Incisional hernia remains one of the most important long-term sequelae after laparotomy.1,2 Although several conditions3,4 have been found to be associated with the development of incisional hernia, the occurrence of this complication in gynecologic oncology patients has not been fully investigated.
Colombo et al5 reported that the occurrence of incisional hernia within 3 years after midline abdominal incision for gynecologic malignancies was strongly associated with obesity, postmenopausal status, and anesthesiologic risk factors, but not with closure technique (continuous versus interrupted suture). Similarly, Sutton and Morgan6 found that abdominal closures in gynecologic patients with either a running, looped, monofilament polybutester suture or Smead-Jones interrupted suture were comparable in their likelihood of wound dehiscence.
Conversely, in patients undergoing laparotomy for nongynecologic diseases, a relation has been described between the development of incisional hernia and both the suture material used for abdominal closure7,8 and the suture technique.9,10 Moreover, early wound infection,11 length of the abdominal incision,12 chest infection,13 and male gender13 were found to be clearly associated with the subsequent formation of an incisional hernia.
In most investigations, patients were followed up for incisional hernia only until 1 year after the operation.7,8,13,14 However, Mudge and Hughes15 noted that in patients undergoing elective major abdominal surgery of the gastrointestinal tract, biliary tree, or colon, the incidences of incisional hernia after 1 and 3 years from the operation were 44% and 79%, respectively. The purpose of our study was to explore the factors associated with the formation of early and late incisional hernia in patients undergoing surgery for uterine malignancies during a 10-year period.
Materials and Methods
Our study included consecutive women who underwent extended abdominal hysterectomy with bilateral salpingo-oophorectomy (BSO) type I–IV of Piver et al,16 for cervical or endometrial malignancies between January 1, 1989 and June 31, 1999 at our department. Only patients who underwent the operation through a midline abdominal incision were considered. The presence of an incisional hernia at the time of surgery was an exclusion criterion. A cross-sectional study was designed, and two groups of subjects were identified according to the development (or not) of incisional hernia.
Preoperative mechanical bowel preparation was used in all patients. Pubic hair removal and vaginal preparation were performed in the surgery room. If antibiotic prophylaxis was given, 1 g amoxicillin and 0.5 g clavulanic acid were administered 1 hour before the initiation of the operation. Laparotomy was performed either with a scalpel from the skin through the fascia or with a scalpel for the skin incision followed by diathermy in coagulation mode from the subcutaneous tissue through the fascia. Bleeding points were controlled by coagulation diathermy in both groups. The fascia was closed with a mass continuous no. 1 polyglactin multifilament suture (Vicryl; Ethicon, Inc., Somerville, NJ), interrupted reabsorbable polydioxanone monofilament sutures (PDS; Ethicon, Inc.), or interrupted reabsorbable polydioxanone monofilament sutures (PDS; Ethicon, Inc.) placed at more than 1.5 cm from the wound edges and 1 cm apart, depending on the surgeon's preference. Subcutaneous drains were not used.
Forms to record information prospectively on the gynecologic oncology patients were used throughout the study period at our institution and were updated at each follow-up visit. Demographic characteristics, type of surgery, intraoperative details, and follow-up information were collected. A specific part of the form was intended to record the occurrence of early and late wound complications.
Wounds were assessed during the postoperative hospital stay, and patients were subsequently followed up at 1, 3, 6, and 12 months after the operation and then yearly according to our oncologic follow-up program. During the follow-up, the women were examined in the horizontal supine position and then were asked to sit up and cough to reveal any incisional hernia. Any abnormality giving rise to a protruding swelling with a palpable fascial defect of at least 3 cm in diameter seen during the follow-up visits was considered an incisional hernia.10 Severe wound infection was defined as the presence of erythema, induration, pain, and serous fluid or pus. According to the number of gynecologic oncology operations performed during the last 5 years, attending and faculty physicians were defined as seniors (at least 50) or juniors (fewer than 50). An incisional hernia was considered to occur late when it developed at least 5 years after surgery. The ethics committee of our institution approved the review of data collected in our gynecologic oncology forms.
