A Retrospective Study of Pilonidal Sinus Healing by Secondary Intention Using Negative Pressure Wound Therapy Versus Alginate or Gauze Dressings
Pilonidal sinus (PS) disease is an inflammatory skin and subcutaneous tissue condition that presents with infection, acute abscess, chronic discharging wounds, and/or pain. Surgery with open healing by secondary intention typically is used to achieve the fastest healing time with minimal recurrence rates. A retrospective analysis was conducted of data extracted from the medical records of 73 consecutive patients who had symptomatic natal cleft PS over a 10-year period to compare use of NPWT to alginate-based/gauze daily dressing (DD) changes in terms of healing time and recurrence.
Variables extracted included age, gender, PS wound diameter (small <1 cm, medium 1 cm to 3 cm, large >3 cm), and time in weeks to achieving the endpoint (epithelialization). Risk factors examined that can affect healing or recurrence of previously operated PS disease included initial drainage before excision and risk factors for impaired healing (morbid obesity as determined by body mass index [BMI] ≥35, chronic infective skin conditions, and ongoing therapy with immuno-modulating drugs or chemotherapy), and loss to follow-up. Data were collected and analyzed using the chi-squared statistic, Kaplan-Meier curves, and Cox regression models. The total time of follow-up was 390 weeks for the DD group and 311 weeks for NPWT group. Patient mean age was 26.5 ± 10.7 years, most (53, 72.6%) were male, and 12 (16.4%) had comorbidities potentially affecting healing. Nine (9) were treated with primary closure and 62 patients were treated with open healing by secondary intention (2 additional patients receiving DD were excluded from the analysis because they had small sinuses that made NPWT unfeasible). Among participants, 30 (48%) received DD and 32 had NPWT. The median time to healing was 10 weeks (95% CI: 7-17) in the DD group and 8 weeks (95% CI: 7-9) in the NPWT group (not significantly different). In patients who healed, the average time to healing was 15.0 ± 18.1 and 9.8 ± 6.3 weeks in the DD and NPWT groups, respectively (not significantly different). The PS wound recurred in 5 patients — 4 (12.5%) in the DD group and 1 (3.1%) in the NPWT group (P = .355). In univariate analysis, only the presence of comorbidities was found to significantly affect time to healing (HR 95%, CI: 0.40 [0.17-0.93]; P = .033]. Prospective, randomized controlled clinical studies are warranted.
Pilonidal sinus (PS) disease is an inflammatory skin and subcutaneous tissue condition that arises in the natal cleft within hair found in sinus tracts in subcutaneous tissue. The disease presents with infection, acute abscess, chronic discharging wounds, and/or pain.1,2 Its etiology has been broadly debated in surgical analysis and teaching as to whether the disease lies primarily in inflamed hair follicles3 due to an infected vestigial remnant or is acquired from hair penetrating through the skin to subcutaneous tissue.1
Definitive treatment usually requires either surgical excision and healing by secondary intention or primary excision with closure by one of several methods, including midline closure or lateral closure using flaps.4-6 The aim of surgical treatment is to achieve the fastest healing time with minimal recurrence. A Cochrane systematic review7 of surgery for PS disease that examined the endpoint of recurrence and healing times found recurrence of PS was less likely after open wound healing (44/828 patients [5.3%]) than after primary closure (73/838 patients [8.7%]). However, in 9 of 13 studies examined, faster healing times were achieved from primary closure. In addition, the authors observed the majority of the data were limited in quality and based on case reviews or comparative studies.
Sondenaa et al8 conducted a randomized trial with 2 respective groups of 60 patients: 1 group was treated with excision of PS with primary closure and the other with excision and healing by secondary intention using daily dressing (DD) changes using gauze after Sitz baths. The study found the PS group experienced an 8% recurrence rate (5/60 patients) after primary closure was performed compared to the secondary intention group’s recurrence rate of 5% (3/60 patients) (P = .49).
