Hysterectomy on benign indication and risk of pelvic organ prolapse surgery: A national matched cohort study

Abstract

Introduction

Hysterectomy is a frequently performed gynecological procedure but long‐term effects remain understudied. Pelvic organ prolapse reduces life quality significantly. The lifetime risk of undergoing pelvic organ prolapse surgery is 20% and parity is known to be the largest risk factor. Studies have shown an increased risk of pelvic organ prolapse surgery after hysterectomy; however, few have studied the compartments which are affected and how this association is affected by surgical route and parity.

Material and methods

In this Danish nationwide cohort study, we identified women born in 1947–2000 who underwent hysterectomy during 1977–2018 who were indexed on the day of hysterectomy. We excluded women who immigrated when older than 15 years, who underwent pelvic organ prolapse surgery prior to index, and who were diagnosed with a gynecological cancer prior to or within 30 days of index. Women who underwent hysterectomy were matched 1:5 to references on age and year of hysterectomy. Women were censored at the time of death, emigration, a gynecological cancer diagnosis, radical or unspecified hysterectomy or December 31, 2018, whichever came first. The risk of pelvic organ prolapse surgery after hysterectomy was computed using Cox proportional hazard ratios (HRs) with 95% confidence intervals (CIs), adjusted for age, calendar year, parity, income and educational level.

Results

We included 80 444 women who underwent hysterectomy and 396 303 reference women. Women who underwent hysterectomy had a significantly higher risk of undergoing pelvic organ prolapse surgery: HR_adjusted 1.4 (95% CI 1.3–1.5). In particular, the risk of a posterior compartment prolapse operation was increased: HR_adjusted 2.2 (95% CI 2.0–2.3). The risk of prolapse surgery increased with increased parity and by an additional 40% after hysterectomy. Cesarean sections did not seem to increase the risk of prolapse surgery.

Conclusions

This study shows that hysterectomy, regardless of surgical route, leads to an increased risk of pelvic organ prolapse surgery, especially in the posterior compartment. The risk of prolapse surgery increased with the number of vaginal births, and not cesarean sections. Women should be thoroughly informed about the risk of pelvic organ prolapse and other treatment options should be considered before choosing hysterectomy to treat benign gynecological diseases –particularly women who have had numerous vaginal births.

This study shows that hysterectomy, regardless of surgical route, leads to an increased risk of pelvic organ prolapse surgery, especially in the posterior compartment. The risk of prolapse surgery increased with the number of vaginal births, but not cesarean sections. Women should be thoroughly informed about risk of pelvic organ prolapse and other treatment options should be considered before choosing hysterectomy to treat benign gynecological diseases—particularly women with numerous vaginal births.

Abbreviations

CI

confidence intervals

HR

hazard ratio

POP

pelvic organ prolapse

Key message.

Women undergoing hysterectomy, regardless of route, have an increased risk of pelvic organ prolapse surgery, especially in the posterior compartment. The risk further increases with increased parity, but only among those with vaginal births.

1. INTRODUCTION

Hysterectomy on benign indication is one of the most frequently performed surgeries, but it may be associated with long‐term complications. The rates of hysterectomies in different countries vary but have remained high over the last decade, with around 180/100 000 women years in Denmark. ¹ , ² Previous studies have shown an increased risk of pelvic organ prolapse (POP) after hysterectomy. ³ , ⁴ POP is a benign yet complicating medical condition which may cause discomfort, urinary tract symptoms, fecal incontinence as well as decreased sexual function. ⁵ Women with POP are often treated surgically and a Danish nationwide study showed that the lifetime risk of undergoing POP surgery was 18.7%. ⁶ A study has predicted that the number of Americans suffering from a symptomatic POP condition will increase by 46%–200% towards 2050. ⁷ While increasing age and obesity are important risk factors for POP, ⁵ the largest known risk factor for developing POP is parity. ⁸ However, little is known about the type of delivery in relation to this increased risk.

In a recent study we showed an increased risk of POP surgery among nulliparous women undergoing hysterectomy. ⁹ In the present nationwide cohort study, our aim was to support this finding in a larger cohort and furthermore investigate how parity and surgical routes affect this association and in which compartment, anterior or posterior, the POP develops. We hypothesized that hysterectomy increases the risk of POP surgery and that vaginal deliveries would further increase this risk.