Statistical analysis was performed with SPSS package (SPSS Inc., Chicago, IL) and with Epistat 4.0 (Epistat Services, Richardson, TX). Student t test or Mann-Whitney U test was used to compare continuous variables, whereas proportions were analyzed with χ2 or Fisher exact test where appropriate. To evaluate incisional hernia formation over time, a Kaplan-Meier survival curve was generated. Patients who died were considered as lost to follow-up with a censored time equal to the interval between the operation and the date of death. Receiver operating characteristic (ROC) curves were constructed to describe the relation between sensitivity and false-positive rate for different values of age and body mass index (BMI) in predicting the occurrence of incisional hernia. The null hypothesis was considered a curve with an area below it of 50%. Logistic regressions were used to investigate the relation between a number of different variables and the occurrence of incisional hernia. Age and BMI were entered into the logistic regression model as dichotomous variables, using cutoffs derived from the ROC curves analysis. The regression relations between the log odds of the presence of incisional hernia and the explanatory variables were examined with overall logistic regressions. Statistical significance was defined as P < .05.
During the study period, 460 women underwent extended abdominal hysterectomy with BSO type I–IV of Piver et al, for cervical or endometrial malignancies at our institution. Five were lost to clinical follow-up within 12 months from the operation and were not included in the analysis, leaving 455 in the study population. The total incidence of incisional hernia during the study period was 16.9% (95% confidence interval [CI] 9.59, 24.21) (77 of 455). Figure 1 shows the cumulative probability of developing an incisional hernia during the 10-year period. Of the 77 incisional hernias, 23, 56, and 67 occurred within 1, 3, and 5 years from the operation, respectively. Eighteen patients died during the study period, of whom one developed an incisional hernia 26 months after the operation. Eleven of 455 women (2.4%) were lost to follow up.
Table 1 shows the demographic and clinical characteristics according to the presence (or not) of incisional hernia. Operative and postoperative characteristics of the patients are presented in Table 2. The ROC curve analysis demonstrated a significant relation between incisional hernia formation and both the BMI (area under the curve = 63%; P < .05) and age at operation (area under the curve = 67%; P < .05). The values located at the knees of the curves were 27 kg/m2 for BMI and 62 years for age, respectively. We decided to use these values for statistical purposes because they had the best sensitivity and specificity for the identification of patients who subsequently developed an incisional hernia.
Logistic regression revealed significant relations between incisional hernia formation within 1 year after surgery and age (odds ratio [OR] 2.73; 95% CI 1.01, 5.62; P < .05), BMI (OR 4.64; 95% CI 1.83, 12.02; P < .01), severe wound infection (OR 7.44; 95% CI 2.22, 24.90; P < .01), and fascial closure with interrupted suture (OR 1.61; 95% CI 1.03, 2.23; P < .05). Multiple logistic regression showed that the only conditions independently associated with the occurrence of incisional hernia within 1 year after the operation were BMI and postoperative wound infection (Table 3). When the analysis was restricted to patients with a follow-up of at least 3 years, the variables independently associated with the development of an incisional hernia at least 3 years after surgery were BMI at least 27 kg/m2 (OR 2.29; 95% CI 1.02, 5.63; P < .05), diabetes (OR 6.66; 95% CI 2.24, 19.83; P < .01), fascial closure with interrupted suture (OR 3.82; 95% CI 1.02, 6.58; P < .05), and severe wound infection (OR 8.12; 95% CI 1.82, 36.51; P < .01). When all of these covariates were entered simultaneously into a logistic regression model, only the presence of diabetes and severe postoperative wound infection remained significantly associated with incisional hernia formation (Table 3).
We further analyzed patients with at least 5 years and 8 years of follow-up to explore a possible relation between conditions present at the time of operation and the occurrence of incisional hernia. Multiple logistic regressions revealed that when the presence of diabetes, occurrence of severe postoperative wound infection, BMI, and type of fascial closure were entered in a model, the only variable significantly associated with the occurrence of an incisional hernia after 5 years (OR 8.32; 95% CI 1.41, 55.65; P < .05) and 8 years (OR 49.52; 95% CI 2.72, 907.14; P < .01) from the operation was diabetes. A significant difference in incision length was found between early and late incisional hernias (median 9 cm, range 4–20; versus median 5 cm, range 3–12; P < .05). Nine patients who developed an incisional hernia required surgical repair; of these hernias, seven occurred within 12 months from the operation, one at 19 months, and one at 12 years after surgery. The last patient was a 65-year-old diabetic woman with a BMI of 28 kg/m2 who underwent an extended hysterectomy for endometrial cancer and who had an uneventful postoperative course.
The development of incisional hernia within 3 years of operation has been attributed to the surgical technique and to the occurrence of postoperative wound infections in individuals with predisposing conditions (ie, postmenopausal status, obesity, metabolic disorders).3,5 Our findings and those of Colombo et al5 confirm the association between incisional hernia formation within 1 year from operation and both an elevated BMI and an advanced age in gynecologic oncology patients.
