hand topic



PUBLISHED: 30-09-2020

AJOPS logo

Different collagenase delivery for Dupuytren’s disease in public hospitals

Jessica A Paynter,1–3 Vicky Tobin PhD,1 James CS Leong MBBS FRACS MS,3,4 Warren M Rozen MBBS PhD FRACS,1,2,4 David J Hunter-Smith MBBS MPH FRACS1,2,4


Peninsula Health


The Alfred Centre


Monash Health


Monash Medical Centre

Name: David J Hunter-Smith
Address: Frankston Hospital
Peninsula Health
2 Hastings Road
Victoria 3199
Email: [email protected]
Phone: +61 3 9784 8416
Citation: Paynter JA, Tobin V, Leong JCS, Rozen WM, Hunter-Smith DJ. Different collagenase delivery for Dupuytren’s disease in public hospitals. Australas J Plast Surg. 2020;3(2):22–31.
Manuscript received: 30 August 2019
Manuscript revised: 20 December 2019
Accepted for publication: 2 June 2020
Copyright © 2020. Authors retain their copyright in the article. This is an open access article distributed under the Creative Commons Attribution 4.0 Licence which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.|
Section: Hand


Background: The delivery protocol of collagenase Clostridium histolyticum (collagenase) injection for Dupuytren’s disease is variable, due to limited evidence for any one approach and widespread ‘off-label’ delivery occurring in Australia. As such, this preliminary study aimed to assess whether different collagenase delivery protocols for treating Dupuytren’s disease have an impact on effectiveness and safety. It was hypothesised that different collagenase delivery would affect outcomes.

Methods: This preliminary, prospective study included a consecutive cohort of adult patients with Dupuytren’s disease being treated with collagenase within two Australian public hospitals to determine whether different collagenase delivery protocols impact on effectiveness and safety. The therapeutic effect was measured objectively using the total passive extension deficit (TPED), clinical success and clinical improvement. Three patient-reported outcome measures (PROMs) were used: Unité Rhumatologique des Affections de la Main (URAM), the Southampton Dupuytren’s Scoring Scheme and the Canadian Occupational Patient-Specific Functional Scale (PSFS).

Results: The delivery of collagenase was variable at both clinics. The number of patients treated with collagenase at Institute I and Institute II was 49 and 18, respectively. Clinical success was achieved in 42 per cent of the Institute I and 35 per cent of the Institute II cohort. A statistically significant reduction in all three PROMs was observed for both cohorts. No significant differences between effectiveness (clinical success, clinical improvement, TPED, PROMs) or safety was found when comparing the two cohorts.

Conclusion: The delivery of collagenase was variable at Institutes I and II, but these differences did not appear to impact the effectiveness or safety of collagenase delivery.

Keywords: microbial collagenase, dupuytren contracture, collagenases, treatment outcome, patient reported outcome measures


Dupuytren’s disease is a common, benign fibrotic disease resulting in finger flexion contractures.1 It can lead to significant functional hand impairment,2 with patients therefore seeking treatment.3 Treatment focuses on four characteristics identified by Tubiana in 1975: ‘to correct deformity, avoid complications, shorten postoperative recovery, and prophylactically prevent recurrences.’4

Surgical correction of Dupuytren’s disease via fasciotomy, fasciectomy and dermofasciectomy has been the mainstay of therapy over recent decades.5,6 Minimally invasive techniques include percutaneous needle aponeurotomy and collagenase injections. Percutaneous needle aponeurotomy is advantageous due to rapid recovery times and the ability for outpatient management.7 However, recurrence rates are higher compared with fasciectomy and dermofasciectomy.8 These two surgical options are associated with a lengthier recovery time and a lower recurrence rate but an increased complication profile.5

In 2013, collagenase was licensed by the Therapeutic Goods Administration in Australia for the treatment of Dupuytren’s disease. The drug is not Medicare indexed, limiting its accessibility for patients.

The aim of this study was to assess differences in effectiveness and safety that may arise due to different collagenase delivery protocols. It was hypothesised that different protocols would have an impact on the effectiveness and safety of collagenase delivery.


A preliminary, prospective study occurred from February to October 2018. It was undertaken at two tertiary hospitals in Australia. Institute II commenced collagenase treatment in 2014 and Institute I in 2015.

