Risk factors for primary pelvic organ prolapse and prolapse recurrence: an updated systematic review and meta-analysis

e 2.62), and levator defect (n ¼ 2, odds ratio, 3.99; 95% conﬁdence interval, 2.57 e 6.18) are statistically signiﬁcant risk factors, and cesarean delivery (n ¼ 2, pooled odds ratio, 0.08; 95% conﬁdence interval, 0.03 e 0.20) and smoking (n ¼ 3, odds ratio, 0.59; 95% conﬁdence interval, 0.46 e 0.75) are protective factors for primary prolapse. Parity, vaginal delivery, and levator hiatal area are identiﬁed as “conﬁrmed risk factors.” For prolapse recurrence, preoperative prolapse stage (n ¼ 5, odds ratio, 2.68; 95% conﬁdence interval, 1.93 e 3.73) and age (n ¼ 2, odds ratio, 3.48; 95% conﬁdence interval, 1.99 e 6.08) are statistically signiﬁcant risk factors. CONCLUSION: Vaginal delivery, parity, birthweight, age, body mass index, levator defect, and levator hiatal area are risk factors, and cesarean delivery and smoking are protective factors for primary prolapse. Preoperative prolapse stage and younger age are risk factors for prolapse recurrence after native tissue surgery.


Introduction
Pelvic organ prolapse (POP) is a common medical condition worldwide impairing many women in their daily life. Although POP is not a lifethreatening disease, it has a significant impact on the quality of life. 1 Studies show that women have a lifetime risk of 12.6% to undergo surgical correction for POP by the age of 80 years. 2 This number indicates not only the burden of POP on society and healthcare systems but also its financial impact on healthcare.
With increasing life-expectancy in general, it is estimated that the number of care-seeking women and surgeries will increase tremendously in the coming 20e40 years. These high rates for POP surgery demand a focus on preventive strategies. 3 The key to finding the right prevention strategies is knowledge about etiology and risk factors. With an eye on the emerging preventive medicine, several studies investigating the risk factors for POP development and POP recurrence after surgery have been carried out. This knowledge about risk factors not only contributes to developing prevention strategies but also helps in the counseling of patients preoperatively and managing expectations. The systematic review by Vergeldt et al 4 identified parity, vaginal delivery, age, and body mass index (BMI) as confirmed risk factors for the development of POP and preoperative stage 3 and 4 as confirmed risk factors for POP recurrence after native tissue repair (on the basis of definition in !2 studies with significant association in multivariable analysis). In the years after this publication, multiple studies have been published on this subject. Among others, the meta-analysis of Cattani et al identified forceps delivery and first vaginal birth as risk factors for anatomic and symptomatic primary POP. 5 For POP recurrence, the meta-analysis of Friedman et al showed that levator defect, preoperative prolapse stage 3 or 4, family history of prolapse, and levator hiatal area are significant risk factors for POP recurrence. 6 In this paper, we will update the review of Vergeldt et al and perform a meta-analysis not only on the risk factors for primary POP but also on POP recurrence for women in the Western developed countries. 7

Methods
This systematic review and metaanalysis was conducted in accordance with a prospectively registered protocol (International Prospective Register of Systematic Reviews [PROSPERO]; PROSPERO number CRD42021230813, March 26, 2021), the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines, and the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) guidelines. 8,9 Information sources and search strategy A database search was performed by the primary reviewer (S.F.S.) and a librarian in PubMed and Embase using the search terms "pelvic organ prolapse" AND "recurrence" and "pelvic organ prolapse" AND "risk factors." The search for the previous publication ended on August 4, 2014. Therefore, we searched from July 1, 2014 until July 5, 2021. The same search terms were used. No language restrictions were used. For the complete search, see appendix A.