Negative pressure wound therapy (NPWT). NPWT functions by providing a sealed dressing to which suction is applied with a portable or stand-alone pump to draw tissue fluid into a canister and improve circulation into the wound.9 In a prospective human trial,9 skin perfusion was increased 3- to 5-fold when NPWT was applied to intact forearm skin and compared to baseline blood flow (P <.001). A Cochrane review10 conducted to assess the effects of NPWT on healing a wide variety of general surgical wounds by secondary intention found NPWT facilitates the development of granulation tissue. Advocates of NPWT propose open wound healing time is significantly reduced and NPWT provides more comfort for the patient.2,10
The previously mentioned Cochrane Review10 concluded the number of trials and patient numbers are lacking and no rigorous, randomized, controlled trial offered evidence of the clinical effectiveness of NPWT. However, faster healing was achieved in a multicenter, randomized clinical trial11 among patients with diabetes with forefoot amputations when NPWT (77 patients) was compared to DD changes using “moist wound dressings with alginates, hydrocolloids, foams, or hydrogels” (85 patients). Forty-three (43, 56%) of the NPWT patients completed healing versus 33 (39%) in the DD group (P = .040), the rate of granulation tissue formation was faster in the NPWT group than in the DD group (P = .002), and wound closure was achieved faster with NPWT than DD (P = .005).
The literature on PS surgical treatment includes 10 case reports of NPWT use for post op wound care in very complex and large PS wounds,12-15 where better-than-expected results were achieved when compared with routine DD. Farrell and Murphy16 reviewed these 10 case reports involving the use of NPWT in PS disease and found 5 to indicate NPWT is emerging as a preferred treatment for pilonidal disease management.
Two (2) prospective studies address NPWT therapy in PS postsurgical wounds.2,17 Biter et al17 randomized 49 patients to receive NWPT (24 patients) or DD (25 patients). The use of NPWT was limited to 2 weeks, after which patients were changed to DD described as “porous silicone wound dressing, with instructions to rinse the wound 3 times daily.” The authors found no difference in time to complete wound healing and time to resume daily life activities, although a statistically significant positive effect on wound size reduction was observed in the first 2 weeks (the mean measured wound size was 0.3 cm in NPWT versus 0.57 cm in DD [P = .02]). Banasiewicz et al2 randomized 19 patients to receive either NPWT (n = 10) or DD (“conventional absorbing dressings”; n = 9) and found follow-up time was reduced (11.8 versus 30.3 days), return to normal activity or “time to healing” was faster (7.3 versus 15.9 days), and less pain was noted after the fourth day postoperatively in the NPWT group versus the DD group, although these differences were not statistically significant.
The purpose of this retrospective study was to compare NPWT and DD (daily alginate or gauze) with respect to the time taken to reach the endpoint (ie, complete epithelialization) and whether the PS recurred (failure to heal).
The study retrospectively analyzed the records of all of the 73 consecutive PS patients treated over a 10-year period (2006 – 2015) through a private urban surgical practice. Cases of PS surgery were identified by the Australian Government Medicare Benefits Schedule coding number (30676) in the patient electronic record database of the surgical practice.18
Data were de-identified using coding in the study database to protect patient privacy; patient records were kept in accordance with the patient confidentiality and privacy laws of Victoria and Australia. The research was conducted according to the Principles of the Declaration of Helsinki.19
Variables extracted for the study were entered into an Excel spreadsheet and included age, gender, PS wound diameter (small <1 cm, medium 1–3 cm, large >3 cm) measured as the largest dimension reported by the histopathologists from the formalin-fixed specimens, time in weeks to achieving endpoint (complete epithelialization as recorded by the surgeon), and failure to heal/early recurrence.
Secondary data collected included patients presenting with recurrence of previously operated PS disease or patients presenting with PS for the first time, initial drainage before excision, risk factors for impaired healing (eg, morbid obesity as determined by body mass index [BMI] ≥35, other diseases affecting healing rates such as infective skin conditions and ongoing therapy with immuno-modulating drugs or chemotherapy, and type 1 diabetes20), and patients who became lost to follow-up (did not attend [DNA] follow-up appointments).
All patients underwent surgical excision of their sinus and were nonrandomly allocated to either NPWT or DD for assessment of postoperative wound management. In the first 3 years, only DD were used; when NPWT was introduced, the larger wounds were selected preferentially for NPWT, a practice occasionally modified for some patients because of their inability to afford NPWT treatment due to lack of health fund coverage. Of the 32 NPWT patients, 29 had ActiV.A.C.™ or V.A.C.Via™ Dressings (KCI, Acelity, San Antonio, TX) and 3 had Renasys GO™ Dressings (Smith & Nephew, London, UK).
Patients who failed to heal were recorded as having an early recurrence. Patients were reviewed until the endpoint was achieved, DNA, or failure to heal (early recurrence of between 1 and 80 weeks). Late recurrences were recorded after the endpoint was achieved only if patients returned with further PS development.