2. MATERIAL AND METHODS

2.1. Setting

The Danish healthcare system is tax‐based and provides healthcare for all residents free of charge. ¹⁰ , ¹¹ At birth or immigration, all residents in Denmark are registered in The Danish Civil Registration System and assigned a unique 10‐digit personal identifier. This allows for exact individual linkage between data sources and epidemiological studies with lifelong follow‐up.

2.2. Data sources

Data on death and migration is provided from the Civil Registration System. ¹² Health data are provided through The Danish National Patient Registry (DNPR). ¹³ This includes information on diagnoses and surgeries on somatic hospital contacts, on inpatient admissions since 1977 and outpatient contacts since 1995. Surgeries were coded according to the Danish Classification of Surgical Procedures and Therapies (1st through 3rd edition) from 1977–1995 and since 1996 according to the Nordic Medico‐Statistical Committee (NOMESCO) classification of Surgical Procedures. Specific diagnostic and surgical codes are provided in the Appendix S1.

Information on parity was retrieved from The Fertility Database (FTDB), established in 1993. ¹⁴ This database provides virtually complete coverage on fertility in women born from 1945 onwards. From Statistics Denmark we collected information on disposable income (from 1980) and educational level (from 1981) for the entire population. ¹⁵ , ¹⁶

2.3. Study population

The study design (Figure 1) was inspired by the study performed for POP surgery among nulliparous women undergoing hysterectomy. ⁹ The source population was women born in 1947–2000 and living in Denmark at some point during 1977–2018. Women who immigrated after the age of 15 or who had twins or more were excluded. Among the source population we identified and included women undergoing hysterectomy for benign indication via the surgical routes of either total abdominal (open or laparoscopic), sub‐total abdominal (open or laparoscopic), vaginal or laparoscopic assisted vaginal hysterectomy.

FIGURE 1.

Study design flow chart.

For each woman undergoing hysterectomy we sampled five women from our source population, matching age (integer year) and calendar year of hysterectomy using the risk set sampling principle. ¹⁷ The index date was the date of hysterectomy for women who underwent hysterectomy and for women in the reference group the same date as their matched counterparts’ date of hysterectomy. We excluded women who underwent POP surgery prior to the index time and women who were diagnosed with a gynecological cancer prior to the index time or within 30 days after hysterectomy. Women in both groups were followed until POP surgery, time of death, emigration, a gynecological cancer diagnosis, radical or unspecified hysterectomy, or December 31, 2018, whichever came first. Furthermore, women in the reference group undergoing hysterectomy on a benign indication were censored and from then on included in the cohort of women undergoing hysterectomy. ¹⁷

2.4. Statistical analyses

Exposure was defined as women undergoing hysterectomy for benign indication without concomitant POP surgery. Our primary outcome was the first POP surgery in any compartment (apical, anterior or posterior) after the index date. Secondary outcomes were first POP surgery in anterior compartment and posterior compartment, respectively. In Denmark, hysterectomies to treat POP are mainly performed vaginally. We registered vaginal hysterectomy performed concomitantly with POP surgery as an outcome (POP surgery) and not as an exposure (hysterectomy). Vaginal hysterectomy without concomitant POP surgery was registered as an exposure (hysterectomy).

All women were assigned an income quintile. This was based on a reference calculated on all women 18 years or older living in Denmark within the age groups of 18–27, 28–40, 41–65 and ≥66 years based on income and age at the time of the index date. Education according to the highest completed level of education on the index date was grouped into high (more than a high school degree), intermediate (high school or vocational education), low (less than a high school degree) or unknown. Missing data was addressed as independent categories.

Comparison of the groups was evaluated using t‐test for follow‐up time and Chi‐square test for trend (disposable income and educational level).

Data were analyzed using Cox proportional hazard regression and presented as hazard ratios (HRs) with 95% confidence intervals (CIs). We performed crude analyses and analyses adjusted for age, calendar year, parity, income and educational level. Furthermore, crude analyses were stratified for age (<40 years, 40–47, >47) and calendar year at index time (before 1998, 1998–2006, after 2006).

We reiterated all analyses in the following subgroups: surgical route of hysterectomy (total abdominal, sub‐total abdominal, vaginal or laparoscopic assisted vaginal hysterectomy), parities (0, 1, 2, 3, ≥4), only vaginal births (0, 1, 2, 3, ≥4) and only cesarean sections (0, ≥1). We did not distinguish between elective and emergency cesarean sections.

Finally, we repeated all analyses for first anterior or posterior compartment surgery, respectively, as outcomes.

