Success and failure are dynamic, recurrent event states after surgical treatment for pelvic organ prolapse

Abstract

Background:

The ideal measure of success after surgery for pelvic organ prolapse (POP) has long been debated. Historically, strict definitions based on anatomic perfection have dominated the literature. However, the importance of patient-centered perception of outcomes are equally or more important when comparing the success of various prolapse surgeries. Understanding the limitations of existing outcome definitions will guide surgical outcome reporting and comparisons of POP surgeries.

Objectives:

To describe the relationships and overlap between participants who met anatomic, subjective, and retreatment definitions of success or failure after POP surgery; demonstrate rates of transition between success and failure over time; and compare scores from the Pelvic Organ Prolapse Distress Inventory (POPDI), Short-Form Six-Dimension health index (SF-6D), and quality-adjusted life years (QALYs) between these definitions.

Study Design:

Definitions of surgical success were evaluated at 3 or 6, 12, 24, 36, 48, and 60 months after surgery for ≥ stage II POP in a cohort of women (N=1,250) from four randomized clinical trials conducted by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Pelvic Floor Disorders Network. Surgical failure was defined by composite measure requiring: 1) anatomic failure (POPQ Ba, Bp or C > 0); 2) subjective failure (presence of bothersome vaginal bulge symptoms); or 3) pessary or surgical retreatment for POP. POPDI, SF-6D and QALYs were compared between participants who met a variety of definitions of success and failure including novel “intermittent” success/failure over time.

Results:

Among the 433/1,250 (34.6%) women who had surgical failure outcomes at ≥1 timepoint, 85.5% (370/433) met only one component of the composite outcome at assessment of initial failure (anatomic failure = 46.7%, 202/433; subjective failure = 36.7%, 159/433; retreatment = 2.1%, 9/433). Only 12.9% (56/433) met criteria for both for anatomic and subjective failure. Despite meeting the criteria for failure in primary study reporting, 24.2% (105/433) of these transitioned between success and failure during follow-up, and 83.8% (88/105) of these intermittent success/failures met criteria for success at their last follow-up. There were significant associations between success/failure group and the 1- and 2- year QALYs as well as a time-varying group effect on POPDI and SF-6D scores.

Conclusions:

True failure rates after prolapse surgery may be overestimated in the current literature. Only 13% of clinical trial subjects initially met both subjective and objective criteria for failure. Approximately one-quarter of failures were intermittent and transitioned between success and failure over time, with the majority of intermittent failures being in a state of ‘surgical success’ at their last follow-up. Current composite definitions of success or failure may result in overestimation of surgical failure rates, potentially explaining, in part, the discordance with low retreatment rates after POP surgery.

Condensation:

In pelvic prolapse surgical trials, one-quarter of participants who met failure criteria fluctuated between success and failure including 84% who met success criteria at last follow-up.

Introduction

There is tremendous variation in reported rates of success after pelvic organ prolapse (POP) surgery.^1–4 This variation results, in part, from the use of multiple definitions and inconsistent guidance from outcomes researchers and professional societies.³ Current recommendations for defining success after POP surgery include reporting subjective and anatomical outcomes and retreatment.^{2, 3} These recommendations emerged from an analysis of 18 different definitions in a single cohort of 322 participants after abdominal sacrocolpopexy using variations of anatomic, symptomatic, or retreatment outcomes.² A definition that used absence of postoperative vaginal bulge symptoms was significantly associated with a patient’s assessment of overall improvement while anatomic success alone was not.² While these findings provided useful guidance, the study was limited by small numbers, a single procedure, lack of consideration to the duration women perceived their surgery as successful, and the fact that success after POP surgery may be dynamic over time.

The first aim of this study was to describe the relationships and overlap between participants who met anatomic, subjective, and retreatment definitions of success/failure after POP surgery using a larger, more diverse cohort. The second aim was to examine these outcomes longitudinally to determine whether participants transition between success and failure over time. We hypothesized that success/failure after prolapse surgery is a dynamic state: while some women achieve and persist in a state of success throughout the duration of follow-up, other women transition between success and failure (“intermittent” success/failure) leading them to be falsely classified as failures in time to event analysis methods. A third aim was to compare symptoms and quality of life in patients with “intermittent” success/failure states versus those with persistent success or failure.