Study selection and eligibility criteria
We used the same evaluation strategy as in the previous review. All the articles were evaluated by title and/or abstract by 2 independent reviewers (S.F.S. and M.C.). In case of disagreement, a third reviewer (K.B.K.) solved conflicts by consensus. Clinical studies reporting on the etiology or risk factors for primary POP or POP recurrence were included. A manual reference check of the included abstracts was performed. The included articles after abstract selection were screened on full text with a standardized in-and exclusion form. The authors were contacted to retrieve the article in case the full text was not available. Randomized controlled trials, crosssectional and cohort studies conducted in the Western developed countries 7 that reported on multivariable analysis with sufficient data (including odds, risk, or hazard ratio [HR] with 95% confidence intervals) of risk factors for POP or POP recurrence were included.
The definition of POP or POP recurrence had to be based on anatomic references or POP-Quantification (POP-Q) ! stage 2. For POP recurrence, only studies that reported on recurrence after native tissue repair with a median follow-up of at least 1 year were included. In case studies used the same population in multiple publications, only the most recent publication was included.

Data extraction
Data extraction was conducted by 2 reviewers (S.F.S. and M.C.) using a predefined data extraction form with data on study design, sample size, study population, definition of outcome, investigated risk factors, and results of the multivariable analysis. The corresponding authors were contacted in case additional information was needed on the study results. To provide a comprehensive overview, the results of the previous review were used again in this paper. The template data collection forms and data extracted from included studies are available on request.
Assessment of risk of bias A quality assessment was performed by 2 independent reviewers (S.F.S. and M.C.) on the final included articles using the Newcastle-Ottawa Scale (NOS) for cross-sectional and cohort studies comprising of the following: participant selection, comparability of study groups, and assessment of outcome or exposure. 10

Data synthesis
In case a risk factor was studied in at least 2 studies using the same type of outcome and adjusted for at least the following confounders: parity, delivery mode, age, and BMI for primary POP and preoperative POP-Q stage for POP recurrence, we pooled the adjusted results with a random-effects meta-analysis using the inverse variance method on the logtransformed ratios and corresponding standard errors and presented the 95% confidence intervals of the backtransformed ratios. If necessary and possible, data conversion was applied (eg, conversion of per 1 year to per 10 years). In the case of a similar outcome but on the basis of different sets of adjustment variables, the results were only pooled in case of sufficiently low between-study heterogeneity (I 2 <50%). Variation across studies (heterogeneity) was estimated with a restricted maximum likelihood estimator for s 2 . If studies could be pooled, an extra line in the forest plot below the studies was added to present the pooled result of the meta-analysis in bold. If the effects of a risk factor were presented in different measures, eg, odds ratio (OR) and HR, these were not pooled but were presented graphically in forest plots separated by effect measure. In addition to the risk factors identified by metaanalyses, we identified "confirmed risk factors" also. A confirmed risk factor was defined as a statistically significant association on the basis of multivariable data analysis that was reported in at least 2 studies that could not be pooled because of heterogenic outcome definitions or effect measures, without other studies reporting contradicting results. This definition was based on the definition used in the previous publication. No subgroup or sensitivity analyses were performed because of the small number of studies per potential risk factor and the large heterogeneity in risk factors. Publication bias was not evaluated, as the meta-analyses were based on 5 studies each at the most. All analyses were performed with the statistical software R version 3.6.3, 11 packages meta 12 version 2.4-0, and forest plot 13 version 1.10.1. The Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach was used to interpret the certainty in the body of the evidence. 14,15 As publication biases were not evaluated because of the small number of available studies per risk factor, the certainty of evidence was not downgraded for this domain.

Study selection
A total of 5284 articles were retrieved by our search update. After the removal of duplicates, 3381 articles were screened by title and/or abstract. The full texts of 112 articles were evaluated using the inand exclusion form. No extra articles were included after cross checking reference lists. After final selection, an additional 14 articles met our inclusion criteria, of which 8 articles were on the risk factors for primary POP and 6 articles were on the risk factors for POP recurrence. One article was excluded in the previous review and now included because of exclusion of an older study with the same population. 16 Three articles that were included in the previous review were now partly 17 or totally 18,19 excluded, because they used the same study population in a more recent publication or the country of investigation was not a Western developed one. 7 Three studies appeared to meet the inclusion criteria but were excluded, because no separate analysis was performed for anatomic POP recurrence. 20e22 In total (with the included articles of the previous publication), we included 16 articles on primary POP and 11 articles on POP recurrence. Figure 1 shows the flow diagram of the selection process. Because of high heterogeneity or differences in definitions and effect measures, not all studies could be pooled. Forest plots were made to visualize the results and to be able to recognize trends; see Figures 2e8. The results of the studies that could not be included in the forest plots are listed in the tables; see appendix B.

