Temporal trends in obstetric trauma and inpatient surgery for pelvic organ prolapse: an age-period-cohort analysis

Background

The rates of cesarean delivery have increased over time in industrialized countries, while the rates of instrumental vaginal delivery have declined. Instrumental vaginal delivery and obstetric trauma are risk factors for pelvic floor disorders.

Objective

We carried out a population-based study to quantify the association between temporal changes in obstetric trauma during childbirth and temporal changes in surgery for pelvic organ prolapse.

Study Design

We designed a retrospective analysis to examine age-specific trends in vaginal and cesarean delivery, obstetric trauma, and surgery for pelvic organ prolapse among all women (pregnant and nonpregnant) in Washington State, from 1987 through 2009. Cases of obstetric trauma (including severe perineal tears and high vaginal lacerations) and inpatient surgery for pelvic organ prolapse were identified among all hospitalizations. Temporal trends and age-period-cohort regression analyses were used to quantify the time period, age, and birth cohort effects among women born from 1920 through 1980.

Results

From 1987 through 2009, cesarean delivery rates among women aged 15-44 years increased from 12.7-18.1 per 1000 women, vaginal delivery rates remained stable, and instrumental vaginal delivery rates declined from 6.3-3.9 per 1000 women. Obstetric trauma decreased from 6.7 in 1987 to 2.5 per 1000 women aged 15-44 years in 2009. Surgery for pelvic organ prolapse decreased from 2.1 in 1987 to 1.4 per 1000 women aged 20-84 years in 2009. Obstetric trauma rates in 1987 through 1999 among women 15-44 years old were strongly correlated with the rates of surgery for pelvic organ prolapse among women 25-54 years of age 10 years later in 1997 through 2009 (correlation coefficient 0.87, P < .001). Similarly, rates of midpelvic forceps delivery in 1987 through 1999 were correlated with the rates of surgery for pelvic organ prolapse 10 years later (correlation coefficient 0.72, P < .01). Regression analyses showed a strong effect of age on surgery for prolapse, temporal decline in surgery, and an effect of birth cohort, as younger cohorts (women born in ≥1965 vs 1940) had lower rates of surgery for pelvic organ prolapse.

Conclusion

Temporal decline in instrumental vaginal delivery and obstetric trauma may have contributed to the reduction in surgery for pelvic organ prolapse.

Introduction

Pelvic floor disorders, including pelvic organ prolapse, urinary incontinence, and fecal incontinence, greatly impact the quality of life of a large number of women and represent a significant public health burden. It is estimated that 25% of adult women in the United States have ≥1 pelvic floor disorders, and that 1 in 4 women will undergo surgery for stress urinary incontinence or pelvic organ prolapse during their lifetime. Routine gynecologic examinations reveal evidence of pelvic organ prolapse in up to 50% of adult women.

While the mechanical causes of pelvic floor disorders remain poorly understood, age, obesity, and obstetric trauma increase the risk of these disorders. Studies have shown that parous women are 3 times more likely to have urinary and fecal incontinence and are twice as likely to experience pelvic organ prolapse compared with nulliparous women. Vaginal birth in particular has been implicated in the risk of pelvic organ prolapse and urinary incontinence later in life. One vaginal delivery is associated with a 2-fold increased risk of urinary incontinence and a 4-fold increased risk of pelvic organ prolapse, while 2 vaginal deliveries increase the risk 2.4-fold for urinary incontinence, and 8-fold for prolapse (as compared with women who have not had a vaginal delivery). Long-term follow-up studies show a 40% increased risk of fecal incontinence among women with at least 1 vaginal delivery (as compared with 1 cesarean delivery), while a significant perineal tear (second-degree tear or higher) doubles the risk. Conversely, cesarean delivery is associated with less need for incontinence or prolapse surgery and is protective against prolapse symptoms. There is substantial epidemiological evidence showing a lower risk of pelvic floor disorders following cesarean delivery without labor as compared with vaginal delivery.

The last 2 decades have witnessed an unprecedented increase in the rate of cesarean delivery in high-income countries. In the United States, the percentage of cesarean deliveries increased by 62.6% from 20.1% in 1996 to 32.7% in 2013. Cesarean delivery is the most common surgical procedure among US women, with close to 1.3 million cesarean deliveries performed annually. While rates of cesarean delivery have increased, the rates of instrumental vaginal delivery have declined in the United States (from 9.0% of live births in 1990 to 3.3% of live births in 2013). We hypothesized that the decrease in instrumental vaginal delivery, especially midpelvic forceps delivery, would have led to a decrease in pelvic floor injury requiring subsequent surgery for pelvic organ prolapse. We therefore carried out a population-based study to examine the temporal changes in instrumental vaginal delivery rates and obstetric trauma rates and their association with temporal trends in surgery for pelvic organ prolapse.

