Pelvic organ fistula disease is a devastating condition that is difficult to treat and often demands repeat surgical interventions.1–4 Fistulae resulting from obstetric trauma remain a major problem in developing countries, whereas fistula disease as a complication of childbirth is rare in industrialized countries due to advances in modern obstetric care.5,6 Instead, pelvic surgery, and hysterectomy in particular, are widely considered a major cause of genital fistula disease.1,2
The majority of hysterectomies are performed for benign uterine disorders,7,8 and estimates suggest that one in three U.S. women9,10 and one in five British women11 will undergo a hysterectomy during their lifetime. It has been suggested that hysterectomy, as well as the mode and route of hysterectomy, influences the risk of fistula formation.4,6 However, support for an association between hysterectomy and pelvic organ fistula disease generally derives from case series and retrospective hospital-based studies with inadequate sample sizes.
In the present nationwide Swedish study, we aim to study the association between hysterectomy for benign indications and pelvic organ fistula disease. We also intend to estimate the effect of various hysterectomy techniques and of age at hysterectomy on the association. To achieve sufficient statistical strength to study this relatively rare disease, we conducted a nationwide cohort study based on Swedish health care registers.
MATERIALS AND METHODS
We defined our exposed cohort by identifying all records in the Swedish Inpatient Register indicating a performed hysterectomy from January 1, 1973, until December 31, 2003 (exposed cohort). For every woman having hysterectomy, we randomly selected three women not undergoing the procedure from the Register of the Total Population (unexposed cohort). The unexposed women were individually matched to the exposed by year of birth, calendar date for the hysterectomy (start of follow-up), and county of residence at year of hysterectomy. Using national registration numbers, our exposed (n=227,385) and unexposed (n=682,129) cohorts were linked to other nationwide Swedish registers, including the Cancer Register, the Emigration Register, and the Cause of Death Register. Exposed and unexposed women were followed up from the time of entry into the cohort (time of hysterectomy for the exposed cohort and matching for the unexposed cohort) until the first occurrence of pelvic organ fistula surgery, pelvic floor surgery, Crohn’s disease, ulcerous colitis, diverticular disease, any type of cancer disease, death, emigration, or end of follow-up (December 31, 2003), whichever came first. Person-time for each person was counted from the time of entry to the cohort until occurrence of any of the above-mentioned events.
Exclusion criteria implemented for both the exposed and unexposed group were 1) if fistula surgery preceded or concurred with study entry (n=624); 2) if the hysterectomy was performed on malignant indications (n=6,749); 3) if the patient had any record in the Cancer Register at the entry to the cohort (n=65,063); 4) if the patient had any record of Crohn’s disease, ulcerous colitis, or diverticular disease, before, or at entry to the cohort (n=8,935); or 5) if the patient was younger than 18 years of age (n=144). Four patients were excluded due to erroneous data coding. When an exposed woman was excluded, the whole matching stratum was excluded.
The Swedish Inpatient Register contains data on individual hospital discharge diagnoses according to the International Classification of Diseases (ICD). International Classification of Diseases-7 was used until 1968, ICD-8 from 1968 through 1986, ICD-9 from 1987 through 1996, and ICD-10 thereafter. In 1973, the register covered more than 60% of the population, in 1983 the coverage was 85%, and from 1987 it has complete national coverage (http://www.socialstyrelsen.se/Publicerat/2008/10174/2008–125–15+summary.htm).The inpatient discharge record also contains the dates of hospital admission and discharge and up to 12 operation codes according to the Swedish Classification of Operations and Major Procedures. Ninety-eight percent of records achieved correct coding for surgical procedures with a less than 1% yearly loss to registration when cross-checked with hospital charts.12
The specific codes from the Swedish Classification of Operations and Major Procedures for 1973–1996 and 1997–2003 were used to identify surgical procedures in the Swedish Inpatient Register. Hysterectomy was classified as abdominal total hysterectomy (operation code 7210 for 1973–1996 and LCD00 for 1997–2005), abdominal subtotal hysterectomy (operation code 7211 for 1973–1996 and LCC10 for 1997–2005), vaginal hysterectomy (operation codes 7261, 7262 for 1973–1996 and LCD10 for 1997–2005), laparoscopy-assisted vaginal hysterectomy (operation code LCD11 for 1997–2005), laparoscopic total hysterectomy (operation codes LCD01–04 for 1997–2005), and hysterectomy with concurrent pelvic organ prolapse repair (operation code 7467 for 1973–1996 and LEF13 for 1997–2005). Laparoscopic total hysterectomy and laparoscopy-assisted vaginal hysterectomy were combined in the statistical analyses.