A p‐value <0.05 was considered significant. All calculations were done using STATA, release 16 (StataCorp.), except the Chi‐square tests for trend, which were performed with EpiInfo, StatCalc.

The study followed the STROBE guidelines for reporting on epidemiological studies. ¹⁸

2.5. Ethics statement

According to the Danish law, ethical approval is not required for non‐intervention register‐based studies. This study was approved by the Capital Region of Denmark, rec. no. P‐2020‐683.

3. RESULTS

We identified 80 444 women who underwent hysterectomy for benign indications who were matched to 396 303 reference women. The groups of women were comparable according to age and follow‐up (Table 1). The group of women who underwent hysterectomy generally had a higher level of education and income.

TABLE 1.

Baseline characteristics of the study population.

	Women undergoing hysterectomy	Reference women	Total
n	n = 80 444	n = 396 303	n = 476 747
Follow‐up time, median years (Q1 – Q3)a	14.98 (7.87–22.06)	14.05 (7.05–21.27)	14.52 (7.16–21.38)
Age, mean years (SD)	43.55 (7.02)	43.50 (7.02)	43.53 (7.02)
Income quintile (%)*
1 (Highest)	9983 (12.4)	43 734 (11.0)	53 717 (11.3)
2	16 061 (20.0)	68 264 (17.2)	84 325 (17.7)
3	18 589 (23.1)	85 147 (21.5)	103 736 (21.8)
4	19 061 (23.7)	94 515 (23.6)	113 576 (23.8)
5 (lowest)	15 614 (19.4)	98 556 (24.9)	114 170 (24.0)
Unknown	1136 (1.4)	6087 (1.5)	7233 (1.5)
Education level (%)*
High	26 741 (33.2)	106 975 (27.0)	133 716 (28.1)
Intermediate	35 374 (44.0)	175 044 (44.2)	210 418 (44.1)
Low	16 493 (20.5)	105 480 (26.6)	121 973 (25.6)
Unknown	1836 (2.3)	8804 (2.2)	10 640 (2.2)

Abbreviations: SD, standard deviation; Q3, 75% quartile; Q1, 25% quartile.

^* p < 0.005.

In the group of women who underwent hysterectomy, 3.6% (n = 2874) underwent POP surgery, compared with 2.4% (n = 9341) in the reference group. The Cox proportional hazard regression showed that women who underwent hysterectomy had a significantly higher risk of undergoing POP surgery than women in the reference group did: HR_crude 1.44 (95% CI 1.37–1.50) and HR_adjusted 1.40 (95% CI 1.34–1.47) (Table 2). No substantial differences were found after stratification for age and calendar year (data not shown). We saw a significantly increased risk in the posterior compartment (HR_adjusted 2.16, 95% CI 2.03–2.29) and only a small difference in the anterior compartment (HR_adjusted 1.08, 95% CI 1.02–1.15) (Figure 2).

TABLE 2.

Hysterectomy and subsequent risk of pelvic organ prolapse surgery by compartment.

Any compartment (apical, anterior, or posterior)
Index operation	Number of women undergoing POP surgery after index date (%)	Median time of risk (years)	Incidence rates (per 100 000 person‐years)	HR crude [95% CI]	HR adjusted a [95% CI]
Reference group	9341 (2.4)	14.05	158.37	1 (Ref.)	1 (Ref.)
Hysterectomized women	2874 (3.6)	14.98	229.11	1.44 [1.37–1.50]	1.40 [1.34–1.47]
Anterior compartment b
Reference group	6853 (1.7)	14.12	115.71	1 (Ref.)	1 (Ref.)
Hysterectomized women	1654 (2.1)	15.17	130.56	1.12 [1.05–1.18]	1.08 [1.02–1.15]
Posterior compartment b
Reference group	4208 (1.1)	14.16	70.87	1 (Ref.)	1 (Ref.)
Hysterectomized women	1966 (2.4)	15.12	155.69	2.19 [2.06–2.32]	2.16 [2.03–2.29]

Abbreviations: CI, confidence interval; HR, hazard ratio.

^a Adjusted for age, calendar year, parity, income and educational level.

^b Analysis for each compartment was made separately, as some women had both an anterior and posterior prolapse operated simultaneously and therefore were included in both analyses.

FIGURE 2.

Cumulative incidence of pelvic organ prolapse surgery.

In the sub‐analyses we found similar risks of undergoing POP surgery according to the different surgical routes of hysterectomies apart from laparoscopic assisted vaginal hysterectomy, which was insignificant (Table 3).

TABLE 3.

