Article. Loop Ileostomy Closure After Restorative Proctocolectomy. Outcome in 1504 Patients. 2005
Article. Loop Ileostomy Closure After Restorative Proctocolectomy. Outcome in 1504 Patients. 2005
Article. Loop Ileostomy Closure After Restorative Proctocolectomy. Outcome in 1504 Patients. 2005
Restorative Proctocolectomy:
Outcome in 1,504 Patients
Kutt-Sing Wong, M.D., F.R.C.S.,1 Feza H. Remzi, M.D., F.A.S.C.R.S.,1
Emre Gorgun, M.D.,1 Susana Arrigain, M.A.,2 James M. Church, M.B.B.Ch., F.R.A.C.S.,1
Miriam Preen, R.N., B.S.N.,1 Victor W. Fazio, M.B., M.S.1
1
Department of Colorectal Surgery, The Cleveland Clinic Foundation, Cleveland, Ohio
2
Department of Biostatistics and Epidemiology, The Cleveland Clinic Foundation, Cleveland, Ohio
PURPOSE: Routine use of a temporary loop ileostomy for strated that ileostomy closure after restorative proctocolec-
diversion after restorative proctocolectomy is controversial tomy can be achieved with a low morbidity and a short
because of reported morbidity associated with its creation hospitalization stay. In addition, we found that complica-
and closure. This study intended to review our experience tion rates and length of hospitalization were similar be-
with loop ileostomy closure after restorative proctocolec- tween handsewn and stapled closures. [Key words: Ileosto-
tomy and determine the complication rates. In addition, my; Ileostomy closure; Restorative proctocolectomy; Pelvic
complication rates between handsewn and stapled closures abscess; Anastomotic dehiscence; Enterocutaneous fistula;
were compared. METHODS: Our Department Pelvic Pouch Bowel obstruction; Wound infection]
Database was queried and charts reviewed for all patients
who had ileostomy closure after restorative proctocolec-
tomy from August 1983 to March 2002. RESULTS: A total of
1,504 patients underwent ileostomy closure after restora-
tive proctocolectomy during a 19-year period. The median
length of hospitalization was three (range, 1–40) days and
A relatively high risk of anastomotic dehiscence
after restorative proctocolectomy has led to the
frequent use of a defunctioning loop ileostomy.1,2 For
the overall complication rate was 11.4 percent. Complica-
tions included small-bowel obstruction (6.4 percent), a defunctioning ileostomy to have any clinical utility,
wound infection (1.5 percent), abdominal septic complica- ileostomy-related complications including those per-
tions (1 percent), and enterocutaneous fistulas (0.6 per-
cent). Handsewn closure was performed in 1,278 patients
taining to ileostomy closure must be less than the risk
(85 percent) and stapled closure in 226 (15 percent). No and complications related to anastomotic dehiscence.
significant differences in complication rates and length of However, reports of high complication rates of ileos-
hospitalization were found between handsewn and stapled tomy closure after restorative proctocolectomy have
closure techniques. CONCLUSIONS: Our results demon-
discouraged the routine use of a defunctioning ileos-
tomy.3–8 Furthermore, there is concern about the
Read at the meeting of The American Society of Colon and Rectal need for a second operation with time in the hospital
Surgeons, New Orleans, Louisiana, June 21 to 26, 2003. and off work.
Correspondence to: Feza H. Remzi, M.D., F.A.S.C.R.S., Depart- This study intended to determine the outcome and
ment of Colorectal Surgery, The Cleveland Clinic Foundation, 9500
Euclid Avenue, Desk A30, Cleveland, Ohio 44195, email: complication rates of a large series of ileostomy clo-
remzif@ccf.org sures after restorative proctocolectomy. In addition,
Dis Colon Rectum 2005; 48: 243–250 this study was designed to compare the outcome and
DOI: 10.1007/s10350-004-0771-0
© The American Society of Colon and Rectal Surgeons complication rates between handsewn and stapled
Published online: 7 February 2005 techniques of ileostomy closure.
243
244 WONG ET AL Dis Colon Rectum, February 2005
We queried our Institution Review Board-approved Before ileostomy closure, the majority of patients
Pelvic Pouch Database for all patients who underwent had a Gastrografin enema performed to demonstrate
ileostomy closure after restorative proctocolectomy in free flow of contrast into the ileostomy bag (indicating
our department from August 1983 to March 2002 in- the absence of distal obstruction) and the absence of
clusive. Hospital and clinic charts of patients who had leakage from the pelvic pouch and the pouch-anal
incomplete data were reviewed. We noted patient de- anastomosis.
mographics, length of hospitalization, incision used Ileostomy closure was generally approached via a
for closure, closure technique, complications, and circumstomal incision and mobilization of the stoma
time interval from ileostomy construction to closure. performed down to the fascia and peritoneal cavity.