Patients from each clinic over 18 years of age who were receiving collagenase treatment for their Dupuytren’s disease were invited to join the study. Exclusion criteria included pregnancy, breastfeeding or chronic neuromuscular disorders. Demographic data, Dupuytren’s disease history and medical history were collected for all participants.

Objective measurements of the Dupuytren’s disease contracture pre- and post-collagenase treatment were undertaken by measuring the degree of finger contracture (passive extension deficit) using a goniometer. The change in contracture was then used to determine the clinical success and clinical improvement as defined in the landmark CORD (Collagenase Option for the Reduction of Dupuytren’s) I study by Hurst, Badalamente and colleagues. Clinical success is defined as a reduction in primary joint contracture to 0–5 degrees of full extension 30 days after the last collagenase injection. It is reported as the percentage of study participants who achieved the 0–5 degree endpoint. Clinical improvement is defined as a more than 50 per cent reduction from baseline contracture 30 days after the last injection.2 The outcome was recorded as the percentage of participants achieving 50 per cent reduction between the cohorts from the two clinics.2

Three patient-reported outcome measures (PROMs) were employed to assess patients’ perceptions of the effectiveness of collagenase treatment. These were the Unité Rhumatologique des Affections de la Main (URAM), the Southampton Dupuytren’s Scoring Scheme and the Canadian Occupational Patient-Specific Functional Scale (PSFS).9–11 This was used to assess each person’s individual concerns. The PSFS scores five activity-related themes.11 This was then used to quantify activity limitations and measure functional outcomes for patients.12 This was abbreviated to three activity limitations due to clinic time constraints.

The collagenase treatment received by each patient was recorded by documenting the collagenase dose and the location of injection using a hand map. Manipulation was recorded by documenting success (failure, partial success, complete success) and any complications on a manipulation form.

Data were then analysed using STATA version 15 (StataCorp, College Station, TX).13 For each clinic, and subsequent data set, continuous data were assessed using the Shapiro-Wilk normality test, and normally distributed and unpaired data were assessed using Student’s t-test. Categorical or interval data expressed as a percentage were compared between the two cohorts using Fisher’s exact test.

Data were assessed for each individual digit and as a percentage of all digits. Furthermore, when clinical success or clinical improvement was detected, the clinical improvement values were used to calculate the median and range of clinical success and clinical improvement. This was repeated for each individual joint injected with collagenase (metacarpophalangeal joint, proximal interphalangeal joint, distal interphalangeal joint). The percentage of digits that achieved clinical success and clinical improvement within the cohorts was compared using Fisher’s exact test.

A Shapiro-Wilk test was implemented to assess for normal distribution of the three PROMs and TPED (continuous, paired data). The data were found to be non-normally distributed. Continuous non-normal data within each clinic that was paired (TPED or PROMs—before versus after collagenase injection) were compared using the Wilcoxon signed rank test. Comparison between the two clinics for pre- and post-collagenase outcomes (TPED and PROMs) was undertaken after generating a new variable (eg, TPED difference = TPED pre-collagenase—TPED post-collagenase). This was performed to simplify the data structure and enable a non-parametric equivalent of a t-test (Mann–Whitney U-test) to be used.

Adverse effects

Adverse effects were recorded by asking or observing patients for any side effects that they had experienced. These were recorded as a percentage of the participant cohort that developed the side effects. Complications were analysed as a total sum across all digits and joints. Complications were classified as treatment-related and serious treatment-related adverse effects.

A p-value less than 0.05 was deemed statistically significant.


Study population

At Institute I and Institute II, 49 and 18 participants, respectively, were recorded at initial visit, manipulation and first review (six weeks post-treatment).

Clinic design comparison

Table 1 shows the key constituents of each clinic design. This is a qualitative tabulation of how collagenase is delivered at each clinic.