Material and Methods

This was a retrospective analysis of women enrolled in four prospective randomized surgical trials for symptomatic stage II-IV POP conducted across 17 centers engaged in the Eunice Kennedy Shriver National Institute of Childhood Health and Human Development (NICHD) Pelvic Floor Disorders Network. The design and results of the four trials have been published.^5–10 The Colpopexy and Urinary Reduction Efforts (CARE) trial evaluated the effectiveness of prophylactic Burch cystourethropexy in reducing postoperative de novo stress urinary incontinence in stress-continent women undergoing abdominal sacrocolpopexy between March 2002 and February 2005.⁵ Those participants who completed 2-year follow-up, were offered enrollment in the extended CARE (e-CARE) study with follow-up for up to 9 years after surgery.⁶ The Outcomes Following Vaginal Prolapse Repair and Midurethral Sling (OPUS) trial evaluated the effectiveness of a prophylactic retropubic midurethral sling versus sham in reducing de novo stress urinary incontinence one year after surgery in stress-continent women undergoing vaginal prolapse surgery between May 2007 and October 2009.⁷ The OPUS trial also included women who declined to undergo randomization but participated in a patient-preference cohort. The Operations and Pelvic Muscle Training in the Management of Apical Support Loss (OPTIMAL) trial evaluated 2-year outcomes in women undergoing native tissue vaginal apical suspension with midurethral sling for symptomatic prolapse and stress urinary incontinence between January 2008 and March 2011.⁸ Participants were randomized in a 2x2 factorial design to: (1) perioperative behavioral therapy with pelvic floor muscle training versus usual care and (2) uterosacral ligament suspension versus sacrospinous ligament suspension. At 2-years, participants were invited to enroll in the extended trial for up to 5 years of follow-up.⁹ The Study of Uterine Prolapse Procedures Randomized Trial (SUPeR) trial compared the efficacy and adverse events of women undergoing vaginal hysterectomy with uterosacral ligament suspension versus transvaginal mesh hysteropexy between April 2013 and February 2015.¹⁰ The primary endpoint was 3 years after the last randomization, and participants were followed for up to 5 years. All studies received institutional review board approval at each site and all participants signed a written informed consent.

The primary outcome for our analyses was surgical success defined as a single composite measure requiring anatomic success, subjective success, and absence of retreatment for POP.² Surgical failure was defined as the occurrence of at least one of: anatomic failure, subjective failure, or retreatment for pelvic organ prolapse. Anatomic failure was defined as prolapse beyond the hymen (POPQ point Ba, Bp, or C > 0). Subjective failure was defined as the presence of a bothersome bulge as indicated by an affirmative response to either of the Pelvic Floor Distress Inventory (PFDI or PFDI-20) questions “Do you usually have a sensation of bulging or protrusion from the vaginal area?” or “Do you usually have a bulge or something falling out that you can see or feel in the vaginal area?” with any degree of bother more than “not at all.” Retreatment included either surgery or pessary. Subjective and anatomic failures could be transient across the visits, but retreatment failure was considered a permanent state of failure. Notably, the composite outcome measure in this analysis slightly differs from those used in the original reports of the OPTIMAL, OPUS and CARE trials.

Success/failure outcome definitions were evaluated in this analysis at 3 or 6, 12, 24, 36, 48, and 60 months after surgery as available. Four groups were defined based on surgical success and failure states over time. The “Persistent Success” group met the definition of success at all time points for which the participant had data available, the “Persistent Failure” group met the definition of failure and their outcomes remained failures at all subsequent visits, and the “Intermittent Success/Failure” group fluctuated between success and failure over time. Participants in the Intermittent Success/Failure group that met the definition of success at their last follow-up were further classified as “Terminal Success” and those who met the definition of failure at their last follow-up were labeled “Terminal Failure.”

Secondary outcomes included POPQ examinations, pelvic organ prolapse distress inventory (POPDI) subscale scores from the PFDI or PFDI-20, and Short-Form Six-Dimension health index (SF-6D) scores. Quality-adjusted life years (QALYs) were calculated using an area under the curve approach following the trapezoidal rule from each subject’s SF-6D Index score reported at baseline, 3 or 6 months, 12 months, and 24 months.^{11,12, 13}

POPDI subscale scores, SF-6D scores and QALYs were compared between participants who met each of the definitions of success and failure over time. Adjusted means, standard errors, and p-values for repeated measures outcomes were obtained from general linear mixed models with fixed effects for surgical failure group, visit, and their interaction, and modeling within-subject correlation across visits with an unstructured correlation structure. These modeling methods assumed that missing data were missing at random. This was considered reasonable, particularly because most missing data were due to differing pre-specified lengths of follow-up for the various studies which was not expected to bias the results. Preoperative risk factors were compared between surgical failure groups using Kruskal-Wallis test for continuous measures and Fisher’s Exact Test for categorical measures.

Results

Participant flow for the four trials is depicted in Figure 1. In summary, 93% (1,250/1,337) of eligible and randomized participants were included in our analysis: 311 of 322 CARE/E-CARE participants, 414 of 466 OPUS participants, 352 of 374 OPTIMAL/E-OPTIMAL participants and 173 of 175 SUPeR participants.

Figure 1.