When difficulty was encountered during mobilization,
the incision was extended vertically. A midline inci-
Statistical Analysis sion was used in patients in whom mobilization was
To assess differences in complication rates by inci- impossible despite extension of the circumstomal in-
sion used and technique of closure, we used chi- cision, or sometimes in patients who had obstructive
squared and Fisher’s exact tests as appropriate. We symptoms or concurrent ventral midline hernias be-
assessed the relationship between incision and length fore closure. After mobilization, the integrity of the
of stay, using Kruskal-Wallis test and Dunn’s multiple afferent and efferent limbs was routinely checked
comparison procedure. We assessed the relationship with instillation of air and saline/Betadine.
of age and gender with closure technique using Wil- For handsewn closure, after the ileostomy had been
coxon rank-sum and chi-squared tests respectively. fully mobilized, the everted edges of the afferent limb
To assess the relationship between closure tech- were freed and thickened bowel edges trimmed. In
nique and length of stay, we used a negative binomial general, handsewn closure was performed with a
generalized linear model, adjusting for the year of single-layer interrupted serosubmucosal Vicryl® (Ethi-
closure because there was a trend toward earlier pa- con, Inc., Somerville, NJ) sutures. In some cases, clo-
tient discharge in more recent years. Because length sure was performed in two layers.
of stay was measured in days, a positive discrete vari- For stapled closure, a 6-cm linear cutter stapler was
able, a number of appropriate generalized linear inserted into each bowel limb followed by firing a
models were explored. A negative binomial model linear stapler across both limbs below the previous
offered much lower deviance and hence a substan- stoma. This resulted in a side-to-side functional, end-
tially improved fit compared with a Poisson model. to-end anastomosis.
To assess the relationship between timing of ileos- After ileostomy closure, the anterior fascia was
tomy closure and development of complications, we closed with 1-0 Vicryl®, PDS® (Ethicon, Inc.), or
used logistic regression analysis. Because of the dis- Prolene® (Ethicon, Inc.) sutures, depending on sur-
tribution of the interval between ileostomy construc- geon preference. The majority of circumstomal inci-
tion and closure, we took the log base 2 of months to sions were left partially or completely open. Midline
closure to model the odds of developing a complica- wounds were closed in the standard fashion: mass
tion. This transformation increased the goodness-of- closure with PDS® or Prolene® followed by staples to
fit for all logistic regression models. skin.
We used the median time of ileostomy closure to
divide patients into early and late closure groups and Definition of Complications
tested differences in the proportion of complications
before and after that time using chi-squared tests. All Small-bowel obstruction was defined by a combi-
complications are mentioned in the RESULTS, al- nation of the following findings: abdominal disten-
though wound infection had the only significant dif- tion, abdominal pain, vomiting, or the presence of
ference. multiple air-fluid levels on a plain abdominal radio-
Statistics were generated using SAS 8 software (SAS graph in the postoperative period. Wound infection
Institute Inc., Cary, NC) and S-PLUS 6.1 (Insightful was defined by the presence of purulent wound dis-
Inc., Seattle, WA). P < 0.05 was considered statistically charge, wound erythema, and induration. Abdominal
significant. septic complications were defined by the presence of
Vol. 48, No. 2 OUTCOME OF ILEOSTOMY CLOSURE 245
Table 3.
Incidence of Small-Bowel Obstruction by Incision Type
Small-Bowel
No Small-Bowel Obstruction
Incision Obstruction (n = 1,407) (n = 97) P Valuea
Midline (n = 80) 73 (91.3) 7 (8.87) 0.23
Circumstomal (n = 1,214) 1,132 (93.3) 82 (6.8)
Extended circumstomal (n = 201) 193 (96) 8 (4)
Data are numbers with percentages in parentheses unless otherwise indicated.
a
Chi-squared test.
Table 5.
Timing of Closure and Complications
Odds Ratio Per
No. of Mean No. of Months Doubling in Time
Complication Level Patients to Closure (SD) to Closure (95% CI) P Valuea
Small-bowel obstruction No 1,407 3.6 (1.9) 0.94 (0.62–1.42) 0.78
Yes 97 3.7 (2.3)
Enterocutaneous fistula/abdominal No 1,480 3.6 (2) 0.73 (0.32–1.7) 0.47
septic complications Yes 24 3.4 (1.2)
Wound infection No 1,481 3.6 (2) 0.39 (0.18–0.85) 0.017
Yes 23 3 (1.1)
SD = standard deviation; CI = confidence intervals.
a
Logistic regression analysis. All odds ratios were calculated using the log2 of months to closure.
13. Hasegawa H, Radley S, Morton DG, Keighley MR. with stapled loop ileostomy closures. Dis Colon Rectum
Stapled versus sutured closure of loop ileostomy: 1996;39:1086–9.
a randomized controlled trial. Ann Surg 2000;231: 16. Sagar PM, Lewis W, Holdsworth PJ, Johnston D. One-
202–24. stage restorative proctocolectomy without temporary
14. Bain IM, Patel R, Keighley MR. Comparison of sutured defunctioning ileostomy. Dis Colon Rectum 1992;35:
and stapled closure of loop ileostomy after restora- 582–8.
tive proctocolectomy. Ann R Coll Surg Engl 1996;78: 17. Galandiuk S, Wolff BG, Dozois RR, Beart R. Ileal pouch-
555–6. anal anastomosis without ileostomy. Dis Colon Rectum
15. Hull TL, Kobe I, Fazio VW. Comparison of handsewn 1991;34:870–3.