The role of postoperative wound infection in the occurrence of incisional hernia is still a matter of controversy. Our study contrasts with that of Colombo et al5 because they did not find an association between early incisional hernia formation and postoperative wound infections. A possible explanation for this discrepancy could be the different populations investigated. Whereas our study included only patients with uterine cancers, in the study by Colombo et al,5 the majority of patients (61%) were affected by ovarian cancer. We did not include patients with ovarian cancers in our study because of the short follow-up time in most of them. In addition, Colombo et al5 did not report the incidence and degree of postoperative wound infections.
The findings of the present study are in keeping with several studies conducted by general surgeons.3,10 Riou et al3 reported that, in patients undergoing major abdominal surgery, the proportion of wound infections was significantly higher in patients who developed a wound dehiscence in the immediate postoperative period than in those who did not (45% versus 2%) Wissing et al10 reported a significant association between postoperative wound infection and the occurrence of an incisional hernia within 1 year from the operation.
The present study and that of Ellis et al,17 who followed patients up to 5.5 years after major laparotomy, did not find an association between late incisional hernia formation and postoperative wound infection. It seems that the relevance of wound infection in the etiology of incisional hernia is greater for hernias occurring in the immediate postoperative period than for those occurring years after the operation.
Several authors5–7,11,18 have investigated whether the type of suture material used and the technique used for fascial repair may influence the development of incisional hernia. Although a consensus exists on the lower risk of incisional hernia formation with reabsorbable sutures,19 it is still debated whether continuous sutures offer advantages over interrupted sutures.1 In most of the studies6,9,14,18 exploring this issue, the contribution of the fascial closure method on incisional hernia formation was not adjusted for confounding variables.
Our study agrees with that of Orr et al,19 who reported that after gynecologic abdominal procedures, the risk of incisional hernia was not affected by the method of closure (continuous versus interrupted fascial closure). Few studies evaluating the occurrence of late incisional hernia are reported in the literature.15,20 Mudge and Hughes15 reported that, during a 9-year follow-up period, 21% of incisional hernias occurred at least 5 years after general surgery. However, these authors did not report whether clinical conditions, surgical modalities, postoperative complications, and sex of the patients were associated with the late appearance of incisional hernia.
A MEDLINE search of the literature from 1966 to August 2000 using the terms “incisional hernia,” “postoperative wound complications,” “wound healing,” and “abdominal closure” revealed that the present experience is the first study investigating the development of an incisional hernia 3 years after the operation in gynecology patients. We found that the longer the time between operation and incisional hernia formation, the lower the importance of the surgical modalities and postoperative complications. Indeed, in our study, the only condition associated with incisional hernia formation at least 5 and 8 years after the operation was the presence of diabetes. Ellis et al17 have also reported the lack of causality between factors implicated in the etiology of early incisional hernia (eg, obesity, wound infection, age) and the occurrence of late incisional hernia.
The association between diabetes and the development of late incisional hernia may be explained by the altered regulation of collagen metabolism at the level of the fascial scar in patients with metabolic disorders. Experimental studies21–23 have demonstrated that when diabetes was optimally controlled, wound strength and extensibility were similar in diabetic and nondiabetic rats. Moreover, delayed insulinemic treatment in diabetic rats was not able to prevent the inhibitory effects of diabetes on wound healing.23 We believe that the association between incisional hernia and diabetes might be explained by the suboptimal glycemic control often found in elderly diabetic subjects and not by the disease per se.
Finally, our study suggests the need for a targeted examination for the identification of incisional hernia at each follow-up visit, even several years after the operation. The clinical importance of late incisional hernia is not the same between gynecologic oncology patients and those undergoing major nongynecologic abdominal surgery. Our study is in accordance with that of Colombo et al,5 who reported that only 3.9% of incisional hernias required surgical reparation, and all of these procedures were performed within 2 years of the operation. In contrast, Read and Yoder24 reported that 21.6% of incisional hernias requiring surgical treatment occurred at least 3 years after a major abdominal operation. This discrepancy can be explained by the different populations investigated. Indeed, in the study by Read and Yoder,24 99.5% of the patients who underwent surgical treatment for incisional hernias were men, 96% were smokers, and 34% had previous multiple laparotomies. It has been reported recently that estrogen plays a pivotal role in collagen metabolism during the early phase of wound healing.25
1. Wantz GE, Chevrel JP, Flament JB, Kingsnorth A, Schumpelik V, Verhaeghe P. Incisional hernia: The problem and the cure. J Am Coll Surg 1999;188:429–47.