Table 1: Hospital comparison of key collagenase Clostridium histolyticum delivery design features:
Institute I vs Institute II
Institute I Institute II
Collagenase criteria
Dupuytren's disease severity High severity High severity
High severity High severity
Number of cords One Multiple
Recurrent cords Yes Yes
Patient autonomy Yes Yes
Recent stroke Yes Yes
Pregnancy/breastfeeding No No
Bleeding disorder Yes Yes 
Chronic neuromuscular condition Yes Yes 
Anticoagulant use Yes Yes 
Tetracycline use Yes Yes
Injection protocol
Frequency Weekly Quarterly
No. of patients 4 20–30
Location  Clinic Clinic
Anaesthetic (Y or N) Yes No
Type of anaesthetic Lidocaine (lignocaine) 1%–2%  Not applicable
Anaesthetic volume 5–20mL Not applicable
Anaesthetic location Palm/ring block Not applicable
Powdered drug amount 0.9 0.9
Diluent volume (mL) 0.5 0.39
Solution volume (mL) 0.5 0.25
Dose per injection 0.9 0.58
Number of injections Single Multiple
Bilateral injections No Yes
Injection direction (at skin) Oblique = 45° Perpendicular = 90°
Multiple mini injections Yes Yes
<30 days since last injection (re-injection) Yes No
Dressing Gauze Gauze
Crepe bandage Yes No
Sling No No
Vitals checked Yes No
Time observing 10 minutes Nil
Staff at injection
Consultants 2 1
Registrars 1 1
Resident 1 1
Clinical coordinator 1 2
Nurse 0 0
Time since collagenase (days) 7 2
Location Clinic Theatre
Anaesthetic type Local Sedation
Anaesthetic classification Palm/ring block Light sedation
Lidocaine (lignocaine) 1–2% Yes No
Propofol 1% No Yes
Fentanyl No Yes
Oxygen mask No Yes
Consultant surgeon 1 1
Surgical registrar 1 2
Surgical resident 0 1
Anaesthetic consultant 0 1
Anaesthetic registrar 0 1
Anaesthetic nurse 0 1
Scrub/scout nurse 0 2
Preoperative/postoperative nurse 0 2
Outpatients nurse 1 0
Theatre technician 0 1
Porter 0 1
Clinic coordinator 0 2
Total staff 3 15
Gauze Yes No
Crepe bandage Yes No
Plaster cast/back slab No No
Skin split protocol
Dressing Jelonet, crepe, gauze Jelonet, crepe, gauze
Antibiotics No Yes—cefalexin
Analgesia No Yes
Follow-up As per normal Earlier
Splint As per normal—volar splint Only for wound healing
Frequency  Worn nightly All the time 
Length 4 months Until wound heals
Restrictions No Full-time until healed
Hand therapy Yes No
Hand exercises provided by Hand therapist Information sheet provided by doctors
Follow-up frequency 3 visits (average) 0 visits (average)
Clinic review
1 week review Yes Yes
First review 6 weeks 1 month

Clinic quantitative results

Demographic data

Seven statistically significant demographic differences were recorded when comparing Institute I and Institute II. Average age (Institute I < Institute II: 68 years old:58 years old), body mass index (BMI) < (Institute I < Institute II: 51%:19%), diabetes history, smoking history, family history, current use of blood thinning medication and weekly alcohol consumption were significantly different between the two cohorts. The most common country of origin was Australia for both cohorts, although Institute II recorded a higher percentage of diversity in country of origin.

Dupuytren’s disease history

Two statistically significant disease history variables were recorded when comparing Institute I and Institute II—disease length and previous hand surgery or injury. The mean length of years with Dupuytren’s disease was 9.7±5.8 years at Institute I versus 7.2±6.8 years at Institute II. A statistically significant greater percentage of Institute II participants recorded a previous hand injury or surgery unrelated to Dupuytren’s disease (75%) compared with Institute I (35%).


Correction of contracture

There was no statistically significant difference in the effectiveness measures between the two clinic results when analysing effectiveness.

There was nil significant difference in the proportion of each digit at each clinic that achieved clinical success or clinical improvement.

Overall, Institute I recorded greater clinical success (42%) compared with Institute II (35%). This was not statistically significant, as shown in Table 2. In comparison, CORD I and CORD II recorded a clinical success cohort percentage of 64 per cent and 44 per cent, respectively.2,14