Surgical procedure. All patients received a general anesthetic. Positioning was usually left lateral, but sometimes the prone position; the lateral position facilitates a less complex anesthetic without compromising surgical access. The excisional technique involved a delineation of the sinus tracts with blue ink instillation through the primary sinuses. In each case, complete surgical incision of the PS complex (the combination of primary sinuses, associated infected cavity and tracts, and secondary sinuses) then was performed with diathermy excising outside the blue ink-stained tissue. Long-acting local anesthetic was infiltrated. Specimens were submitted to histopathology in formalin.
Postsurgical procedure. The choice of primary closure, NPWT, or DD techniques was determined by several factors. Although healing by secondary intention was regarded as the preferred method of treatment for any PS of significant size (>1 cm) or complexity, excision and primary closure was used for a small number of patients with small (< 1 cm) PS wounds. As stated previously, open wound treatment was provided using DD only for the first 3 years of the study — principally, alginate-based packing postoperatively and gauze, changed daily. From the fourth year of data collection, NPWT was introduced as an alternative postoperative option; large wounds (>3 cm in longest dimension), patients with recurrent disease from prior surgery, and patients with comorbid diseases linked to poor healing were selected for NPWT if possible. Circumstantial factors such as treatment in a government hospital or lack of appropriate community nursing availability excluded NPWT as an option. The resulting NPWT and DD groups comprised a mix of sinus wound sizes, recurrent disease, and patients with comorbidities.
For NPWT dressing application, the wound cavity was filled with black foam dressing and an airtight adhesive dressing was applied and connected to a unit providing continuous negative pressure of 125 mm Hg. To avoid skin irritation, DuoDERM™ strips (ConvaTec, Greensboro, NC) or a similar product were applied around the wound edges and underneath foam inserts where the suction line tracks away from the wound. To achieve an airtight seal in the natal cleft, Hollister Adapt™ strips (Hollister, Libertyville, IL) or a similar product were used to fill the natal cleft before applying the adhesive plastic dressing. NPWT dressings were changed every 3 to 4 days until granulation was level with the skin and then alginate or gauze dressings were used.
Postoperatively, the patients were followed by the surgeon every 2 weeks. The time to healing was quantified in weeks from the date of surgery. The endpoint (epithelialization) was determined by the surgeon during follow-up. The DNA patients were examined as a subgroup.
Data collection and analysis. Data were extracted, de-identified, and transferred to an Excel Spreadsheet by the principal author. Demographic, wound treatment, and outcome data were descriptively analyzed. Categorical variables were reported as frequencies and their percentages and compared between treatment groups using the chi-squared statistic. Kaplan-Meier curves together with the average treatment effect (ATE) were used to compare the time to the endpoint in the 2 treatment groups. Cox regression models were run to assess the probability of achieving the endpoint.21 The inverse-probability-weighted regression adjustment was used to estimate the ATE. The Cox model and ATE were adjusted for age, gender, prior surgery, sinus size, obesity, and comorbid diseases. Stata Statistical Software version 14.1 (StataCorp, College Station, TX) was used for analyses of healing times and rates of early recurrence, comparing the 2 groups (NPWT or DD). An alpha of 0.05 was used for level of statistical significance.
Patient characteristics: A total of 73 patients were included in the study: 53 (72.6%) men, 20 (27.4%) women; mean age 26.5 ± 10.7 years, range 12–69 years; 25 with a past history of surgery for PS at other hospitals. Definitive surgery as either excision with primary closure or excision with healing by secondary intention had been attempted for 14 patients (19.2%); another 11 patients (15%) had only drainage procedures and no excisional surgery. These 25 patients had all re-presented some years after these procedures into the current patient group. Twelve (12, 16.4%) patients had comorbidities potentially affecting tissue healing: 9 were morbidly obese, 1 was receiving chemotherapy for a germ cell tumor, 1 had multiple sclerosis and was being treated with immuno-modulator drugs, and 1 had an extensive and recurrent condition of folliculitis and hidradenitis. Other predisposing factors to infective processes such as insulin-dependent type 1 diabetics did not feature in this group of patients. All patients in this study cohort had experienced recurrent episodes of infection requiring antibiotic treatment on at least 1 occasion, and 11 patients (15%) had prior incision and drainage (see Table 1). Thirteen (13) patients (21%) had recurrent PS disease following prior excisional surgery.