Different surgical routes of hysterectomy and subsequent risk of pelvic organ prolapse surgery.

Procedure	Number of women	Pelvic organ prolapse surgeries	Incidence rate	HRcrude [95% CI]	HRadjusted a [95% CI]
Reference group	396 303	9341	158.37	1 (Ref.)	1 (Ref.)
Total hysterectomy b	58 264	2074	224.61	1.38 [1.31–1.45]	1.38 [1.32–1.46]
Subtotal hysterectomy b	11 273	474	244.76	1.56 [1.42–1.72]	1.55 [1.41–1.71]
Vaginal hysterectomy	8471	267	254.15	1.80[1.59–2.03]	1.41 [1.24–1.59]
Laparoscopic assisted vaginal hysterectomy	2436	59	182.56	1.25 [0.97–1.62]	1.11 [0.86–1.44]

Abbreviations: CI, confidence interval; HR, hazard ratio. Incidence rates are per 100 000 person years.

^a Adjusted for age, calendar year, vaginal deliveries, income and educational level.

^b Performed abdominally with either laparoscopy or laparotomy.

We saw an increased risk of POP surgery among parous women as compared with nulliparous women. The higher the number of parities, the more the risk increased. Within each parity (1, 2, 3, ≥4), women undergoing hysterectomy had a 40% increased risk of POP surgery compared with women with their uterus in situ (Table 4). For women with four or more births previously, a hysterectomy will increase the risk of POP surgery by 40%, from 21 times to an increased risk of 29 times compared with nulliparous women with their uterus in situ. The increased risk of POP surgery with parity was strongly associated with vaginal births, whereas cesarean sections did not seem to increase the risk of POP surgery.

TABLE 4.

Hysterectomy and subsequent risk of pelvic organ prolapse surgery by increasing parity.

	Women undergoing hysterectomy		Reference group
Parity	Number of women (% of women undergoing POP surgery)	Hazard ratio [95% CI]	Number of women (% of women undergoing POP surgery)	Hazard ratio [95% CI]
All women
0	9545 (0.30)	1.62 [1.03–2.57]	50 178 (0.18)	1 (ref.)
1	12 980 (2.12)	10.15 [7.49–13.75]	65 629 (1.46)	7.49 [5.60–10.02]
2	37 831 (3.97)	18.70 [14.07–24.86]	187 810 (2.62)	13.36 [10.07–17.71]
3	15 863 (5.28)	25.91 [19.42–34.58]	73 040 (3.52)	18.52 [13.94–24.61]
≥4	4225 (5.42)	28.50 [20.91–38.85]	19 646 (4.08)	21.03 [15.69–28.19]
Vaginal births only
1	11 084 (2.38)	10.98 [8.09–14.90]	55 895 (1.69)	8.43 [6.30–11.28]
2	32 193 (4.45)	20.21 [15.20–26.88]	161 761 (2.90)	14.39 [10.85–19.08]
3	12 874 (6.02)	28.45 [21.30–37.99]	61 629 (3.90)	19.95 [15.01–26.52]
≥4	3233 (6.21)	30.84 [22.53–42.21]	16 022 (4.53)	22.64 [16.86–30.39]
Cesarean sections only
≥1	4384 (0.41)	3.14 [1.79–5.50]	19 628 (0.10)	0.75 [0.38–1.48]

4. DISCUSSION

This nationwide cohort study shows that women undergoing hysterectomy have a higher risk of undergoing POP surgery than women with their uterus in situ. We found similar risks across different surgical routes, which could indicate that it is not the method of hysterectomy but the absence of the uterus per se which increases the risk of POP.

The increased risk of POP surgery after undergoing hysterectomy was supported by previous studies. ³ , ⁴ Our study specifically looked at prolapse by compartment and found that hysterectomy was associated with an increased risk of POP surgery in the posterior compartment, whereas the anterior compartment was only slightly affected. This result is supported by the literature, indicating that a uterus in situ protects against POP surgery, primarily in the posterior compartment. ¹⁹ , ²⁰ , ²¹