Table 2: Effect of collagenase: clinical success and clinical improvement per digit at first review: Institute I vs Institute II
Clinical success Clinical improvement
Digit Institute I Institute II Test: Institute I vs Institute II Institute I Institute II Test: Institute I vs Institute II
Total number of digits injected  at 6 weeks (% of total number of digits injected) at 6 weeks Total number of digits injected  at 1 month  (% of total number of digits injected) at 1 month Fisher’s exact test Fisher’s exact test
Thumb 4/7  57% 3/7  43%
Index  3/4  75% 2/4 50%
Middle  8/10 80% 1/5 20% p=0.09 8/10  80% 5/5  100% p=0.52
Ring  22/31  71% 9/12 75% p=1.00 28/31  90% 11/12 92% p=1.00
Little  12/37 32% 12/37 25% p=0.731 33/37  89% 10/12 83% p=0.63
Overall  49/89  55% 13/29 45% p=0.395 74/89 83% 26/29 90% p=0.56

Institute II achieved a greater clinical improvement (86%) compared with Institute I (78%), which was not statistically significant, as highlighted in Table 2. In comparison, CORD I and CORD II recorded a clinical success cohort percentage of 85 per cent and 78 per cent, respectively.2,14

The change in total passive extension deficit (TPED) at first review following collagenase was found to be statistically significant in both cohorts (Figure 1). As shown in tables 2 and 3, in both cohorts the fourth and fifth digit had a better response to treatment. To determine whether differences existed between the TPED at each clinic a Mann–Whitney U-test was used. Nil statistically significant difference in the median TPED values from the two clinics (p=0.09) (Figure 1) was recorded.

Fig 1. Total effect of collagenase Clostridium histolyticum; overall total passive extension digit at Institute I vs Institute II. Mann–Whitney U-test for first review for difference in total passive extension deficit (TPED), p=0.09.

Table 3: Results of patient reported outcome measures pre and first review post-collagenase injection
Institute I Institute II
Test N Median Range (min max) N Median Range (min max)
Test: Wilcoxon signed rank test 46 Z = 5.8 p=0.0000 17 Z = -2.9 p=0.003
Southampton difference 
Institute I vs Institute II
Mann Whitney test 
Test: Wilcoxon signed rank test 49 Z = 6.1 p=0.000 17 –3.6 p<0.001
URAMs difference 
Institute I vs Institute II p=0.12
Mann Whitney test 
Test: Wilcoxon signed rank test 43 Z = 5.7 p=0.000 17 3.5 p=0.0004
PSFS difference p=0.53
Institute I vs Institute II

First review = 6 weeks for Institute I and 1 month for Institute II.
Wilcoxon signed rank test: used to assess whether significant change was achieved in each PROM following collagenase treatment at first review at each clinic.
Mann Whitney test: used to assess significant differences in PROM outcomes between the two hospitals.

Patient-reported outcome measures

A statistically significant reduction in all three PROMs occurred within Institutes I and II at six weeks post-collagenase treatment (Table 3). The three most common patient-reported responses to the PSFS were difficulty, concerns or grievances with the following tasks:

    • ‘catching on things’ 
    • ‘getting gloves on’ 
    • ‘progression’.  

A Mann–Whitney U Test was employed to determine whether significant differences existed in the three PROMs when comparing the two clinics. There was no statistically significance difference in each PROM (URAM, Southampton score and PSFS) reduction that occurred when comparing the two, as shown in Table 3.


Adverse effects following collagenase treatment at both clinics were comparable with results from the CORD I and II studies (Table 4).2,14 Of the Institute I and II cohorts, 88 per cent and 95 per cent, respectively, developed a treatment-related adverse effect. Nil serious adverse effects were recorded at Institute II, compared with one per cent of the Institute I cohort. These included two admissions for pain. Skin tears were more frequently observed, compared with the trial cohorts. The most common adverse effect recorded was oedema, reported in 71 per cent and 91 per cent and of the Institute I and Institute II cohort, respectively.2,14

Table 4: Adverse effects post-injection Institute I vs Institute II vs CORD I vs CORD II results2,14
Adverse effect Institute I Institute II CORD I CORD II
Treatment-related adverse effects 88% 95% 96.60% 100%
Serious adverse effects 1% 0 1% 2.20%
Skin tears 56% 48% 10.80% Not documented
Pain 22% 76% 26.50% Not documented
Bruising 40% 76% 51% 73.30%
Oedema 71% 91% 72.50% 77.80%
Lymphadenopathy 2% 10% 9.80% 24.40%
Blisters  13% 24% Not documented Not documented
Other 7% 0 Not documented Not documented
Self-manipulation 14% 5% Not documented Not documented


This study explored the application of collagenase as a therapeutic for people with Dupuytren’s disease within two public hospital clinics using different delivery protocols. The preliminary data show that collagenase is a safe and effective treatment for people with Dupuytren’s disease, irrespective of clinic design and delivery of collagenase. All outcomes of effectiveness (clinical success, clinical improvement, TPED and the three PROMs) showed a statistically significant improvement of contracture and hand function following collagenase treatment at six weeks post-injection in both clinics, and no significant difference was found when comparing between the two.