STROBE participant flow

Surgical outcomes and overlap of different definitions of success among all participants at first failure are demonstrated in Figure 2. Among the 433/1250 (34.6%) of participants who met failure criteria at least once during follow-up, at the initial time of failure 12.9% (56/433) of participants met criteria for both anatomic and subjective failure, and 85.5% (370/433) met only one component of the composite failure outcome (anatomic failure = 46.7%, 202/433; subjective failure = 36.7%, 159/433; retreatment = 2.1%, 9/433).

Figure 2.

Overlap of Surgical Outcome Definitions at Time of Initial Failure (N=433)

Longitudinal patterns of success/failure using the recommended composite definition are demonstrated in Figure 3. The majority of participants (817/1250, 65.4%) met the definition of success at all time points and were classified into the “persistent success” group. Participants that met criteria for failure and stayed in a state of failure, the “persistent failure” group, included 328/1250 (26.2%) participants. The median time to failure in this group was 1.0 year (interquartile range [IQR] 0.5 to 2.0 years). The 328 women in the persistent failure group met criteria for surgical failure at 617 visits including: subjective failure only: 199/617 (32.3%), anatomic failure only: 234/617 (37.9%), and anatomic and subjective failure: 123/617 (19.9%). POP retreatment with surgery or pessary occurred in 61/328 (18.6%) of women in the persistent failure group.

Figure 3.

Dynamic success and failure states after pelvic organ prolapse surgery over time for all participants. A state of success is indicated by a black circle, state of failure indicated by a red X and retreatment for pelvic organ prolapse indicated by a blue circle. The top group are the ‘persistent success’ participants who meet the definition of success at all time points. The bottom group are ‘persistent failures’ designated by red X at all follow-up visits. The middle group are participants that move back and forth between success and failure states over the follow-up period.

The longitudinal patterns of the ‘intermittent’ success/failures are demonstrated in Figure 4. Despite being considered failures in the primary study analyses using time-to-event methods, 24.2% (105/433) of participants who met the criteria for failure transitioned between success and failure during follow-up and 83.8% (88/105) of these were successes at their last follow-up (i.e. “terminal successes”) while 17/105 (16.2%) were “terminal failures”. The median time to initial failure for “terminal failures” was 1.0 year (IQR 0.5 to 1.0 years) and the median time to initial failure for “terminal successes” was 1.0 year (IQR 0.5 to 2.0 years). The intermittent success/failure group had a total of 175 visits with surgical failure including: subjective failure only: 97/175 (55.4%), anatomic failure only: 69/175 (39.4%), and anatomic and subjective failure: 8/175 (4.6%). Only 1 of the 105 women in the terminal failure group underwent POP retreatment.

Figure 4.

Dynamic success and failure states after pelvic organ prolapse surgery over time in participants in the intermittent failure/success group. A state of success is indicated by a black circle, state of failure indicated by a red X and retreatment for pelvic organ prolapse indicated by a blue circle. The top group are the ‘terminal success’ participants who meet the definition of success at the last follow-up visit. The bottom group are ‘terminal failures’ who meet the definition of failure at their last follow-up visit.

Characteristics of participants who met definitions of persistent success, persistent failure and intermittent failure/success are demonstrated in Table 1. Compared to the persistent success group, women in the intermittent success/failure group were more likely to be obese and to have had any Cesarean birth. They were less likely to be non-Hispanic White, and they had worse baseline POPDI total score and SF-6D Index scores. The intermittent success/failure group was also less likely to have undergone abdominal sacral colpopexy.

Table 1.