A Swedish and a Taiwanese study showed an increased risk of POP surgery after vaginal hysterectomy compared with other surgical routes. ³ , ⁴ Conversely, we found vaginal hysterectomy and risk of prolapse similar to other surgical routes. Vaginal hysterectomy is used as a treatment for both POP and other benign gynecological diseases. ²² , ²³ Consequently, vaginal hysterectomy constitutes an important methodological problem, as it can be seen as both an exposure and an outcome. In the Swedish and Taiwanese studies, women undergoing POP surgery concomitant with hysterectomy were excluded, which may have reduced the number of outcomes. ³ , ⁴ This could lead to an overestimated difference between the study group and reference group. In our study a vaginal hysterectomy without concomitant POP surgery was classified as a hysterectomy (exposure), whereas a vaginal hysterectomy with concomitant POP surgery was classified as a POP surgery (outcome) in the reference group. Furthermore, when we repeated our analyses by excluding all vaginal hysterectomies, the results were materially the same (data not shown). In Denmark, laparoscopic assisted vaginal hysterectomy is not frequently performed and we were not able to draw any conclusions about this surgical route (Table 3).

This result highlights the need for improved prophylactic surgical techniques concomitant with hysterectomy in the posterior compartment specifically. Suspension of the vaginal vault during hysterectomy is one possible area which has been researched. ²⁴

While the absolute risk of POP surgery for nulliparous women who have undergone hysterectomy is low, ⁹ the risk increased considerably with higher parity, as shown previously. ³ Our results reinforced this association and showed that the increased risk was constant for all parities (1, 2, 3, ≥4). From our results we saw that the association between parity and increased risk of prolapse was associated with vaginal births and not with cesarean sections (Table 4). This indicates that it is the vaginal delivery, and not the pregnancy itself, that is associated with an increased risk. A recent Swedish study looking at pregnancy, parity and mode of delivery has shown a similar result. ²⁵

The strengths of the study include the use of nationwide health registers of high quality with complete follow‐up, ¹³ which has been shown to be efficient for studying POP. ²⁶ Secondly, we observed the women in our study for up to 42 years. Thirdly, we were able to adjust for parity, income and educational level. Fourthly, the large cohort enabled several sub‐analyses such as assessing which compartment was most affected after hysterectomy and analyzing for cesarean sections and vaginal births separately. Lastly, we were able to differentiate between vaginal hysterectomy as an exposure and as an outcome.

This study also had limitations worth considering. First, we did not have clinical assessments such as POP‐Q, BMI, smoking habits, alcohol consumption or use of pessary. ⁵ , ²⁷ , ²⁸ , ²⁹ Women who underwent hysterectomy are more likely to experience pessary failure and might thereby more likely undergo POP surgery. Secondly, the number of women with non‐surgically treated symptomatic POP might be substantially larger. Thirdly, there could be a potential health‐seeking bias between the two groups, with hysterectomized women more likely to undergo a second surgery, thereby overestimating the differences between the groups. This was, however, addressed in our earlier study and we found no difference in health‐seeking behavior between women who underwent hysterectomy and the reference group. ⁹ Fourthly, the lack of information regarding indication for hysterectomy is a potential bias. Two previous studies found that hysterectomy performed on POP indication was associated with a higher rate of subsequent POP surgeries compared with hysterectomy performed on other indications. ²⁰ , ³⁰ We attempted to avoid this potential bias by distinguishing between vaginal hysterectomies performed concomitant with a POP operation (reflecting POP indication) and vaginal hysterectomies performed without POP operation (reflecting other indications). Lastly, we were not able to assess the risk of surgery in the apical compartment due to the non‐comparability of surgical opportunities, as well as operation codes for women with or without the uterus in situ.

5. CONCLUSION

Our study reinforces that hysterectomy on benign indication is a risk factor for POP surgery. The risk was highest in the posterior compartment and was regardless of surgical route. The risk increases with increased parity but was only associated with the number of vaginal births, not cesarean sections. Considering the relative indications for benign hysterectomy, doctors and patients should be aware of the risk of POP after hysterectomy and other treatment options should be carefully discussed, especially among women with numerous vaginal births.

AUTHOR CONTRIBUTIONS

AHA, main author, planned the methodology, analyzed the data, interpreted the results and wrote the paper. NK formulated the research idea, planned the methodology, interpreted the results and revised the paper. KOG planned the methodology, data analysis, interpretation of the results and revised the paper. KRH formulated the research idea, planned the methodology, carried out the data analysis, interpreted the results and revised the paper.

CONFLICT OF INTEREST STATEMENT

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

Supporting information

Appendix S1.

Aagesen AH, Klarskov N, Gradel KO, Husby KR. Hysterectomy on benign indication and risk of pelvic organ prolapse surgery: A national matched cohort study. Acta Obstet Gynecol Scand. 2023;102:774‐781. doi: 10.1111/aogs.14561