2. Hsiao WC, Young KC, Wang ST, Lin PW. Incisional hernia after laparotomy: Prospective randomized comparison between early-absorbable and late-absorbable suture materials. World J Surg 2000;24:747–51.
3. Riou JA, Cohen JR, Johnson H. Factors influencing wound dehiscence. Am J Surg 1992;163:324–30.
4. Nygaard IE, Squatrito RC. Abdominal incisions from creation to closure. Obstet Gynecol Surv 1996;51:429–36.
5. Colombo M, Maggioni A, Parma G, Scalambrino S, Milani R. A randomized comparison of continuous versus interrupted mass closure of midline incisions in patients with gynecologic cancer. Obstet Gynecol 1997;89:684–9.
6. Sutton G, Morgan S. Abdominal wound closure using a running, looped monofilament polybutester suture: Comparison to Smead-Jones closure in historic controls. Obstet Gynecol 1992;80:650–4.
7. Sahlin S, Ahlberg J, Granström L, Ljungström KG. Monofilament versus multifilament absorbable sutures for abdominal closure. Br J Surg 1993;80:322–4.
8. Fagniez PL, Hay JM, Lacàine F, Thomsen C. Abdominal midline incision closure. A multicentric randomized prospective trial of 3,135 patients, comparing continuous vs interrupted Polyglycolic acid sutures. Arch Surg 1985;120:1351–3.
9. Israelsson LA, Jonsson T. Closure of midline laparotomy incision with polydioxanone and nylon: The importance of suture technique. Br J Surg 1994;81:1606–8.
10. Wissing J, van Vroonhoven TJMV, Schattenkerk ME, Veen HF, Ponsen RJG, Jeekel J. Fascia closure after midline laparotomy: Results of a randomized trial. Br J Surg 1987;74:738–41.
11. Krukowski ZH, Cusick EL, Engeset J, Matheson NA. Polydioxanone or polypropylene for closure of midline abdominal incisions: A prospective comparative clinical trial. Br J Surg 1987;74:828–30.
12. Pollock AV, Greenall MJ, Evans M. Single layer mass closure of major laparotomies by continuous suturing. J R Soc Med 1979;72:889–93.
13. Bucknall TE, Ellis H. Abdominal wound closure—a comparison of monofilament nylon and polyglycolic acid. Surgery 1981;89:672–6.
14. Gallup DG, Nolan TE, Smith RP. Primary mass closure of midline incision with a continuous Polyglyconate monofilament absorbable suture. Obstet Gynecol 1990;76:872–5.
15. Mudge M, Hughes LE. Incisional hernia: A 10 year prospective study of incidence and attitudes. Br J Surg 1985;72:70–1.
16. Piver MS, Rutledge F, Smith JP. Five classes of extended hysterectomy for women with cervical cancer. Obstet Gynecol 1974;44:265–72.
17. Ellis H, Gajraj H, George D. Incisional hernias: When do they occur? Br J Surg 1983;70:290–1.
18. Trimbos JB, Smit IB, Holm JP, Hermans J. A randomized clinical trial comparing two methods of fascia closure following midline laparotomy. Arch Surg 1992;127:1232–4.
19. Orr JW Jr, Orr PF, Barrett JM, Ellington JR Jr, Jennings RH, Paredes KB, et al. Continuous or interrupted fascial closure: A prospective evaluation of no. 1 Maxon suture in 402 gynecologic procedures. Am J Obstet Gynecol 1990;163:1485–9.
20. Harding KG, Mudge M, Leinster SJ, Hughes LE. Late development of incisional hernia: An unrecognized problem. BMJ 1983;286:519–20.
21. Yue DK, McLennan S, Marsh M, May YW, Spaliviero J, Delbridge L, et al. Effects of experimental diabetes, uremia and malnutrition on wound healing. Diabetes 1987;36:295–9.
22. McMurry JF Jr. Wound healing with diabetes mellitus. Better glucose control for better wound healing in diabetes. Surg Clin North Am 1984;64:769–78.
23. Andreassen TT, Oxlund H. The influence of experimental diabetes and insulin treatments on biochemical properties of rat skin incisional wound. Acta Chir Scand 1987;153:405–9.
24. Read RC, Yoder G. Recent trends in the management of incisional herniation. Arch Surg 1989;124:485–8.
© 2001 The American College of Obstetricians and Gynecologists
25. Calvin M. Oestrogens and wound healing. Maturitas 2000;34:195–210.