Comparison of Negative Pressure Wound Therapy with or without a Split-Thickness Skin Graft in the Surgical Management of Axillary Hidradenitis Suppurativa: A Retrospective Cohort Study

Abstract

Introduction: Negative pressure wound therapy (NPWT) alone or with the addition of a split-thickness skin graft (STSG) are 2 reconstructive options available after surgical excision of axillary hidradenitis suppurativa (HS). The aim of this study was to retrospectively examine patients undergoing these treatments and to assess clinical and patient-related outcome measures. Methods: A single-centre, retrospective analysis was conducted, evaluating surgical excision of axillary HS, with STSG and NPWT, or NPWT alone. Data collected included No. of post-operative clinic visits, time to heal, size of wound, disease recurrence, follow-up time, Dermatology Life Quality Index (DLQI), the Generalised Anxiety Disorder Assessment (GAD-7), the Patient Health Questionnaire Depression Scale (PHQ-9), Pain Visual Analogue Scale (PAINVAS2), the Brief Illness Perception Questionnaire (BIPQ), and Dermatology Visual Analogue Scale (DERMVAS). Two-tailed t-test and Mann–Whitney Wilcoxon U-tests were used to assess for significant relationships. Results: One hundred five patients were included in the study, 44 who received NPWT alone, and 61 who received NPWT + STSG. There was no significant difference in follow-up time (P = .934) or No. of follow-up appointments between groups (P = .287). There was a significant difference in time to heal between groups, with STSG + NPWT observing a mean time of 2.77 months and NPWT alone observing a mean time of 4.40 months (P = .0006). There was no difference in patient-reported outcomes between the 2 groups. Conclusion: There is no difference in patient-reported outcomes with the addition of an STSG to NPWT after surgical excision of HS. Wide excision and use of NPWT alone is an effective procedure for the treatment of axillary HS.

Introduction

Hidradenitis suppurativa (HS) is a chronic, idiopathic condition, characterised by inflammation in areas of skin containing apocrine sweat glands, leading to abscess formation, and scarring.^{1, 2} The disease can occur at multiple sites including axillae, groin, inframammary fold, and perineum with the axilla being most commonly affected. ³ Estimates of the prevalence of HS have ranged from 3.3 to 100 per 10 000 population^4–6. HS carries a considerable psychosocial burden, and there is often a significant delay in diagnosis. ⁷ The functional impairment associated with HS can be similar or greater than other chronic dermatological and medical conditions, such as psoriasis, atopic dermatitis, heart disease, asthma, and diabetes.^8,9

HS is a complex disease ideally managed with a multidisciplinary approach. ¹⁰ Medical treatment is typically the initial management but does not provide local cure, just disease control. Surgical excision, however, offers patients a curative treatment option for affected areas. Given the significant psychological and economic burden of the condition, early surgical treatment, in combination with systemic therapies and lifestyle changes, should be considered for poorly controlled disease.^10,11

There are several surgical treatment options. These include incision and drainage of abscesses, excision of lesions with direct closure, and wide excision with reconstruction. Incision and drainage or excision of small lesions with direct closure can offer short-term relief but are associated with high levels of recurrence.^12,13 Numerous reconstructive procedures have been described, including local flaps, wide local excision with skin grafting, wide local excision and negative pressure wound therapy (NPWT), and direct closure^14–25. Recent guidelines from the British Association of Dermatologists have recommended either secondary intentional healing or thoracodorsal artery perforator (TDAP) flap for axillary HS. ¹⁰

The use of NPWT is well documented in the management of HS^16,18,22,23 although, most commonly in combination with a split-thickness skin graft (STSG). There is other evidence reporting the outcomes of secondary intention healing alone, but no application of NPWT. ²⁴ The addition of NPWT to a skin graft, particularly in highly mobile areas such as the axilla, is intended to reduce graft failure. However, this does introduce a second surgical site with the resultant scar and potential for complications.

Outcomes for surgical treatment of HS within the literature primarily focus on objective measures such as healing time, recurrence, and complications. Although surgical and clinician-assessed outcomes are more easily measured, these often correlate poorly with self-reported outcomes. ²⁶ Given the disease's psychosocial consequences, it is important to assess patients’ impressions of outcomes. Patient-reported outcome measures (PROMs) provide a patient's perspective of potential treatment benefits and harms. In a disease process such as HS, PROMs provide a key insight into treatment efficacy.

