstrated no statistically significant associations between the use of NPWT-T or NPWT-I and the respective outcomes of each in the trials included in this review (Figures 1-4).

In the current systematic review and meta-analysis of RCTs investigating diabetic foot outcomes following treatment with either NPWT-T or NPWT-I, no significant differences were found between NPWT-T and NPWT-I for BWC, time to wound closure, LOS, or AEs.

Approximately 100 000 major leg amputations are performed annually in the United States alone,¹⁸ with 30-day mortality ranging from 7% to 22% and 5-year mortality ranging from 50% to 80%.¹⁹ In addition to the high mortality, approximately 55% of patients who undergo an amputation related to DM or peripheral vascular disease will become permanently disabled and never return to ambulatory status.²⁰ Amputation not only affects quality of life^21,22 and life expectancy,⁶ but also adds substantial financial burden to the individual patient and government health systems. The US Medicare program spends an average of $4.3 billion annually for the treatment of nontraumatic lower extremity amputations, with costs of approximately $100 000 per individual for the index procedure and outpatient care.^23-25

The use of NPWT has been proven to reduce secondary amputations and has been shown to reduce costs through decreased LOS, decreased total number of surgeries, and decreased time to healing compared with NPWT-T.^26,27 The advent of NPWT demonstrated an advance in limb salvage and has proved superior to SOC in several studies.^28,29 Animal studies have demonstrated a reduction in QBCs with the use of NPWT; however, these findings have not translated into an anticipated reduction in clinical infection.^9,16

Animal studies have shown a significant reduction in QBCs with NPWT-I compared with NPWT alone.^{5,10-13,30,31} It was expected that lower QBCs would translate into reduced infections, higher rates of healing, and faster times to complete closure. The results of the current meta-analysis indicated no significant differences, however, with the addition of irrigation.

Many studies highlighting the efficacy of NPWT-I have small sample sizes with varying wound etiology and/or location as well as unknown glycemic control, vascular status, dwell times, and irrigation solutions.^17,32-42 The antiseptic has since been observed to offer no benefit compared with normal saline, with one RCT determining no difference in clinical outcomes.⁴³ The addition of periodic irrigation mechanically decreases the wound fluid viscosity and facilitates more efficient removal of exudate, superficial and deep cellular debris, and infectious material.^44,45 The reduction in QBCs is likely the result of mechanical debridement provided by the alternating dwell and negative pressure cycles rather than the solution used.

Decreasing the level of bioburden has been touted as one of the pillars of wound healing and a central element in dressing selection and treatment approaches in the management of complex wounds. The rationale has been that there is a critical level of bacterial colonization that is high enough to impede healing, but not high enough to cause clinical signs of infection.⁴⁶

While previous studies have noted a reduction in QBC with NPWT-I,^{5,10-13,30,31} one study included in the current meta-analysis observed no difference in bacterial count at the first dressing change compared with predebridement between NPWT-T and NPWT-I.¹⁶ A major limitation of the RCTs in the current study is the subjective nature of successful debridement. Aggressive debridement and binary success of first debridement is a subjective variable determined by the operating physician. The similarities in outcomes between NPWT-I and NPWT-T may be attributed to the effectiveness of the initial debridement. It is reasonable that should an aggressive debridement eradicate the infection, the proposed benefit of periodic irrigation on the reducing QBCs may be muted. In addition, the dose and duration of irrigation may have been insufficient to observe a clinical benefit; more RCTs are needed to evaluate optimal dose and dwell durations for DFUs and DFIs.

The limitations of this study include the subjective nature of successful wound debridement. Whether or not wound closure was appropriate was at the discretion of the surgeon, with no clear criteria. In addition, the irrigation solution varied across the 3 studies included in this meta-analysis. One major limitation was the large number of studies discarded during the screening and full-text review process. Many articles contained data that were irrelevant to the objective of this analysis, did not assess the primary outcomes the authors of the current study were examining, or did not have extractable 2 × 2 data. Despite using only RCTs as level 1 evidence, the sample size remains limited, and further RCTs are required to assess the role of NPWT-I for the treatment of DFU and DFI.

NPWT-I for DFU and DFI does not improve outcomes compared with NPWT-T. The use of NPWT-I is not significantly different from NPWT-T for BWC, time to wound closure, LOS, or AEs. Further RCTs are required to assess the role, if any, of NPWT-I in the management of DFU and DFI.