Baseline Characteristics of Study Participants by Surgical Outcome Group

					Odds Ratio/Location Shift & 95% CI (Reference=Persistent Success) a
Characteristic	Persistent Failure (N=328)	Intermittent Failure/Success (N=105)	Persistent Success (N=817)	P-value a	Persistent Failure	Intermittent Failure/Success
Baseline Characteristics
Age (years)	61 (54, 69)	60 (54, 70)	62 (55, 68)	0.723	−0.6 (−1.9 – 0.8)	−0.3 (−2.5 – 2.0)
Body Mass Index (kg/m2)	28 (25, 31)	29 (26, 33)	27 (24, 30)	<0.001	0.8 (0.2 – 1.4)	1.7 (0.8 – 2.7)
Obese (BMI ≥ 30 kg/m2)	101/327 (30.9)	44/105 (41.9)	220/817 (26.9)	0.006	1.2 (0.9 – 1.6)	2.0 (1.3 – 3.0)
Non-Hispanic Caucasian	234/327 (71.6)	70/105 (66.7)	689/813 (84.7)	<0.001	0.5 (0.3 – 0.6)	0.4 (0.2 – 0.6)
Some College or Greater	179/310 (57.7)	53/98 (54.1)	485/799 (60.7)	0.358	0.9 (0.7 – 1.2)	0.8 (0.5 – 1.2)
Married	199/311 (64.0)	69/99 (69.7)	584/798 (73.2)	0.011	0.7 (0.5 – 0.9)	0.8 (0.5 – 1.4)
Nulliparous	3/327 (0.9)	3/105 (2.9)	13/816 (1.6)	0.317	0.6 (0.1 – 2.1)	1.8 (0.3 – 6.8)
Any Cesarean Delivery	21/326 (6.4)	15/105 (14.3)	50/814 (6.1)	0.015	1.1 (0.6 – 1.8)	2.5 (1.3 – 4.8)
Any Vaginal Delivery	320/327 (97.9)	100/105 (95.2)	801/816 (98.2)	0.141	0.9 (0.3 – 2.5)	0.4 (0.1 – 1.3)
Post-Menopausal	278/328 (84.8)	87/105 (82.9)	697/817 (85.3)	0.768	1.0 (0.7 – 1.4)	0.8 (0.5 – 1.5)
Currently using Vaginal Estrogen	82/327 (25.1)	29/105 (27.6)	237/817 (29.0)	0.415	0.8 (0.6 – 1.1)	0.9 (0.6 – 1.5)
Prior Urinary Incontinence Surgery	13/327 (4.0)	8/105 (7.6)	33/817 (4.0)	0.235	1.0 (0.5 – 1.9)	2.0 (0.8 – 4.5)
Prior Prolapse Surgery	46/327 (14.1)	19/105 (18.1)	143/817 (17.5)	0.343	0.8 (0.5 – 1.1)	1.0 (0.6 – 1.8)
Prior Hysterectomy	109/327 (33.3)	38/105 (36.2)	319/817 (39.0)	0.191	0.8 (0.6 – 1.0)	0.9 (0.6 – 1.4)
Current Smoker	24/327 (7.3)	5/105 (4.8)	49/817 (6.0)	0.592	1.2 (0.7 – 2.1)	0.8 (0.2 – 2.0)
Diabetes	37/325 (11.4)	13/102 (12.7)	93/810 (11.5)	0.919	1.0 (0.6 – 1.5)	1.1 (0.6 – 2.1)
Connective Tissue Disease	6/321 (1.9)	1/103 (1.0)	6/812 (0.7)	0.234	2.6 (0.7 – 9.6)	1.3 (0.0 – 11.0)
POPQ Measurements
POPQ Ba	3.0 (1.0, 4.0)	2.0 (1.0, 4.0)	2.0 (1.0, 4.0)	0.007	0.5 (0.0 – 1.0)	0.0 (0.0 – 1.0)
POPQ Bp	−1.0 (−2.0, 1.5)	−1.0 (−2.0, 1.0)	−1.0 (−2.0, 1.0)	0.653	0.0 (0.0 – 0.0)	0.0 (0.0 – 0.5)
POPQ C	0.0 (−3.0, 4.0)	−2.0 (−4.0, 2.5)	−1.0 (−4.0, 3.0)	0.015	1.0 (0.0 – 1.0)	0.0 (−1.0 – 0.0)
POPQ Stage 3 or 4 b	256/328 (78.0)	76/105 (72.4)	578/817 (70.7)	0.041	1.5 (1.1 – 2.0)	1.1 (0.7 – 1.8)
Anatomic POP beyond hymen or POP symptoms c	325/327 (99.4)	103/104 (99.0)	794/814 (97.5)	0.085	4.1 (1.0 – 36.3)	2.6 (0.4 – 108.6)
Anatomic POP beyond hymen c	294/328 (89.6)	97/105 (92.4)	715/817 (87.5)	0.282	1.2 (0.8 – 1.9)	1.7 (0.8 – 4.2)
POP symptoms c	295/313 (94.2)	94/99 (94.9)	733/797 (92.0)	0.331	1.4 (0.8 – 2.6)	1.6 (0.6 – 5.4)
Surgical Repairs
Antero-apical POP Repair (AAR)				<0.001
AAR Native Tissue	250/328 (76.2)	79/105 (75.2)	475/817 (58.1)		Reference	Reference
AAR (Mesh/Graft)	24/328 (7.3)	8/105 (7.6)	103/817 (12.6)		0.4 (0.3 – 0.7)	0.5 (0.2 – 1.0)
Abdominal Sacral Colpopexy (Mesh/Graft)	54/328 (16.5)	18/105 (17.1)	239/817 (29.3)		0.4 (0.3 – 0.6)	0.5 (0.2 – 0.8)
Posterior POP Repair (PR)				0.257
Not Performed	178/327 (54.4)	55/105 (52.4)	395/816 (48.4)		1.3 (1.0 – 1.7)	1.1 (0.7 – 1.7)
PR (Native Tissue)	143/327 (43.7)	50/105 (47.6)	408/816 (50.0)		Reference	Reference
PR (Mesh/Graft)	6/327 (1.8)	0/105 (0.0)	13/816 (1.6)		1.3 (0.4 – 3.8)	0.0 (0.0 – 2.2)
Concomitant Urinary Incontinence Surgery (MUS)	222/267 (83.1)	82/101 (81.2)	483/672 (71.9)	<0.001	1.9 (1.3 – 2.8)	1.7 (1.0 – 3.0)
Concomitant Hysterectomy	203/328 (61.9)	61/105 (58.1)	425/817 (52.0)	0.008	1.5 (1.1 – 2.0)	1.3 (0.8 – 2.0)
Patient Reported Outcomes (Baseline)
SF-6D Index Score	0.7 (0.6, 0.8)	0.6 (0.6, 0.8)	0.7 (0.6, 0.8)	0.001	−0.0 (−0.0 – 0.0)	−0.1 (−0.1 – −0.0)
SF-6D Physical Function Score	2.0 (2.0, 3.0)	3.0 (2.0, 4.0)	2.0 (2.0, 3.0)	0.004	0.0 (0.0 – 0.0)	0.0 (0.0 – 1.0)
SF-6D Role Limitation Score	2.0 (1.0, 4.0)	3.0 (2.0, 4.0)	2.0 (1.0, 4.0)	0.031	0.0 (0.0 – 0.0)	0.0 (0.0 – 0.0)
SF-6D Social Function Score	2.0 (1.0, 3.0)	2.0 (1.0, 3.0)	1.0 (1.0, 2.0)	<0.001	0.0 (0.0 – 0.0)	0.0 (0.0 – 0.0)
SF-6D Pain Score	3.0 (2.0, 4.0)	3.0 (2.0, 4.0)	3.0 (2.0, 4.0)	0.025	0.0 (0.0 – 0.0)	0.0 (0.0 – 1.0)
SF-6D Mental Health Score	2.0 (2.0, 3.0)	2.0 (2.0, 3.0)	2.0 (1.0, 3.0)	0.004	0.0 (0.0 – 0.0)	0.0 (0.0 – 1.0)
SF-6D Vitality Score	3.0 (2.0, 3.0)	3.0 (2.0, 4.0)	3.0 (2.0, 3.0)	0.106	0.0 (0.0 – 0.0)	0.0 (0.0 – 0.0)
POPDI Total Score	92.9 (56.5, 153.0)	105.4 (54.2, 152.4)	80.4 (45.2, 131.5)	0.002	11.9 (3.6 – 19.6)	17.3 (4.8 – 30.4)