Our aim was to retrospectively analyse all patients undergoing wide excision of axillary HS at our unit and compare the surgical and patient-reported outcomes of wide excision with STSG and NPWT or wide excision and NPWT alone.

Methods

A single-centre, retrospective analysis was conducted, evaluating outcomes for all patients undergoing surgical excision of axillary HS, with STSG and NPWT, or NPWT alone, in a cohort study. The study centre is a tertiary referral unit for HS and data was collected in August 2019 for patients treated between January 2014 and June 2018 inclusive. Patients referred to the service are most commonly Hurley Stage 2 or 3 and this, therefore, constituted inclusion criteria. All patients undergoing treatment were considered eligible and follow-up was undertaken as part of routine care.

PROMs were obtained through Integrating Mental and Physical Healthcare Research, Training and Services (IMPARTS) questionnaires, which were routinely completed by patients before and after surgery. Data collected included the Dermatology Life Quality Index (DLQI), the Generalised Anxiety Disorder Assessment (GAD-7), the Patient Health Questionnaire Depression Scale (PHQ-9), Pain Visual Analogue Scale (PAINVAS2), the Brief Illness Perception Questionnaire (BIPQ), and Dermatology Visual Analogue Scale (DERMVAS), which includes disease severity, odour, discharge, pain, and fatigue scales.

A database was compiled that consisted of information on No. of post-operative clinic visits (surgical team and specialist nurses), time to heal, size of wound, disease recurrence, follow-up time, and IMPARTS scoring. Descriptive statistics, in the form of means + standard deviation, were utilised for the above data values. For those patients who underwent NPWT with an STSG, donor site complications were recorded. Follow-up intervals were determined by the clinical team.

Post-operative IMPARTS data, where possible, was collected from a questionnaire taken at least 6 weeks after the procedure to minimise the influence of immediate post-operative effects. The maximum post-operative timeframe for IMPARTS data inclusion was 6 months. To minimise potential selection bias, authors not directly involved in the care of the patients undertook the data collection and analysis.

Patients were then assigned to either of 2 cohorts for analysis; NPWT with an STSG (NPWT + STSG) or NPWT alone (NPWT alone). For the analysis, the data interface RStudio ²⁷ was used. For the analysis of non-IMPARTS outcomes, two-tailed t-test was used as a way of observing changes and statistical significance. For analysis of IMPARTs outcomes, Mann–Whitney Wilcoxon U-tests were used to assess for significant relationships. Wilcoxon signed-rank tests were used to assess for differences between pre-operative and post-operative scores for each closure method. As this was an exploratory analysis, no correction was utilised. Multiple imputation analysis was not undertaken. This study adheres to current STROBE guidelines. ²⁸

Results

A total of 298 records were identified in the initial clinical coding process. One hundred five patients underwent reconstructive procedures for axillary HS (35 male and 70 female patients), 44 received NPWT alone, and 61 patients received NPWT plus an STSG (Figure 1).

Figure 1.

A flowchart illustrating the exclusion and inclusion criteria for cases in the study and showing where patients were removed and the number who were removed or included.

The mean age of the STSG + NPWT group was 39.0 and the mean age of the NPWT alone group was 35.8. Of the 44 NPWT alone patients, 23 had completed IMPARTS data recorded pre- and post-operation. Of the 61 NPWT plus STSG cohort, 24 patients had completed IMPARTS data.

For wound area, there was no significant difference between the 2 groups, with STSG + NPWT observing a mean wound area of 90.7cm² (SD = 58.4) and NPWT alone observing a mean wound area of 124.1cm² (SD = 92.2) (P = .280). Mean NPWT application time was 19.9 days (SD = 19.8).

The mean follow-up time was 13.7 (SD = 15.7) months and 14.1 months (SD 12.6) for STSG + NPWT and NPWT alone respectively. The mean No. of follow-up appointments were 3.07 (SD = 1.57) and 3.77 (SD = 3.19) for STSG + NPWT and NPWT alone, respectively.

There was a total of 3 recurrences, 2 in the NPWT alone group (4.5%) and 1 in the NPWT + STSG group (1.6%). Six donor site complications (9.8%) were noted in the NPWT + STSG cohort (Delayed healing—2, Scar hypertrophy—2, and Scar hyperpigmentation—2).