Study characteristics
Studies on primary pelvic organ prolapse The characteristics of the studies concerning the risk factors for primary POP are summarized in Table 1. In total, data on 43,333 women were analyzed in 8 prospective cohort studies and 8 crosssectional studies. POP was defined as POP-Q stage 2 or higher in 7 studies, 23e29 POP beyond the hymen in 5 studies, 30e34 degree 2 or 3 in the Baden-Walker classification in 1 study, 35 the most descended point of the vaginal wall to the introitus or outside of the vagina (according to the Women's Health Initiative classification system) in 1 study, 36 the most dependent point of the vaginal wall or the cervix to or beyond the hymen 16 and the most descended point of the vaginal wall À0.5 cm above the hymenal remnants in 1 study. 37

Studies on pelvic organ prolapse recurrence
The characteristics of the studies concerning POP recurrence are summarized in Table 2. In total, data on 2132 women were analyzed in 4 retrospective and 6 prospective cohort studies and 1 secondary analysis of an RCT. POP recurrence was defined as POP-Q stage 2 or higher in 10 studies 38e46 and as descent of the vaginal apex (point C of the POP-Q system) more than one-third into the vaginal canal or anterior or posterior vaginal wall descent beyond the hymen in 1 study. 47

Risk of bias of included studies
Studies on primary prolapse The overall quality of the articles was adequate; all the included articles had a sufficient description of in-and exclusion criteria and outcomes. In 15 articles, the number of risk factors for analysis was limited to 10% of the number of events; in 1 article, the 10% was exceeded. 33 Blinding was applied in 9 articles. Quality assessment showed NOS scores of 5 studies being 9, 24,25,27,28,34 that of another 5 studies being 8, 26,30,31,35,37 of 3 studies being 7, 29,33,36 of 1 study being 6, 16 and of 2 studies being 5. 23,32 A score of 7 or higher is considered high quality.

Studies on prolapse recurrence
The overall quality of the articles was adequate; all the included articles had a sufficient description of the in-and exclusion criteria and outcomes, and the median follow-ups of the studies varied between 1 and 12 years. In 2 studies, selection bias because of selective loss to follow-up could not be ruled out, because both studies reported >50% loss to follow-up without further reporting a comparison between the groups. 45,46 Three studies did not apply the limitation of the number of risk factors to be 10% of the number of events. 38

Obstetrical factors
Parity was reported by 7 studies, of which 2 reported parity as a continuous variable (per 1) and 5 as categorical. For parity as a categorical variable, a parity of 2 or higher compared with 0 or 1 was a significant risk factor in 4 studies. 30,34e36 The categorical variables for parity could not be pooled because of differences in effect measures. Therefore, it is identified as a confirmed risk factor. The pooled OR for parity per 1 was not statistically significant (n¼2, OR, 1.06; 95% CI, 0.39e2.89); 27,37 see Figure 2.
Age at first delivery was reported by 3 studies, of which 1 study 34 reported ages above 30 as a risk factor compared with age 24; see Figure 2.
Vaginal delivery was reported by 4 studies, of which 2 reported vaginal delivery as a continuous variable and 2 as a categorical variable. Compared with nulliparity, vaginal delivery was a significant risk factor in 2 studies 24,29 and could therefore be identified as a confirmed risk factor. The pooled OR for vaginal delivery (per 1) was not statistically significant (n¼2, OR, 1.33; 95% CI, 0.73e2.41); 24,37 Figure 3.
Forceps delivery was reported as a significant risk factor and as a significant protective factor for primary POP compared with normal vaginal delivery. 28,34 The results were pooled, because both studies corrected for the same confounders. The pooled OR was not statistically significant (n¼2, pooled OR, 1.05; 95% CI, 0.57e1.94). 28,34 In combination with levator defect, forceps delivery was a strong significant risk factor in 1 study. 31 One study reported on "operative vaginal delivery" (ie, forceps and vacuum delivery), which was a risk factor when compared with normal vaginal delivery. 30 No significant association was found between vacuum delivery and primary POP (n¼2, pooled OR, 0.88; 95% CI, 0.45e1.73); 28,34 Figure 3.
Cesarean delivery was reported by 5 studies, of which 3 compared it with normal vaginal delivery and reported a Forest plot and meta-analyses for primary POP in association with the obstetrical risk factors parity, birthweight, and age at first delivery For each study, the estimate, that is, odds ratio or hazard ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes total number of births, superscript letter b denotes data of univariable analysis.
For prolapse recurrence, parity per 1 was not statistically significant (n¼2, pooled OR, 0.96; 95% CI, 0.76e1.2). 17,43 Birthweight >4500 g was a significant risk factor for prolapse recurrence in 1 of the 2 studies. 39 Birthweight could not be pooled because of different definitions, but there seems to be a trend. For complicated delivery (ie, instrumental delivery with vacuum or forceps and/or a large vaginal laceration), only 2 of 3 studies could be pooled because of not correcting for preoperative POP-Q stage in 1 study. 42 The pooled OR for complicated delivery was not statistically significant (n¼2, pooled OR, 0.90; 95% CI, 0.34e2.37); 39,44 Figure 4.
Lifestyle factors BMI as a risk factor for primary POP was reported in 8 studies. Higher BMI as a categorical variable was a significant risk factor for primary POP in 4 studies 29,35e37 , and 2 studies showed no statistically significant association. 30,34 The pooled ORs for BMI 25e30 vs <25 kg/m 2 and !30 vs <25 kg/m 2 were statistically significant (OR, 1.52; 95% CI, 1.07e2.15 34,37 and OR, 1.75; 95% CI, 1.17e2.62, 34,37 respectively). One study showed BMI !30 kg/m 2 to be a statistically significant protective factor compared with that <25 kg/m 2 , but an index of 25e30 kg/m 2 was not significant when compared with <25 kg/m 224 ( Figure 5).
Physical activity was reported by 3 studies, of which 1 reported a borderline significant effect of more activity to be a risk factor for primary POP. 33 The results of the studies could not be pooled because of differences in definitions ( Figure 5).