The 9 patients with small wounds (<1 cm in diameter) were treated with primary closure; these wounds healed between 2 and 6 weeks post op and were not analyzed further. Two (2) additional patients receiving DD were excluded from the analysis because they had small sinuses that made NPWT unfeasible.
Of the remaining 62 patients, 30 were treated using DD (48%) and 32 were treated using NPWT (see Table 1). DDs were changed daily and NPWT dressings were changed every 3 to 4 days. With the exception of presence of comorbidities (P = .021), the characteristics of patients in the 2 treatment groups were statistically similar at the time of treatment.
The total time of follow-up was 390 weeks for the DD group and 311 weeks for the NPWT group. Two (2) patients (1 from each treatment group) were missing all information at follow-up time. Among patients receiving NPWT, 29 (90.6%) were followed until they reached the endpoint, compared with 21 (70.0%) of those receiving DD. Patients lost to follow-up (DNA) included 3 out of 32 (9.4%) NPWT patients and 9 out of 30 (30%) DD patients. These patients are depicted as censored (ie, dots) in Figure 1.
Time to healing. The median time to healing in the group receiving DD was 10 weeks (95% CI: 7-17) compared with 8 weeks (95% CI: 7-9) in the group receiving NPWT. The log-rank test indicated the 2 curves were not significantly different (P = .116). Average time among patients followed-up until healing was 15.0 ± 18.1SD weeks in the DD treatment group (n = 21) and 9.8 ± 6.3SD in the NPWT group (n = 29). Using inverse-probability-weighted-regression adjustment methods to assume average treatment effect among patients followed-up until healing, the results showed if it was assumed that every patient received NPWT, the estimated average time to healing would be 4.8 weeks less than the average of 13.1 weeks that would be observed if everyone had received DD, a difference that failed to achieve statistical significance (ATE = -4.79; 95% CI: -10.39-0.82; P = .094).
Probability of healing. The adjusted hazard ratio (probability) of healing in the group of patients receiving NPWT was 1.21 times higher than in the group receiving DD and not statistically significant (see Table 2).
Comorbidities. Although the number of patients with comorbidities was small (n= 12; 10 in DD and 2 in NPWT group), unadjusted cox-regression suggested the hazard of healing was significantly lower among patients with comorbidities (P = .033). However, this result disappeared after adjusting for other patient characteristics (see Table 2). Additional subgroup analyses indicated the treatment effect among patients with comorbidities was different than that observed in persons without comorbidities (P = .011). The hazard ratio among patients with comorbidities was 13.10 (95% CI 3.90-44.04; P <.001). No statistically significant differences were observed for different sinus sizes or whether patients had undergone prior surgery.
Probability of recurrence. The PS wound recurred in 5 patients, 4 (12.5%) in the DD group and 1 (3.1%) in the NPWT group (P = .355). The follow-up time for early recurrence patients ranged from 25 to 80 weeks. Two (2) patients returned for treatment with late recurrence at 2.5 years post endpoint.
PS disease causes considerable discomfort and local morbidity for patients due to infective complications. In this study featuring a cohort of consecutive patients in a single surgeon’s practice over 10 years, the prime mode of treatment selected was excision with healing by secondary intention (open wound) for wounds >1 cm diameter. The results from this study suggested clinical outcomes were better for patients in the NPWT than in the DD group, although the differences were not statistically significant. These benefits were indicated by the median time to healing, which was 2 weeks shorter for the NPWT group than the DD group, and ATE was 5 weeks less in the NPWT group. When stratified by patient characteristics, subgroup treatment effect was similar for all characteristics except comorbidity subgroups, where Cox regression among patients with comorbidities indicated healing was 13 times better among patients in the NPWT group compared with the DD group.
The adjusted hazard probability of healing with NPWT was 1.21 times greater than with DD. This observation is similar to those of Biter et al17 and Banasiewicz et al2 who documented NPWT benefits over DD among randomized patients with PS excisional wounds. Biter et al17 reported a positive effect on wound size reduction in the first 2 weeks with NPWT in 24 of 49 patients, compared to the 25 with DD. Banasewicz et al2 reported 10 patients with NPWT had better time to wound healing and recovery and less pain than the 9 patients with DD.
The results also showed healing is 13 times more likely with NPWT in patients with comorbidities. This supports the observation by Farrell and Murphy16 in their analysis of complex case studies in which NPWT was used that healing was achieved with NPWT when it had not expected using other routine dressings.