When collagenase was first introduced as a treatment option, the literature recommended its use for low-severity disease, with a maximum of only three doses per patient.2 However, since the drug’s approval, and its increasing use within the surgical profession, clinicians have become increasingly familiar and professionally comfortable with providing collagenase. This is evident as Institutes I and II currently offer the treatment for any level of disease severity.

Institute II offers treatment to bilateral hands at the same sitting if warranted, whereas Institute I only treats one hand per sitting. Both clinics treat recurrent cords (those that have previously been treated with collagenase or by another non-surgical/surgical method).

Injection delivery

Injection of collagenase was variable—the dose of collagenase, number of injections, injection direction and neo-adjunctive use of anaesthesia all differed between the two clinics. Institute I provided local anaesthesia with lidocaine (lignocaine), whereas Institute II solely injected the collagenase. Both clinics provided multiple small injections of the collagenase dose, which involved injecting a small amount of solution at a point along the cord, and continually injecting with the same solution along the cord.

The dosage delivered between the two clinics varied. Institute I provided an on-label solution dose of 0.25mg, in a diluent volume of 0.39mL, which equates to a dose of 0.58mg/mL. Institute I administered a dose of 0.9mg/mL across an average of 2.5 joints per ampoule. Institute II would frequently use 0.58mg over more than one joint. Institute II provided multiple doses of collagenase compared with Institute I. It is important to highlight that as this was a preliminary study, a conclusion cannot currently be made regarding different doses, their impact upon effectiveness and adverse effects.



The timeframe for manipulation was two days after the collagenase injection at Institute II, compared with seven days at Institute I. Institute I recorded a higher percentage of self-manipulation, potentially due to the increased timeframe, which may have enabled this to occur.


Manipulation was performed in an operating theatre at Institute II and an inpatient clinic at Institute I. Manipulation in the operating theatre was primarily attributed to the ability to provide intravenous light sedation.


Institute I provided local anaesthetic for patients undergoing manipulation, whereas at Institute II patients received intravenous sedation (so the patients maintained their airway under anaesthetic supervision).

Hand therapy

Institute I provided individualised, thermal hand splints and an appointment with a hand therapist to learn exercises aimed at maintaining the contracture correction. Institute II provided all patients with a paper instruction sheet explaining and illustrating the hand exercises.

Collagenase effectiveness

Contracture degree

A statistically significant reduction in median contracture degree (clinical success, clinical improvement and TPED) for both cohorts was recorded post-collagenase treatment at the first review. Furthermore, the range and interquartile across outcomes also decreased. In both cohorts, an improvement in the degree of flexion contracture was recorded at first review.

Across both cohorts, no digit was more responsive to collagenase treatment than another. Although Dupuytren’s disease is more prevalent in the fourth and fifth digits, it does not appear that collagenase has a pertinent response to a particular digit (for all effectiveness outcome measures).

Patient-reported outcome measures

A statistically significant reduction in functional impairment in both cohorts was recorded for all three PROMS. Similarly, patients in both cohorts reported a significant decrease in their functional impairment for the URAM, Southampton score and PSFS.9–11

Collagenase safety

Adverse effects occur with treatment. Nil serious adverse effects were recorded at Institute II, while one per cent of the Institute I cohort experienced a serious adverse effect.2 This is comparative to the existing literature, whereby in CORD I Hurst and colleagues reported one per cent of the treatment cohort developing a serious treatment adverse effect.2 The most common treatment-related adverse effects recorded were oedema, bruising and pain, which are acute, self-limiting occurrences that indicate an inflammatory response to the collagenase treatment.

The Institute II cohort recorded a greater proportion of adverse effects, which may be attributed to timing manipulation two days after injection and manipulation occurring under intravenous sedation.

Most adverse effects relied upon patient recall or reporting. Consequently, the validity of adverse effects is lessened by the impact of recall or reporting bias. This could be further heightened by the memory and hearing impairments of the ageing cohort.