Materials and Methods

We carried out a population-based study to assess the association between obstetric events, including midpelvic forceps and obstetric trauma, and surgery for pelvic organ prolapse. We examined temporal trends in cesarean and vaginal delivery; instrumental vaginal delivery, including midpelvic forceps; and obstetric trauma among women who resided in Washington State during the period from 1987 through 2009. We also examined temporal trends in surgery for pelvic organ prolapse. All women (both pregnant and nonpregnant) in the appropriate age group were included in the analysis to assess the effect of childbirth and related events on population rates of pelvic organ prolapse.

Information on the mode of delivery was obtained from the Comprehensive Discharge Abstract Database, which included all hospitalizations in Washington State from 1987 through 2009. International Classification of Diseases, Ninth Revision , Clinical Modification ( ICD-9-CM ) diagnostic and procedure codes were used to identify childbirth ( Appendix Table 1 ); procedure codes 74. ^∧∧were used to identify cesarean delivery and all other deliveries were considered vaginal. ICD-9-CM codes were used for identifying women who had an instrumental vaginal delivery and the subset with a midpelvic forceps delivery ( Appendix Table 1 ). Women with a diagnosis of pelvic floor trauma during the delivery hospitalization, including third- and fourth-degree perineal laceration, anal sphincter tear, obstetric laceration of cervix, and high vaginal laceration were also identified using ICD-9-CM diagnostic codes 664.2, 664.3, 664.6, 665.3, and 665.4, respectively. In addition, we examined temporal changes in the rates of prolonged labor, identified on hospital discharge abstracts by ICD-9-CM diagnostic codes 662.20, 662.21, 662.22, and 662.23.

ICD-9-CM procedure and diagnostic codes were also used to identify inpatient surgery related to pelvic organ prolapse among all women in the Comprehensive Discharge Abstract Database ( Appendix Table 2 ). This included prolapse surgery among all women 20-84 years of age. Among women with multiple surgeries for the same indication, only the first surgery was used to calculate rates (identified though an internal linkage of hospital records). US census data for Washington State for the years 1990 through 2000 and yearly intercensal age-specific population estimates for women were used to calculate population rates of cesarean and vaginal delivery, instrumental vaginal delivery, pelvic floor injury during childbirth, and surgery for pelvic organ prolapse. For calculation of the overall rates of childbirth-related events, the number of women aged 15-44 years residing in Washington State was used as the denominator, while for calculation of surgery for pelvic organ prolapse, the number of women aged 20-84 years was used.

We used age-period-cohort analyses to analyze temporal changes in the rates of childbirth-related events and pelvic organ prolapse surgery among various birth cohorts of women. Such analyses are important for describing the effects of age, period, and birth cohort simultaneously, as age effects can be confounded if period and/or cohort effects occur. Thus in our analyses, women aged 20 years in 1990 belonged to the cohort of women born in 1970. This cohort of women may have experienced the events of interest as 25-year-old women during the period 1995, and as 30-year-old women during the year 2000.

Age-period-cohort effects on pelvic organ prolapse surgery were modeled for each year from 1990 through 2009. As age, period, and cohort are linearly dependent (cohort = period-age), we used a regression model that first estimated an overall linear trend in surgery rates that reflected the sum of period and cohort effects (a drift parameter). Deviation from linearity uniquely attributable to period and cohort effects was then modeled to estimate independent period and cohort effects. These estimates of curvature, or deviations from linearity, were interpreted as a measure of change in the linear trend for period and cohort.

Temporal trends were assessed using the Cochran-Armitage test for a linear trend in proportions. Pearson correlation coefficients were used to assess the correlation between the rates of obstetric events among women 15-44 years old in the years from 1987 through 1999 and the rates of prolapse surgery 10 years later (from 1997 through 2009) among women aged 25-54 years. In addition, temporal trends in the number of births to primiparous women were examined to assess the potential effect of changes in parity. Data on the number of births by birth order and maternal age were obtained for years 1990 through 2009 from the Washington State Department of Health. Information on the total number of first births per year from 1987 through 2009 was also available from public vital statistics files (through the Washington State Department of Health).

Sensitivity analyses were carried out to examine the potential impact of changes in insurance status among women with surgery for pelvic organ prolapse. The distribution and types of primary payers were evaluated to assess if changes in medical insurance contributed to temporal changes in the number of procedures performed.

Since all analyses were performed on publicly accessible deidentified data, an exemption from ethics approval was granted by the Department of Social and Health Services, State of Washington. Analyses were carried out using software (SAS, Version 9.3; SAS Institute Inc, Cary, NC). Age-period-cohort models were fitted using the apc.fit function in the Epi package of the R program (Version 2.14.2).