Surgical procedures for pelvic organ fistula disease were divided into five categories based on the primary ICD code for fistula surgery: urogenital fistula (ICD-8 code 599B, ICD-9 code 619A, ICD-10 code N360, N820, N821), intestinogenital fistula (ICD-8 code 629,80, 565B, ICD-9 code 619B, ICD-10 code N822, N823, N824), urointestinal fistula (ICD-8 code 596B, 596C, ICD-9 code 619A, ICD-10 code N321, N322), anorectal fistula (ICD-8 code 565,10, ICD-9 code 565B, ICD-10 code K603, K604, K605), and unspecified pelvic organ fistula (ICD-8 and 9 code 619C, 619W, 619X, ICD-10 code N825, N828, N829).
In this matched cohort study, we used survival analysis. Person-time at risk was calculated from the date of hysterectomy (exposed cohort) or from the date of matching (unexposed cohort) until first occurrence of pelvic organ fistula surgery or censoring. Women were censored at the time of diagnosis of Crohn’s disease, ulcerous colitis, diverticular disease, other pelvic surgery, emigration, cancer, death, or end of study (December 31, 2005). We calculated unadjusted incidence rates for fistula disease as the number of fistula operations per 100,000 person-years, with 95% confidence intervals (CIs) based on the Poisson distribution. Only the first occurrence of fistula in each woman was counted as an event. Incidence rates were calculated separately for the various fistula types in the exposed and the unexposed cohort. Subsequently, incidence rates for the various hysterectomy techniques were calculated.
The association between pelvic organ fistula disease and hysterectomy was estimated using a Cox proportional-hazards model, stratified on the matching variables age, county, and calendar time for all analyses. The hazard ratio (HR) was used as a measure of association and is presented with 95% CIs. Strata with nonunique values on the matching variables (age 1-year intervals, calendar time 1-year intervals, and county) were collapsed into the same strata, thus maintaining the same ratio of exposed to nonexposed (1:3). The time scale used in all analyses was time since study entry. Likelihood ratio tests were used to formally assess effect of exposure. Because the hazard rates were significantly nonproportional over follow-up time, we included interactions with time in the model by splitting each record on time and using interaction terms with time bands (cutoffs at 1, 2, 3, 5, and 10 years). Interaction between exposure and age was formally tested using the likelihood ratio test. The statistical analyses were undertaken with Stata Intercooled (Release 10. StataCorp LP, College Station, TX). The study was approved by the Research Ethics Committee at Karolinska.
The final study population available for analysis included 182,641 women having a hysterectomy (exposed cohort) and 525,826 women not undergoing hysterectomy (unexposed cohort), encompassing more than 8 million person-years at risk. During the follow-up period, 469 cases of fistula surgery were identified in the exposed cohort and 384 in the unexposed cohort (Table 1). The overall number needed to harm was 5,700; ie, if 5,700 individuals are exposed to hysterectomy for benign indications, one will develop the disease that they would not have otherwise.
The rates for specific fistula diagnosis with regard to hysterectomy status and mode of hysterectomy are presented in Table 2. The most common type of fistula observed in the exposed cohort was urogenital fistula, followed by intestinogenital fistula. In the unexposed cohort, intestinogenital and anorectal fistula were the most common types, whereas urogenital fistulae were infrequent (Table 2). Laparoscopic hysterectomy was associated with the highest rate of overall fistula surgery, more than three times higher than the corresponding rate for total abdominal hysterectomy. The lowest rate of fistula surgery was seen in the subtotal abdominal hysterectomy group. Fistula surgery after laparoscopic, vaginal, or abdominal hysterectomy was dominated by urogenital fistulae, whereas intestinogenital fistulae were the most common type in the subtotal abdominal hysterectomy group (Table 2).
Table 3 presents the relative risk (RR) (HR) of pelvic organ fistula disease depending on mode of hysterectomy and time since surgery. During the first 2 years after surgery, abdominal hysterectomy was associated with the highest RR of fistula disease; thereafter, the risk did not essentially differ between the various types of hysterectomy (Table 3).
In Table 4, HRs for pelvic organ fistula disease are presented stratified by age at hysterectomy (or at study entry) and time since surgery. Regardless of age at hysterectomy, the overall RR for fistula disease was more than 20 times higher in women with hysterectomy than in women without the procedure during the first year after surgery. When stratifying for age at hysterectomy, we found that the RR of fistula disease 1 to 2 years after hysterectomy was higher among women who were aged at least 50 years at hysterectomy, compared with younger women (HR 7.7, 95% CI 3.0–20.3 and HR 1.8, 95% CI 0.9–3.5, respectively).