Where wounds had not healed, close to 20% of patients were DNA; 10 were in the DD treatment group. These data may reflect the difficulty of treating this patient group with this treatment regimen or the potential value of a dressing regimen that requires follow-up care by a health care professional. The 13 patients whose wounds had not healed were followed-up for 0 to 28 weeks before failing to attend. NPWT dressing management warrants careful postoperative attention by nurses trained in PS wound care and by the managing surgeon. Repeated follow-up visits allow nurses and surgeons to provide minor interventions in wound management that may affect results, such as silver nitrate or iodide application or the use of oral antibiotics. Although both NPWT and DD are noted to produce some patient discomfort,2,17 patients in the NPWT group were more likely to comply with follow-up than persons in the DD group. If patients removed themselves from follow-up (DNA), the outcome was unknown.
The authors of the current study observed that for NPWT to have the maximum effect, it must remain in place and that NPWT dressing changes continue every 3 to 4 days until granulation has almost completely filled the wound cavity (between 3 to 10 weeks in this study). If removed before granulation was level with the skin, the benefit of NPWT could be reduced.17 Biter et al17 compared 24 patients randomized to NPWT left in place for 2 weeks to 25 patients using porous silicone dressings that were changed daily and irrigated. NPWT was stopped after 2 weeks when a higher wound healing rate for NPWT was noted (ie, wound size was reduced 0.3 to 0.57 cm, P = .02) However, little difference was noted in the time to complete wound healing (84 versus 93 days, respectively; P = .44).
As wound healing progresses into the later phases, the surgeon and experienced wound care nurse need to perform ongoing assessment. Although no relevant scientific studies confirm their observations, the authors have found hair may enter the unhealed wound (this needs regular shaving) and skin bridging that can lead to a recurrent sinus can occur. In addition, specific bacterial infections may require courses of antibiotics; commonly, topical silver nitrate may be used to treat low-grade infections that often inhibit the later stages of healing.
Evidence7,8 supports that healing PS wounds by secondary intention leads to less recurrence than primary closure. Primary closure has been reported to have faster healing times than DD with open wounds.7 The potential of faster healing and granulation tissue formation with NPWT2,7,9,11 compared to DD changes and of lower recurrence rates2,7 is reinforced by the identification in this study of a significant benefit from NPWT in the small group of patients with comorbid disease.
A prospective, randomized study of NPWT versus DD in open wound healing is needed and might include a third arm (ie, primary surgical closure). A prospective study could be used to analyze outcomes that might be achieved by using different (nonNPWT) dressings that do not require daily dressing changes; this study would ideally incorporate at least 2 years of postoperative follow-up to assess for late recurrences. In addition, the implications of comorbid diseases (including morbid obesity) in healing also should be examined in future research.
The authors recognize the limitations of this retrospective study and the relatively small numbers reported in patient subsets, including the comorbidities group. Appropriate statistical methods were applied to accommodate these limitations. The study also employed one standardized method of surgical technique for all patients, administered as described by one surgeon.
In this study, the PS wounds of 21 of 29 (72.4%) patients in the DD group and 29 of 31 (93.6%) patients in the NPWT group healed. The median time to healing PS wounds in the group receiving DD was 10 weeks (95% CI: 7-17) compared with 8 weeks (95% CI: 7-9) in the group receiving NPWT. Among patients who healed, the difference in average time to healing was 5.2 weeks. However, these differences were not statistically significant. Although the comorbidities subset was small, a significant difference in healing was noted in the NPWT group compared with the DD group (hazard ratio 13.1, 95% CI). These encouraging results reinforce the authors’ clinical observations that NPWT confers advantages to PS wound healing. A prospective study with a large cohort of patients is warranted. n
1. Goligher JC. Surgery of the Anus, Rectum and Colon, 2nd ed. London, UK: Bailliere, Tindall and Cassell Ltd;1967:254–260.
2. Banasiewicz T, Bobkiewicz A, Borejsza-Wysocki M, et al. Portable VAC therapy can improve the results of the treatment of the pilonidal sinus — a randomised prospective study. Pol Przegl Chir. 2013;85(7):371–376.
3. Bascom J. Pilonidal sinus disease. In: Fazio V. Current Therapy in Colon and Rectal Surgery. Toronto, Canada: B.C. Decker Inc;1990:32.