OR=Odds Ratio, CI=Confidence Interval, P25=25^th Percentile, P75=75^th Percentile, BMI=Body Mass Index, POPQ = Pelvic Organ Prolapse Quantification, POP=Pelvic Organ Prolapse. Data are median (P25, P75) for continuous measures and n/N (%) for categorical measures unless otherwise specified.

^aFor categorical measures, odds ratios, 95% confidence intervals, and p-values were obtained from Fisher’s Exact Test. For continuous measures p-values were obtained using Kruskal-Wallis test and location shift and 95% confidence intervals were obtained using Wilcoxon Rank-Sum test with a Hodges-Lehmann estimation of location shift. All tests were conducted at a significance level of 0.05 and no adjustments for multiple comparisons were made.

^bPelvic Organ Prolapse Quantification (POPQ) Stages: Stage 2-The vagina is prolapsed between 1 cm above the hymen and 1 cm below the hymen; Stage 3-The vagina is prolapsed more than 1 cm beyond the hymen but is less than totally everted; Stage 4-The vagina is everted to within 2 cm of its length.

^cAnatomic pelvic organ prolapse (POP) beyond hymen is defined as a POPQ point Ba, Bp, or C > 0. Pelvic organ prolapse (POP) symptoms is defined as the presence of a bothersome bulge (i.e. a positive response to PFDI-46 items 4 and 5/PFDI-20 item 3 with a degree of bother indicated

Adjusted associations between success/failure group and patient-reported outcomes by year of follow-up are demonstrated in Table 2. There were significant associations between success/failure group and the 1- and 2- year QALYs as well as a time-varying group effect on POPDI, SF-6D Index and each of the sub-scales except social function and mental health. Women with intermittent success/failure had mean POPDI scores in-between those meeting persistent failure and persistent success definitions at each year of follow-up.

Table 2.