The RR estimates for fistula disease over time from hysterectomy, or study entry for the unexposed cohort, are presented in Table 5. The highest risk for fistula surgery after hysterectomy was observed during the first and second year after surgery (HR 21.2, 95% CI 14.9–30.2 and HR 3.0, 95% CI 1.8–5.1, respectively). Thereafter, the risk for fistula surgery remained higher among exposed compared with unexposed women, although the risk increase was moderate and remained stable for the duration of the observational period.
In this nationwide cohort study extending over three decades, we found that the rate of fistula surgery was four times higher among women having a hysterectomy than among women not undergoing the procedure. The highest overall fistula rates were observed after laparoscopic and total abdominal hysterectomy, among older women, and during the first year after surgery, where the overall risk for fistula disease was more than 20 times higher after a hysterectomy than for women with an intact uterus.
Our main findings are in agreement with previous studies where gynecologic surgery and hysterectomy have been identified as risk factors for fistula disease in developed countries.1,2 It is biologically plausible that the observed association can be attributed to iatrogenic injury or complications of surgery such as undetected bowel, bladder, or ureteral lesions or postoperative infections.13 The overall most common types of fistulae were urogenital fistulae and intestinogenital fistulae. Urogenital fistulae are thought to originate from dissection of the bladder, from sutures incorporated into the bladder, or direct injuries to the lower urinary tract.4,14 Intestinogenital fistulae may originate from injuries to the bowels, most commonly the rectum, which has been reported during vaginal surgery and hysterectomy.15,16 Without the dense smooth muscle uterus in place to act as a protective buffer, postoperative infections of the vagina or paravaginal tissues may act as a source for fistula formation soon after surgery, a process which may be precipitated by injury to the surrounding pelvic organs. This would explain the striking risk increase for fistula surgery observed within the first year of hysterectomy.
In concurrence with other researchers,2,4,6 we found that the rate of fistula disease was higher in women who had total abdominal hysterectomy than those who had a vaginal hysterectomy. Hysterectomy by way of laparotomy may introduce a higher risk for pelvic organ fistula disease because of its greater invasiveness and higher incidence of infections compared with the vaginal approach.17 For both abdominal and vaginal hysterectomy, fistulae were predominantly of the urogenital type, although intestinogenital fistulae were nearly as common. The overall lowest rate of fistula surgery was seen in women after subtotal abdominal hysterectomy, and in contrast to total abdominal hysterectomy, intestinogenital fistulae were the most common type, with a relatively low rate of urogenital fistula. During subtotal vaginal hysterectomy, the cervix is preserved and no blunt dissection of the bladder from the cervix or vagina is performed. This part of the surgical procedure has been indicated as the key step for the occurrence of bladder injury at hysterectomy.13 As a result, a combination of leaving the cervix intact and minimal bladder dissection may explain the lower rate of urogenital fistulae after subtotal abdominal hysterectomy.
Laparoscopic hysterectomy has been associated with a low perioperative morbidity and shorter recovery period than abdominal hysterectomy,18,19 and is therefore advocated as the technique of choice if vaginal hysterectomy is not possible.17,20 However, epidemiologic studies suggest that laparoscopic techniques are associated with increased risks for both stress urinary incontinence and pelvic organ prolapse surgery, compared with abdominal and vaginal methods.21,22 In accordance with these findings we found the highest rates of pelvic organ fistula surgery in the laparoscopic hysterectomy group. The majority of fistulae in the laparoscopic hysterectomy group were urogenital fistulae, confirming results from previous studies of higher rates of bladder and ureteral lesions after laparoscopic hysterectomy.20,23 The results are also in agreement with the presumption that the majority of fistulae originate from undetected bladder or ureteral lesions. It should, however, be pointed out that the number of fistula cases in the laparoscopic group were few, and the statistical precision was limited.
The risk of pelvic fistula disease was higher in women having had their hysterectomy at 50 years or older. This can presumably be attributed to perimenopausal and postmenopausal estrogen deficiency resulting in degenerative changes of the vaginal tissues, which in turn may affect tissue susceptibility to infectious and/or inflammatory processes. Older women may also have a higher incidence of other diseases associated with poor tissue healing and susceptibility for infections. We did not have access to information on some potential confounders or effect modifiers such as previous abdominal surgery or diabetes, which is treated primarily in community care. Also, the influence of life style factors such as smoking and hormone replacement could not be taken into account. Due to the highly invasive nature of fistula surgery, misclassification due to undetected cases performed as outpatient procedures should be of limited consequence. The high quality of the registers reduces the risk of misclassification of both exposure (hysterectomy) and the outcome (fistula surgery), whereas exclusion of all fistula cases occurring before study entry effectively increases the internal validity. The nationwide, large-scale study design, using prospectively collected data over three decades, further strengthens our conclusions.