For the time to heal between the 2 closure methods, there was a statistically significant difference, with STSG + NPWT observing a mean time of 2.77 (SD = 1.25) months and NPWT alone observing a mean time of 4.40 months (SD = 2.38) (P = .0006).

Table 1 summarises the mean values for the difference between preoperative and postoperative outcome scores for both cohorts. Wilcoxon signed-rank tests did not show any significant difference between pre-op and post-op scores for both NPWT alone and STSG + NPWT. A score of zero indicates no difference between pre- and postoperative scores.

Table 1.

A Table Showing the Mean Difference Between Pre-Operative and Post-Operative PROMs for NPWT Alone and STSG + NPWT.

	Mean Difference Between pre-op and Post-op
	Closure Method
Outcome Measure	NPWT Alone	STSG + NPWT	P value
DLQI	5.15	6.04	.271
PHQ-9	20.7	19.8	.745
GAD-7	13.5	14.8	.265
PAINVAS2	11	0	1.000
BIPQ	10	5	1.000
DERMVAS	16.6	17.1	.968
Odour Severity	26.7	18.4	.744
Disease Severity	21.8	21.8	.689
Pain Severity	17.2	14.6	.514
Fatigue Severity	27.8	14.3	.102

Abbreviations: BIPQ, Brief Illness Perception Questionnaire; DERMVAS, Dermatology Visual Analogue Scale; DLQI, Dermatology Life Quality Index; GAD-7, Generalised Anxiety Disorder Assessment; NPWT, negative pressure wound therapy; PAINVAS2, Pain Visual Analogue Scale; PHQ-9, Patient Health Questionnaire Depression Scale; PROMS, patient-reported outcome measures; STSG, split-thickness skin graft.

There was no significant difference in follow-up time (P = .934) or the No. of postoperative appointments (P = .287) between groups.

Discussion

There is no difference in patient-reported outcomes with the addition of a STSG to NPWT after surgical excision of HS. The addition of a STSG does not appear to make a significant difference to patients when utilised for wide excision axillary defects.

There was also no significant difference in the No. of follow-up appointments required or follow-up time. The addition of an STSG did reduce the time to heal. However, STSG can result in donor site complications, which were observed in this study, and causes a secondary scar site.

Although the time to heal was significantly greater for patients treated with NPWT only, our results have shown that there was no significant difference in the No. of follow-up appointments attended. Follow-up with the surgical team and specialist nursing staff for dressing changes were included. The cost of follow-up appointments in a specialist care setting is often greater than that of primary care. ²⁹ It is therefore important to note that the longer period of healing resulting from not utilising an STSG did not appear to increase the No. of specialist follow-up appointments required. As this was a retrospective study, there was no standardised follow-up protocol. Requirement for dressing changes was determined by specialist nurses. Only tertiary centre follow-up appointments were recorded.

It is important to highlight that all patients were done as single, day-case procedures. This reduces disruption for the patient while also minimising costs and the potential complications associated with inpatient admission. More complex procedures involving local perforator flaps or free tissue transfer and multiple, staged procedures are more resource-intensive and carry a greater cost. ²¹ It is known that day-case surgery increases patient satisfaction, reduces the rates of hospital-acquired infection and venous thromboembolism rates with a lower risk of cancellation for the patient. ³⁰ Unlike other literature, the vast majority of patients only required a single procedure. ¹⁵ Due to their low-risk nature, along with the lack of requirement for complex reconstruction, these procedures do not necessarily have to be undertaken at a specialist referral centre. Resultantly, a much wider patient cohort can be offered surgical management of their disease.

One advantage to this surgical option is that there is a reduced temptation for the operator to minimise the defect size and therefore increase the potential of recurrence from surrounding areas. This phenomenon has previously been described concerning the surgical treatment of HS and recurrence rates with direct closure versus skin grafts or flaps. ⁶

Although there is a paucity of evidence in the current published literature, there is strong anecdotal evidence that surgical reduction of the disease burden in HS augments the effect of other systemic treatments. A multidisciplinary approach, as previously mentioned, is key in the management of HS and lesions must be quiescent at the time of operation.