Unmodifiable factors
Age per 10 years was a statistically significant risk factor for primary POP in 3 Forest plot and meta-analyses for primary POP in association with the obstetrical risk factor delivery mode For each study, the estimate, that is, odds ratio, hazard ratio, or prevalence ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes exclusive vaginal delivery, superscript letter b denotes delivery mode not further specified, superscript letter c denotes mixed delivery modes, superscript letter d denotes exclusive cesarean delivery.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022.
For ethnicity, 1 study showed Black ethnicity to be protective against POP, 36 and 3 studies showed no association. 29,30,37 Of the 2 studies that could be pooled, the OR showed a borderline significant but small effect for Black ethnicity to be protective against primary POP (n¼2, pooled OR, 0.96; 95% CI, 0.92e1.00), 29,37 Figure 6.
Menopausal status was reported by 2 studies. 25,37 It was not statistically associated with primary POP, and the results could not be pooled because of high heterogeneity, though a slight trend can be observed ( Figure 6).

Comorbidity
Hormone replacement therapy was reported in 2 studies and was only once positively associated with primary POP (Figure 7). 36 Urinary incontinence (UI) was reported by 2 studies, of which 1 reported mixed and urge UI as significant risk factors for primary POP (Figure 7). 36 Pulmonary disease was reported by 2 studies and was not associated with primary POP (Figure 7). 33,36 Hysterectomy status was reported by 2 studies and was not associated with primary POP (n¼2, pooled OR, 1.06; 95% CI, 0.73e1.54), 27,37 Figure 7.
Previous POP surgery was a significant risk factor for POP recurrence in 1 study, 42 but only the data of univariable analyses were available (Figure 8).

Social factors
Education was reported by 5 studies, of which 2 studies reported that a higher form of education is protective against primary POP, 24,35 but different definitions were used (Figure 7).