4. Horwood J, Hanratty D, Chandran P, Billings P. Primary closure or rhomboid excision and Limberg flap for the management of primary sacrococcygeal pilonidal disease? A meta-analysis of randomised controlled trials. Colorectal Dis. 2012;14(2):143–151.
5. Ertan T, Koc M, Gocmen E, Aslar AK, Keskek M, Kilic M. Does technique alter quality of life after pilonidal sinus surgery? Am J Surg. 2005;190(3):388–392.
6. Enriquez-Navascues J, Emperanza J, Alkorta M, Placer C. Meta-analysis of randomised controlled trials comparing different techniques with primary closure for chronic pilonidal sinus. Tech Coloproctol. 2014;18(10):863–872.
7. Al-Khamis A, McCallum I, King PM, Bruce J. Healing by primary versus secondary intention after surgical treatment for pilonidal sinus. Cochrane Database Syst Rev. 2010;20(1):CD006213. doi: 10.1002/14651858.CD006213.
8. Sondenaa K, Nesvik I, Anderson E, Soreide JA. Recurrent pilonidal sinus after excision with closed or open treatment: final result of a randomised trial. Eur J Surg. 1996;162:237-240.
9. Timmers MS, Le Cessie S, Banwell P, Jukema GN. The effects of varying degrees of pressure delivered by negative pressure wound therapy on skin perfusion. Ann Plast Surg. 2005;55(6):665–671.
10. Dumville JC, Owens GL, Crosbie EJ, Peinemann F, Liu Z. Negative pressure wound therapy for treating surgical wounds healing by secondary intention. Cochrane Database Sys Rev. 2015;6:CD011278. doi:10.1002/14651858.CD011278.pub2.
11. Armstrong DG, Lavery LA; Diabetic Foot Consortium. Negative pressure wound therapy after partial diabetic foot amputation: a multicentre, randomised controlled trial. Lancet. 2005;366(9498):1704–1710.
12. Duxbury MS, Finlay IG, Butcher M, Lambert AW. Use of a vacuum-assisted closure device in pilonidal disease. J Wound Care. 2003;12(9):355.
13. McGuinness JG, Winter DC, O’Connell PR. Vacuum-assisted closure of a complex pilonidal sinus. Dis Colon Rectum. 2003;46(2):274–276.
14. Lynch JB, Laing AJ, Regan PJ. Vacuum-assisted closure therapy: a new treatment option for recurrent pilonidal sinus disease. Report of three cases. Dis Colon Rectum. 2004;47(6):929–932.
15. Bendewald FP, Cima RR, Metcalf DR, Hassan I. Using negative pressure wound therapy following surgery for complex pilonidal disease: a case series. Ostomy Wound Manage. 2007;53(5):40–46.
16. Farrell D, Murphy S. Negative pressure wound therapy for recurrent pilonidal disease: a review of the literature. J Wound Osteomy Continence Nurs. 2011;38(4):373–378.
17. Biter LU, Beck GM, Mannaerts GH, Stok MM, van der Ham AC, Grotenhuis BA. The use of negative-pressure wound therapy in pilonidal sinus disease: a randomised controlled trial comparing negative-pressure wound therapy versus standard open wound care after surgical excision. Dis Colon Rectum. 2014;57(12):1406–1411.
18. Medical Benefits Schedule Book. Australian Government Department of Health. Available at: www.health.gov.au/internet/mbsonline/publishing.nsf/Content. Accessed March 1, 2017.
19. World Medical Association. Declaration of Helsinki; Ethical principles for medical research involving human subjects. JAMA. 2013;310(20);2191–2194.
20. Committee on Pre and Post Operative Care, American College of Surgeons. Evaluation of operative risk. In: Committee on Pre and Post Operative Care, American College of Surgeons. Elective Care, Vol. 12. 1989.
21. Hosmer DW, Lemeshow S, May S. Applied Survival Analysis: Regression Modelling of Time – Two - Event Data, 2nd ed. New York, NY: Wiley;2008.
Potential Conflicts of Interest: none disclosed
Dr. Julia Danne is a surgical registrar, St. Vincent’s Hospital Melbourne, Melbourne, Victoria, Australia. Dr. Gwini and Dr. McKenzie are biostatisticians; and Dr. Peter Danne is a general/trauma surgeon at Epworth Healthcare, Richmond, Australia. Please address correspondence to: Dr. Julia L. Danne, Suite 8, Level 8, The Epworth Centre, 32 Erin Street, Richmond, Victoria 3121 Australia; email: email@example.com.