Local flaps after wide surgical excision of HS are popular in the literature and have shown good results^{15,20,31–34}. However, while these techniques can bring healthy tissue coverage into these areas, there are several limitations. First, quiescent diseased tissue may be moved into the defect, resulting in recurrence at a later point. This is supported by higher recurrence rates with local flap surgery shown in the literature. ¹³ Second, these flap options leave patients with scars that extend outside of the original area of disease. Patients can find it difficult to hide these scars when wearing particular clothing. Wide excision and application of NPWT does not require specific microsurgical or reconstructive skills and therefore can be performed by a wider group of surgeons.

Relative immobilisation of the shoulder may be required in patients undergoing STSG to reduce shear forces. The addition of NPWT achieves this to a degree, however, NPWT alone, without STSG can allow earlier shoulder mobilisation without any risk to graft take.

In our experience of performing these procedures, we have found anecdotally that patients are pleased with either of the treatment options and those with bilateral axillary involvement commonly request the procedure on the contralateral side.

While this is a small sample of patients in a single hospital setting, this is one of the UK's central hospitals and therefore is more likely to encompass a more representative sample regarding ethnicity and age demographics. The female:male ratio was also similar to that of ratios reported in the literature. ³⁵ As the effect sizes are small, further prospective research, with larger data sets is required to determine if this data is reproducible.

The results highlight a defined small dataset. For this sample, the DLQI, PHQ-9, and Fatigue Severity conclusions drawn are most likely to be representative of the entire patient population for HS. Given the small sample size, the results for PAINVAS2, BIPQ, DERMVAS and Pain Severity may be less robust.

In a wider context, the complex interaction between pathophysiological and psychosocial factors means that patients with HS are best managed in a multidisciplinary team environment where they can receive holistic treatment. Given the nature of the healthcare setting more widely, as well as the incidence of the condition, this can be generalised to the other hospitals in the UK.

As this was a retrospective observational study, there are several weaknesses. First, there was no set time at which the follow-up IMPARTS surveys were completed by patients. This had the potential to affect some of the patient-reported measures. If these were completed close to the surgical date, changes may be less representative of overall results. Conversely, if they were completed further from the surgical date, the responses would be subject to response shift bias. Only IMPARTS information completed between 6 weeks and 6 months after the procedure was used to try and reduce this effect. IMPARTS data was not available for all patients; however, they were still included to provide a representative outlook of the patient demographics. Patients were asked to complete the IMPARTS questionnaire as part of their routine clinic visit. No incentive was provided for completion and the questionnaires were only available at the clinic visit, with patients unable to complete it from home. These factors will likely have limited the response rate and explain the missing IMPARTS data in this study. Missing data was primarily from post-operative visits. Additionally, this may introduce non-response bias, which should be considered when interpreting the study results. Data capture did not allow for the recording of primary care interventions which may have influenced the No. of follow-up appointments.

This study focused on those undergoing procedures for axillary disease. Although the disease process is uniform throughout commonly affected sites, the results shown here may not be transferable to HS in other areas. Although the recurrence rate was low in our observed cohort, it is important to state that follow-up times were relatively short compared with other studies. ¹³ The small sample size may also have resulted in a type-II error which could be addressed with larger, prospective studies in the future.

The PROMs used in this study, although validated, may not necessarily be best placed to assess surgical outcomes. For example, the PHQ-9 and GAD-7 were developed to assess for depression^36–38 and generalised anxiety ³⁷ which, although potentially influenced by disease burden, may not be able to detect meaningful differences in single-site surgical procedures. This is also true for the DLQI, which was developed with patients mainly suffering from more widespread dermatological conditions such as acne, psoriasis, and eczema. ³⁹ Many of the patients included in this study had HS at multiple body sites and therefore any changes due to surgery in one area may be masked by those of other sites. Although improvements in patient wellbeing should be the aim of these procedures, the use of a more specific method to assess surgical outcomes may be useful. The prospective use of a specific scar scoring scale, such as the SCAR-Q, ⁴⁰ may produce more meaningful comparisons for differences in the surgical management of HS.

Conclusion

Multiple reconstructive options are available for patients with severe axillary HS, many of which produce excellent outcomes. This retrospective study has shown that in patients with axillary HS requiring surgical excision, the use of NPWT alone is not inferior to NPWT and STSG for patient-reported outcomes. This is an effective treatment and can be performed as a day case.