Surgical factors
The preoperative POP-Q stage was reported in 7 studies, of which 5 studies showed that preoperative stage III or IV was a statistically significant risk factor for POP recurrence when compared with stage II (n¼5, pooled OR, 2.68; 95% CI, 1.93e3.73). 38,40,41,44,45 One study reported that preoperative stage ! Forest plot and meta-analyses for POP recurrence in association with the risk factors parity, complicated delivery, birthweight, BMI, and age For each study, the estimate, that is, odds ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes data of univariable analysis.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022. Systematic Reviews ajog.org II was a significant risk factor when compared with stage < II (Figure 4). 39 Concomitant surgery was reported by 3 studies, of which 1 showed a borderline significant protective effect of posterior colporrhaphy 39 on developing POP recurrence and 1 showed sacrospinous hysteropexy to be a significant risk factor. 17 Pelvic floor factors For primary POP, 3 studies reported that levator defect was a statistically significant risk factor (n¼2, pooled OR, 3.99; 95% CI, 2.57e6.18), 27,31 Figure 7.
An increased levator hiatal area on Valsalva was a statistically significant risk factor for primary POP in 2 out of 2 studies. 23,27 The results could not be pooled because of differences in definitions, but it can be identified as a confirmed risk factor (Figure 7).
Regarding POP recurrence, levator defect was a statistically significant risk factor in the study by Weemhoff et al, 17 but in the combined study with the results of another database, 44 the OR was no longer statistically significant. Levator defect was not statistically associated with POP recurrence in the 2 studies in our meta-analysis, but the results of our meta-analysis did show a borderline significant effect (n¼2, pooled OR, 1.5; 95% CI 1.00e2.25), 40,44 Figure 8. Other pelvic floor factors were only investigated once; see appendix B.

GRADE certainty of the evidence
The GRADE approach was applied on the statistically significant pooled results. 14,15 The certainty of the evidence for primary POP was judged to be "very low" for BMI (OR, 25e30 vs <25 kg/m 2 ) and smoking (currently vs never). It was judged to be "low" for age, BMI (HR, 25e30 vs <25 kg/m 2 and OR/HR, !30 vs <25 kg/m 2 ), smoking (past vs never), and birthweight. The certainty of evidence was judged as "moderate" for cesarean delivery and levator defect. Regarding POP recurrence, the certainty of evidence for age and preoperative POP-Q stage was judged as "low." The evidence was downgraded for serious imprecision (small sample size) and Forest plot and meta-analyses for primary POP in association with the nonobstetrical risk factors BMI, smoking, and physical activity For each study, the estimate, that is, odds ratio or hazard ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown.
BMI, body mass index; I2, the estimated between-study heterogeneity; LL, lower limit of 95% confidence interval; MVPA, moderate to vigorous physical activity; POP, pelvic organ prolapse; UL, upper limit of 95% confidence interval.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022.
ajog.org inconsistency (because of high heterogeneity or effect estimates in contradicting directions and crossing the line of no effect). The evidence for cesarean delivery, levator defect, and age (for POP recurrence) were upgraded because of a large magnitude of effect (OR>3 or <0.3). Details about the GRADE assessment are provided in appendix 3.

Comment
Principal findings By updating the systematic review and performing meta-analyses, we were able to present a comprehensive overview of the currently available literature on the risk factors for primary POP and POP recurrence. The results of our metaanalyses show that age, BMI, birthweight, and levator defect are identified as statistically significant risk factors for primary POP and vaginal delivery, and parity and levator hiatal area are identified as confirmed risk factors for primary POP. Cesarean delivery and smoking are significant protective factors for primary POP. For POP recurrence, younger age and preoperative POP-Q stage 3 or 4 are statistically significant risk factors.