Adjusted Analyses of Quality of Life Outcomes by Surgical Outcome Group

Outcome Measures	Persistent Failure (N=328)	Intermittent Success/Failure (N=105)	Persistent Success (N=817)	P-value a
POPDI Score
3/6 Months b	38.72 (310, 2.30)	45.35 (97, 4.08)	22.76 (759, 1.46)	<0.001
1 Year	45.47 (287, 2.28)	42.63 (99, 3.93)	19.55 (733, 1.43)	<0.001
2 Years	60.39 (159, 3.28)	51.24 (89, 4.89)	21.64 (431, 2.02)	<0.001
3 Years	64.48 (104, 3.73)	50.39 (75, 5.07)	22.06 (325, 2.20)	<0.001
4 Years	77.70 (56, 5.02)	59.87 (47, 6.18)	21.51 (173, 2.93)	<0.001
5 Years	82.08 (59, 5.50)	51.94 (52, 6.33)	23.85 (170, 3.26)	<0.001
SF-6D Physical Function Score
3/6 Months b	2.31 (303, 0.07)	2.38 (98, 0.12)	2.27 (743, 0.04)	0.664
1 Year	2.28 (288, 0.07)	2.45 (99, 0.11)	2.00 (717, 0.04)	<0.001
2 Years	2.29 (160, 0.08)	2.46 (85, 0.12)	1.97 (415, 0.05)	<0.001
3 Years	2.43 (102, 0.09)	2.44 (73, 0.12)	2.02 (317, 0.05)	<0.001
4 Years	2.51 (57, 0.13)	2.28 (48, 0.15)	2.15 (170, 0.08)	0.059
5 Years	2.59 (59, 0.13)	2.54 (52, 0.14)	2.14 (168, 0.08)	0.002
SF-6D Role Limitation Score
3/6 Months b	2.25 (305, 0.07)	2.49 (98, 0.12)	2.23 (756, 0.04)	0.135
1 Year	2.21 (289, 0.07)	2.40 (99, 0.12)	1.83 (729, 0.04)	<0.001
2 Years	2.20 (162, 0.09)	2.31 (89, 0.13)	1.98 (423, 0.06)	0.018
3 Years	2.50 (103, 0.11)	2.53 (74, 0.14)	2.09 (324, 0.06)	<0.001
4 Years	2.73 (57, 0.15)	2.66 (48, 0.17)	2.06 (175, 0.08)	<0.001
5 Years	2.70 (58, 0.14)	2.63 (53, 0.16)	2.28 (169, 0.09)	0.017
SF-6D Social Function Score
3/6 Months b	1.70 (307, 0.06)	1.80 (99, 0.10)	1.59 (758, 0.04)	0.052
1 Year	1.73 (290, 0.05)	1.79 (99, 0.09)	1.49 (733, 0.03)	<0.001
2 Years	1.86 (162, 0.07)	1.81 (89, 0.11)	1.55 (429, 0.05)	<0.001
3 Years	1.86 (104, 0.09)	1.81 (75, 0.11)	1.58 (324, 0.05)	0.011
4 Years	1.80 (57, 0.11)	1.96 (48, 0.12)	1.52 (175, 0.06)	0.002
5 Years	2.01 (59, 0.12)	1.91 (53, 0.13)	1.62 (169, 0.07)	0.010
SF-6D Pain Score
3/6 Months b	2.38 (307, 0.07)	2.49 (98, 0.13)	2.30 (756, 0.05)	0.341
1 Year	2.37 (289, 0.08)	2.56 (99, 0.13)	2.18 (732, 0.05)	0.007
2 Years	2.61 (162, 0.09)	2.56 (89, 0.13)	2.20 (427, 0.06)	<0.001
3 Years	2.79 (104, 0.11)	2.75 (75, 0.15)	2.30 (323, 0.07)	<0.001
4 Years	2.94 (57, 0.14)	3.02 (48, 0.17)	2.38 (175, 0.08)	<0.001
5 Years	2.90 (59, 0.14)	3.06 (53, 0.16)	2.31 (170, 0.08)	<0.001
SF-6D Mental Health Score
3/6 Months b	2.08 (307, 0.06)	2.23 (98, 0.10)	1.94 (757, 0.04)	0.007
1 Year	2.05 (290, 0.06)	2.21 (99, 0.10)	1.92 (733, 0.04)	0.009
2 Years	2.16 (162, 0.08)	2.22 (89, 0.11)	1.94 (426, 0.05)	0.007
3 Years	2.06 (104, 0.08)	2.12 (75, 0.11)	1.89 (324, 0.05)	0.060
4 Years	1.90 (57, 0.11)	2.28 (48, 0.12)	1.91 (175, 0.06)	0.020
5 Years	2.12 (59, 0.11)	2.22 (53, 0.13)	1.89 (170, 0.07)	0.030
SF-6D Vitality Score
3/6 Months b	2.70 (307, 0.05)	2.75 (99, 0.10)	2.54 (756, 0.03)	0.014
1 Year	2.60 (289, 0.06)	2.56 (99, 0.10)	2.51 (732, 0.04)	0.407
2 Years	2.73 (162, 0.07)	2.68 (89, 0.09)	2.54 (429, 0.04)	0.034
3 Years	2.80 (104, 0.08)	2.71 (75, 0.10)	2.51 (324, 0.05)	0.004
4 Years	2.76 (57, 0.10)	2.80 (48, 0.11)	2.47 (174, 0.06)	0.004
5 Years	3.15 (59, 0.10)	2.77 (53, 0.12)	2.57 (170, 0.06)	<0.001
SF-6D Index Score
3/6 Months b	0.75 (301, 0.01)	0.73 (96, 0.01)	0.76 (737, 0.00)	0.021
1 Year	0.75 (285, 0.01)	0.74 (99, 0.01)	0.79 (713, 0.00)	<0.001
2 Years	0.73 (160, 0.01)	0.74 (85, 0.01)	0.78 (405, 0.01)	<0.001
3 Years	0.73 (101, 0.01)	0.73 (72, 0.01)	0.78 (316, 0.01)	<0.001
4 Years	0.73 (57, 0.01)	0.71 (48, 0.02)	0.77 (169, 0.01)	<0.001
5 Years	0.70 (58, 0.01)	0.71 (52, 0.01)	0.76 (166, 0.01)	<0.001
1-Year QALY c	0.75 (272, 0.01)	0.72 (94, 0.01)	0.76 (678, 0.00)	<0.001
2-Year QALY c	0.75 (289, 0.01)	0.72 (100, 0.01)	0.78 (708, 0.00)	<0.001