Hysterectomy remains the treatment of choice for many benign gynecologic conditions for which the surgical indications are relative. This study demonstrates that surgical removal of the uterus may be associated with sequelae in the form of pelvic organ fistula. Although rarely encountered in a general population, pelvic organ fistula disease may have devastating effects on all aspects of quality of life and should be considered in surgical decision making.
1. Symmonds RE. Incontinence: vesical and urethral fistulas. Clin Obstet Gynecol 1984;27:499–514.
2. Lee RA, Symmonds RE, Williams TJ. Current status of genitourinary fistula. Obstet Gynecol 1988;72:313–9.
3. Livengood CH 3rd. Colovaginal fistula. Report of a case with failure of transvaginal repair. J Reprod Med 1996;41:291–3.
4. Eilber KS, Kavaler E, Rodriguez LV, Rosenblum N, Raz S. Ten-year experience with transvaginal vesicovaginal fistula repair using tissue interposition. J Urol 2003;169:1033–6.
5. Kelly J, Winter HR. Reflections on the knowledge base for obstetric fistula. Int J Gynaecol Obstet 2007;99:S21–4.
6. Harkki-Siren P, Sjoberg J, Tiitinen A. Urinary tract injuries after hysterectomy. Obstet Gynecol 1998;92:113–8.
7. Carlson KJ, Nichols DH, Schiff I. Indications for hysterectomy. N Engl J Med 1993;328:856–60.
8. Lundholm C, Forsgren C, Johansson AL, Cnattingius S, Altman D. Hysterectomy on benign indications in Sweden 1987–2003: a nationwide trend analysis. Acta Obstet Gynecol Scand 2009;88:52–8.
9. Farquhar CM, Steiner CA. Hysterectomy rates in the United States 1990–1997. Obstet Gynecol 2002;99:229–34.
10. Whiteman MK, Hillis SD, Jamieson DJ, Morrow B, Podgornik MN, Brett KM, et al. Inpatient hysterectomy surveillance in the United States, 2000–2004. Am J Obstet Gynecol 2008;198:34.e1–7.
11. Coulter A, McPherson K, Vessey M. Do British women undergo too many or too few hysterectomies? Soc Sci Med 1988;27:987–94.
12. Nilsson A, Spetz C, Carsjö K, Nightingale R, Smedby B. Reliability of the hospital registry. The diagnostic data are better than their reputation [in Swedish]. Läkartidningen 1994;91:603–5.
13. Armenakas NA, Pareek G, Fracchia JA. Iatrogenic bladder perforations: longterm followup of 65 patients. J Am Coll Surg 2004;198:78–82.
14. Flynn MK, Amundsen CL. Delayed presentation and successful repair of a recurrent vesicovaginal fistula after hysterectomy and primary abdominal repair. Int Urogynecol J Pelvic Floor Dysfunct 2004;15:53–5.
15. Hoffman MS, Lynch C, Lockhart J, Knapp R. Injury of the rectum during vaginal surgery. Am J Obstet Gynecol 1999;181:274–7.
16. Ng CC, Chern BS, Siow AY. Retrospective study of the success rates and complications associated with total laparoscopic hysterectomy. J Obstet Gynaecol Res 2007;33:512–8.
17. Johnson N, Barlow D, Lethaby A, Tavender E, Curr E, Garry R. Surgical approach to hysterectomy for benign gynaecological disease. The Cochrane Database of Systematic Reviews 2005, Issue 1. Art. No.: CD003677. DOI: 10.1002/14651858.CD003677.pub3.
18. Chou DC, Rosen DM, Cario GM, Carlton MA, Lam AM, Chapman M, et al. Home within 24h of laparoscopic hysterectomy. Aust N Z J Obstet Gynaecol 1999;39:234–8.
19. Garry R, Fountain J, Brown J, Manca A, Mason S, Sculpher M, et al. EVALUATE hysterectomy trial: a multicentre randomised trial comparing abdominal, vaginal and laparoscopic methods of hysterectomy. Health Technol Assess 2004;8:1–154.
20. Johnson N, Barlow D, Lethaby A, Tavender E, Curr L, Garry R. Methods of hysterectomy: systematic review and meta-analysis of randomised controlled trials. BMJ 2005;330:1478–86.
21. Altman D, Granath F, Cnattingius S, Falconer C. Hysterectomy and risk for stress-urinary-incontinence surgery: nationwide cohort study. Lancet 2007;370:1494–9.
22. Altman D, Falconer C, Cnattingius S, Granath F. Pelvic organ prolapse surgery following hysterectomy on benign indications. Am J Obstet Gynecol 2008;198:572.e1–6.
23. Härkki-Siren P, Sjöberg J, Kurki T. Major complications of laparoscopy: a follow-up Finnish study. Obstet Gynecol 1999;94:94–8.