Comparison with existing literature
In the previous publication, risk factors were labeled as "confirmed risk factors" if the factors were significantly associated with POP or POP recurrence in a multivariable analysis in at least 2 studies. 4 In this current article, we supplemented the results by providing forest plots and meta-analyses. These forest plots gave more insight into several risk factors. For example, the forest plots showed a clear trend for a larger levator hiatal area to be a risk factor for primary POP, which was also labeled as a "confirmed risk factor." In addition, if a risk factor could not be identified as a confirmed risk factor, eg, levator defect for POP recurrence, the forest plot still illustrates a borderline significant effect of the pooled result. By providing comprehensive forest plots, we give more insight into the results, and the effect of potential risk factors that could not be pooled because of differences in definitions and effect measures (ie, odds ratios, prevalence ratios, and hazard ratios) can still be easily interpreted.
On the contrary, the forest plots also show inconsistencies between studies that used the same definitions for risk factors. This inconsistency is clearly present for the potential risk factor forceps delivery. In one study, it is significantly associated with developing primary POP, but in the other study, it is significantly protective against developing POP. Comparable inconsistencies were found in the meta-analysis by Leng et al 48 even though they analyzed vacuum and forceps delivery in a combined factor. The systematic review by Cattani et al 5 stated that forceps delivery is a risk factor for primary POP, but in contrast to our present review, they included studies that did not make a clear distinction between women with primary POP and women who already had a history of POP surgery. Although the results of our meta-analysis on forceps delivery were not conclusive in showing Forest plot and meta-analyses for primary POP in association with the nonobstetrical risk factors age, ethnicity, and menopausal status For each study, the estimate, that is, odds ratio, hazard ratio, or prevalence ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes data of univariable analysis. I 2 , the estimated between-study heterogeneity; LL, lower limit of 95% confidence interval; POP, pelvic organ prolapse; UL, upper limit of 95% confidence interval.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022. Systematic Reviews ajog.org a direct link between forceps delivery and POP, several studies outside the scope of this systematic review have identified forceps delivery as a risk factor for levator injury. 49 In turn, on the basis of our results, levator defect is identified as a risk factor for primary POP.
The results for BMI were also contradictory. For primary POP, it is a clear risk factor, but this could not be concluded for POP recurrence. The role of BMI on POP and POP recurrence has been investigated several times in the past decade. For primary POP, 2 studies reported BMI as a risk factor. 50,51 On the contrary, the systematic review by Zenebe et al 52 found no association between BMI and primary POP. This difference might be because of broader inclusion criteria for the definition of POP. For POP recurrence, Friedman et al 6 performed a systematic review and metaanalysis on BMI as a risk factor. They found that a higher BMI was not statistically significantly associated with POP recurrence, but a trend could be observed, which is in line with our results. In contrast, the large cohort study of Weltz et al 53 showed that BMI was a risk factor for reoperation in the anterior compartment. They found a trend for the apical and posterior compartment. We did not include studies with reoperation as an outcome measure, and therefore we cannot compare the results.
Obstetrical risk factors have been researched in multiple studies. Firstly, vaginal delivery seems to be an indisputable risk factor, which is confirmed by several reviews. 5,48,54e56 The literature on birthweight as a risk factor for POP has been contradictory so far. Our review shows that there is a significant association, which confirms the findings of the large study by Martinho et al. 57 They performed multivariable analysis for the effect of birthweight on levator defect, which was statistically significant and could eventually cause POP. Although the effect of birthweight in our meta-analysis seemed small (OR, 1.04), this was only the effect of an increase of 100 g. If we consider the effect of birthweight of 500 g instead of 100 g, the OR increases to 1.22, which indicates a clear effect with clinical significance. Levator defect has been a widely investigated subject, both as a risk factor and as an outcome measure. Our review is the first review confirming that levator defect is a risk factor for primary POP and POP recurrence by pooling the results into a meta-analysis. Not all studies concerning levator defect were included in this review because of insufficient data.
Concerning unmodifiable risk factors, in contradiction to the meta-analysis of Friedman et al, we pooled the results of age as a potential risk factor for prolapse recurrence. 6 In our forest plots, younger age was a clear risk factor for POP recurrence, and older age was a risk factor for developing primary POP. Women who are older simply have had more time to develop POP. As mentioned in the previous publication, hereditary tissue weakness could cause POP at a younger age and therefore cause Forest plot and meta-analyses for primary POP in association with the risk factors HRT, urinary incontinence, pulmonary disease, hysterectomy status, education, levator defect, and levator hiatal area For each study, the estimate, that is, odds ratio, hazard ratio, or prevalence ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes data of univariable analysis.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022. ajog.org recurrences at a younger age as well. Two recent meta-analyses reported family history as a risk factor for primary POP and POP recurrence. 6,58 In contrast to our review, these meta-analyses also included caseecontrol studies, data of univariable analyses, and studies about POP recurrence after mesh surgery. On the basis of our inclusion criteria, we could not include other studies that reported this potential risk factor.
Despite the fact that the pooled OR for smoking was statistically significant, the protective effect of smoking should be interpreted with care. The results of the studies that reported smoking as nonsignificant could not be pooled because of differences in definitions or lacking data.