SE=Standard Error. Data are adjusted mean (n, SE).

^aAdjusted means, standard errors, and p-values for repeated measures outcomes were obtained from general linear models adjusting for surgical outcome group, visit ^b, interaction between surgical outcome group and visit while controlling for within subject correlation across visits with an unstructured working correlation structure. Adjusted means, standard errors, and p-values for all other outcomes were obtained from general linear models adjusting for surgical outcome group. All tests were conducted at a significance level of 0.05.

^b3-month visit for subjects in OPUS and CARE studies, 6-month visit for subjects in OPTIMAL and SUPER studies

^cQuality adjusted life years (QALYs) are calculated using an area under the curve approach following the trapezoidal rule from each subject’s SF-6D Index score reported at baseline, 3/6 months b, 12 months, and 24 months. Each subject must have at least an SF-6D Index score reported at baseline and 12 months in order for a 1-year QALY to be calculated. Each subject must have at least an SF-6D Index score reported at baseline and either at 12 or 24 months in order for a 2-year QALY to be calculated. If the SF-6D Index score is missing at 24 months, it was assumed to be the same as at 12 months.

Structured Discussion

Principal Findings

Outcomes after surgery for prolapse have largely been evaluated, analyzed, and interpreted as dichotomous: success or failure. This analysis of a large, diverse cohort from data collected within four rigorously conducted surgical trials suggests that outcomes after surgery for POP are considerably more nuanced and complex. Although most participants in these trials had successful surgery or had initial success that failed over time, a significant proportion demonstrated dynamic outcomes: Their vaginal support and/or POP symptoms transitioned back and forth between success and failure outcomes at various time points without treatment.

Results

Over 80% of those who fluctuated (intermittent success/failure) met all criteria for success at their last outcome assessment (terminal success). Those with ‘intermittent success/failure’ also report intermediate subjective outcomes between those who met persistent success or failure definitions. These findings have important implications for how we understand, measure, and analyze outcomes after POP surgery.

Another area of complexity in POP surgery outcomes highlighted in this analysis is the discordance between anatomic and subjective outcomes. The vast majority of women included in this analysis who met criteria for failure met only one of three components of the composite failure definition at the initial assessment of failure (85.5%) with only 12.9% reporting both symptomatic and anatomic failure and less than 3% having undergone retreatment. This discordance is even greater than that seen in the original analysis of the CARE study population where 64% of patients with prolapse beyond the hymen at 2 years postoperatively also had vaginal bulge symptoms.² Understanding both rates of anatomic and subjective recurrence after POP surgery are important in understanding the effectiveness and mechanisms of failure of these procedures.