These preliminary data demonstrate that collagenase results in improved functional outcomes, both objectively and as perceived by the patient. Collagenase leads to reduced contracture degree, as reported by the outcomes of clinical success, clinical improvement and TPED scoring as measured by a goniometer. This small, preliminary study investigated two different collagenase clinics and found that collagenase clinic design did not impact upon effectiveness or safety outcomes, rejecting the hypothesis. Both clinics deliver collagenase off-label and unconventionally, with nil statistically significant difference in outcomes, and outcomes that were comparable with CORD II (which was undertaken in Australia).14

This preliminary study demonstrated that there is no best way to approach Dupuytren’s disease management. Different collagenase delivery protocols at public hospitals can effectively and safely treat the disease.

Ethics approval

This study was approved by both clinics’ human research ethics and governance committees with reference numbers LRR17PH7 and 13226, respectively. Informed consent was obtained for data and photographs collected for this study.


The authors received no financial support for the research, authorship and/or publication of this article. David J Hunter-Smith and Warren M Rozen were consultants for Actelion (previous distributor of XIAFLEX® in Australia). The authors have no other conflicts of interest to declare.


  1. Shih B, Bayat A. Scientific understanding and clinical management of dupuytren disease. Nat Rev Rheumatol 2010;6(12):715–26. PMid:21060335 
  2. Hurst LC, Badalamente MA, Hentz VR, Hotchkiss RN, Kaplan FT, Meals RA, Smith TM, Rodzvilla J. Injectable collagenase Clostridium histolyticum for Dupuytren’s contracture. N Engl J Med 2009;361(10):968–79 PMid:19726771 
  3. Ball C, Pratt AL, Nanchahal J. Optimal functional outcome measures for assessing treatment for Dupuytren’s disease: a systematic review and recommendations for future practice. BMC Musculoskelet Disord. 2013;14(1):131 PMid:23575442 PMCid:PMC3637830 
  4. Tubiana R. Planning of surgical treatment. The Hand 1975;7(3):223–27. 
  5. Shaw RB, Chong AK, Zhang A, Hentz VR, Chang J. Dupuytren’s disease: history, diagnosis, and treatment. Plast Reconstr Surg. 2007;120(3):44e–54e. PMid:17700106 
  6. Eaton C. Evidence-based medicine: Dupuytren contracture. Plast Reconstr Surg. 2014;133(5):1241–251. PMid:24776555 
  7. Lee Matthew VK, Hunter-Smith D. Needle fasciotomy for Dupuytren’s disease: an Australian perspective. ANZ J Surg. 2009;79(11):776–78. PMid:20078523 
  8. Eaton C. Percutaneous fasciotomy for Dupuytren’s contracture. J Hand Surg Am. 2011;36(5):910–15. PMid:21527145 
  9. Beaudreuil J, Allard A, Zerkak D, Gerber RA, Cappelleri JC, Quintero N, Lasbleiz S, Bernabé B, Orcel P, Bardin T, URAM Study Group. Unité Rhumatologique des Affections de la Main (URAM) scale: development and validation of a tool to assess Dupuytren’s disease-specific disability. Arthritis Care Res. 2011;63(10):1448–455. PMid:21786431 
  10. Mohan A, Vadher J, Ismail H, Warwick D. The Southampton Dupuytren’s Scoring Scheme. J Plast Surg Hand Surg. 2014;48(1):28–33. PMid:24428161 
  11. Chatman AB, Hyams SP, Neel JM, Binkley JM, Stratford PW, Schomberg A, Stabler M. The Patient-Specific Functional Scale: measurement properties in patients with knee dysfunction. Phys Ther. 1997;77(8):820–29. PMid:9256870 
  12. Stratford P, Gill C, Westaway M, Binkley J. Assessing disability and change on individual patients: a report of a patient specific measure. Physiother Can. 1995;47(4):258–63. 
  13. STATA Statistical Software. Release fifteen [web page]. StataCorp [2017]. Available from:
  14. Gilpin D, Coleman S, Hall S, Houston A, Karrasch J, Jones N. Injectable collagenase Clostridium histolyticum: a new nonsurgical treatment for Dupuytren’s disease. Hand Surg Am. 2010;35(12): 2027–38. PMid:21134613