Strengths and limitations
A strength of this review is the comprehensiveness of the review and metaanalyses with illustrating forest plots to summarize the best available evidence in this field. To the best of our knowledge, this is the first review to provide forest plots to give insight into the possible trends if the risk factors could not be pooled. Where most systematic reviews focus solely on one risk factor category, this systematic review included studies with all types of risk factors. We applied strict in-and exclusion criteria and only included studies with clear populations and outcome measures, multivariable analysis, and adequate follow-up to assure the best quality of the evidence.
We acknowledge some limitations of this review. First, we excluded studies that reported merely on composite outcome measures (ie, the combination of POP-Q, bothersome bulge, and/or retreatment) without reporting analyses for anatomic outcome measures separately. 20e22 Our review focused on anatomic outcomes, because we conformed to the previously set outcome measure. Subjective outcome measures may be confounded by spectrum bias. 52 In fact, approximately 8% of women report bothersome bulge symptoms without having an anatomic or objective prolapse. 59,60 It would be interesting to see whether the risk factors for subjective prolapse are applicable for anatomic prolapse also. Reoperation, on the other hand, is a reliable outcome measure that could have been included in the systematic review. Considering the extent of the systematic review, we decided to use the same outcome measures for primary POP and POP recurrence.
Another possible limitation of our review is that in contrast with other systematic reviews, we excluded the studies that reported on risk factors for recurrence after mesh surgeries. In our opinion, most cases in which mesh surgery is applied are complex and should not be compared with the group of women who undergo their first native tissue surgery. We also excluded studies from non-Western developed countries, because the population of women and exposition to risk factors in developing countries may not be comparable with those in Western developed countries. For example, a recent study reported that among others, anemia and carrying heavy objects for more than 5 hours a day are risk factors for developing POP 61 . Furthermore, we excluded caseecontrol studies to select the studies with the best evidence and smaller risk of selection bias. Therefore, studies on Forest plot and meta-analyses for POP recurrence in association with the risk factors pulmonary disease, constipation, previous POP surgery, preoperative POP-Q stage, concomitant surgery, and levator defect For each study, the estimate, that is, odds ratio or hazard ratio with the corresponding 95% confidence interval, and if appropriate, the pooled results of random-effects meta-analysis, are shown. The superscript letter a denotes data of univariable analysis.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022. Systematic Reviews ajog.org    2 y POP-Q point D Not described BMI, body mass index in kg/m 2 ; N/n, number of women included in the study who underwent physical examination/number of women with POP recurrence; LH, levator hiatus; POP, pelvic organ prolapse; POP-Q, POP-Quantification system; SUI, stress urinary incontinence; UI, urinary incontinence.
Schulten. Risk factors for primary prolapse and prolapse recurrence. Am J Obstet Gynecol 2022.

Systematic Reviews
ajog.org genetic risk factors, which are mostly caseecontrol studies, were excluded. Lastly, a limitation not only in this review but also in the whole field of risk factor research is heterogeneity in definitions, outcome measures, and correction for confounders. Because of this heterogeneity, studies cannot be easily compared or pooled, which makes it hard to draw solid conclusions. We tried to overcome this limitation by providing the forest plots.

Conclusions and implications
In this review, we summarize the evidence on the selection of publications with the strongest evidence on the risk factors for POP. Age, BMI, birthweight, and levator defect are statistically significant risk factors for primary POP, and delivery mode, parity, and levator hiatal area are confirmed risk factors for primary POP. Cesarean delivery and smoking are significant protective factors against primary POP. For POP recurrence, younger age and higher preoperative POP-Q stage 3 and 4 are statistically significant risk factors.
Future research should focus on the identification of risk factors for POP recurrence. Although several studies have been performed identifying the risk factors for recurrence after mesh surgery, profound knowledge on risk factors after native tissue surgery is lacking. Future studies should also focus on the comparability of risk factors for anatomic outcome measures vs subjective or composite outcome measures. Furthermore, heterogeneity should be avoided by using definitions and outcome measures as used in previous studies. Thereby, future meta-analyses can be performed more accurately, and conclusions could be drawn with more certainty.
This meta-analysis may give clinicians and patients better insight into the individual risk of developing POP and POP recurrence after primary native tissue surgery. This knowledge can be helpful in the identification of high-risk patients and the development of preventive strategies. High-risk patients may need adjustment of counseling or treatment options and management of expectations. -