The substantial discordance between symptoms and vaginal anatomy seen in our analysis and others suggests that it may be more valuable to evaluate these outcomes independently, rather than as a composite outcome for POP surgery, since a single composite outcome may muddle our understanding of a patient’s experience. Moreover, our data suggest that single composite outcomes likely overestimate failure relative to the entirety of the patient’s postoperative experience. This discordance also highlights the need for future research about what outcomes matter most to patients – the available data suggests it is subjective outcomes.¹⁴

In prospective studies of POP repair, surgical outcomes can be assessed either at discrete timepoints (irrespective of outcomes at earlier timepoints) or they can be assessed over time using the “once a failure, always a failure” approach common to survival analyses. Our data suggest that the latter approach may not truly reflect the patients’ experience for the assessment of the long-term outcomes of POP procedures. Reports on the natural history of prolapse in women suggest that the majority of prolapse does not progress with time and even 3% may demonstrate ≥2 cm of regression in the absence of treatment.^15–17 While anatomic regression represents a minority, others have reported improved symptoms of POP in 64% with only 6% worsening over 5 years.¹⁸ Our data are consistent with these reports and support that prolapse may be a “recurrent event” even after surgical intervention. Our data make a strong argument that traditional time to (first) event analysis methods are not ideal for evaluating POP surgeries; analytic techniques that account for different outcome states in a single individual over time may provide more valid comparisons.

Clinical Implications

Our data also support that the intermittent success/failure states matter to participants. Women whose outcomes are in these ‘intermittent’ failure states report different quality of life and satisfaction rates compared to those who are persistently in a state of success or failure over time. Women with anatomic failures, but no symptoms, are commonly seen in the clinical setting and considered “satisfied failures” because they do not request additional retreatment. Possible explanations for the phenomenon may be the perception that, although not perfect, the degree of prolapse is significantly improved compared to preoperatively. They may also have had other symptoms resolved with the repair (such as incontinence or retention) that improved their quality of life despite not having a perfect anatomic result. Alternately, fear of disappointing their surgeon or denial regarding the success of the operation may prompt a participant to deny symptoms in setting of prolapse beyond the hymen. Women who are labeled ‘failures’ in the absence of anatomic prolapse beyond the hymen may have reported symptoms of pressure and bulge at one time point due to other conditions, such as constipation or urinary tract infection. Finally, while questions related to prolapse symptoms are highly specific with reports ranging from 79 to 99% ruling “in” the diagnosis, the sensitivity can be low particularly in populations with low prevalence making the absence of prolapse based on questionnaire alone likely overestimated.^19–21 The characteristics of the questionnaires themselves may contribute to the “intermittent” success/failure based on symptoms alone. Similarly, the measurement error inherent in the POPQ, including a patient’s Valsalva effort, or body mass index may account for intermittent anatomic success and failure over time.

Research Implications

Implications of this study are that the currently accepted methods likely underestimate success rates after prolapse surgery when using single composite definitions and time to event approaches. This may result in overestimation of surgical failure rates and potentially, in part, explain low retreatment rates. Future research considerations include the need to characterize failure types over time and to update recommendations on defining success/failure that report anatomic, subjective and retreatment outcomes separately; and to follow-up over a minimum amount of time with multiple measures to account for the dynamic nature of POP. Researchers may consider reanalyzing pivotal trials using these new recommendations to guide clinicians in estimating average treatment effects over time for common procedures in the field. Qualitative research into the discordance between subjective and objective outcomes may help inform which outcomes to use when comparing optimal surgical treatment approaches. Finally, additional research should be performed to identify women who may experience intermittent failed outcomes since it could be useful to counsel these women preoperatively and potentially alter treatment recommendations.

Strengths and Limitations

Strengths of this study include the use of randomized controlled trials, consistently implemented surgical techniques, standardized collection of objective and subjective outcome measures across multiple sites, and masked examiners and interviewers. Limitations to this work include the fact that multiple trials were analyzed in aggregate potentially resulting in biases due to varying study designs and differing missing values. Additionally, two of the four trials were not designed to compare outcomes of POP surgery itself but rather urinary symptoms after POP surgery and there were varying follow-up time points for each trial with one only having 1-year follow-up.⁷ Lastly, the long time-span during which these trials were conducted does not account for evolving practice patterns.

Conclusions

True failure rates after prolapse surgery may be overestimated in the current literature. In this population, approximately, one in four failures were intermittent and fluctuated between success and failure over time. Only 13% initially met failure definitions by both subjective and objective criteria. The majority of intermittent failures were in a state of ‘surgical success’ at their last follow-up. Both anatomic and subjective outcomes after POP surgery appear to be dynamic in nature and traditional survival analyses where “once a failure, always a failure” may not be ideal.

AJOG at a Glance:

1. Why was the study conducted? Understanding whether anatomic and subjective outcomes after pelvic organ prolapse surgery are concordant and whether ‘failure’ as a dynamic state should guide future surgical outcome definitions and analyses.

2. What are the key findings? Most initial failures after prolapse surgery meet only one component of a composite definition and one-quarter of ‘failures’ transition during follow-up, with most being in a state of success at their last visit. Women with ‘intermittent’ outcomes report symptoms and quality of life scores in-between women with persistent success and persistent failure outcome states.

3. What does this study add to what is already known? Composite success definitions and time-to-event analysis methods after prolapse repair may overestimate failure rates and impede true understanding of a woman’s long-term perception of success.