The CARE trial included 322 stress-continent women with stages II–IV pelvic organ prolapse who were planning abdominal sacrocolpopexy.12,13
This trial was performed through the Pelvic Floor Disorders Network (PFDN), a cooperative agreement network sponsored by the National Institute of Child Health and Human Development. Each clinical site received institutional review board approval, and all women provided written informed consent. The enrolling sites were tertiary care centers, and pelvic floor reconstructive specialists performed the sacrocolpopexy procedures. Some aspects of intraoperative and postoperative care were mandated by protocol, such as deep venous thrombosis prophylaxis, as standard for that site. Other aspects of care were typical (although not specified in the protocol), like early ambulation and early feeding to regular diet before discharge.
This prospectively planned analysis included baseline and 2-year postoperative outcome data and complication and safety data collected throughout the 2-year postoperative period. Demographics and medical history were collected by interview at baseline. We quantified the burden of chronic illness at baseline using the Cumulative Illness Rating Scale (CIRS).14
At baseline and 2-year postoperative visits, participants underwent the Pelvic Organ Prolapse Quantification (POP-Q) examination15
by examiners blinded to group assignment. POP-Q points Ba, Bp, C, and the most dependent POP-Q point were used to assess descent of the anterior compartment, posterior compartment, apex, and maximum point of descent, respectively. A standardized stress test was conducted at maximal capacity or 300 mL, whichever was less, in the sitting or standing position, using Valsalva and/or cough as the stress maneuver. A catheterized postvoid residual volume (PVR) was measured after spontaneous voiding.
Validated questionnaires were centrally administered by telephone by trained staff at the Quality of Life (QOL) Interviewing Center, including the Pelvic Floor Distress Inventory (PFDI), Pelvic Floor Impact Questionnaire (PFIQ), and the 36-item Medical Outcomes Study Short-Form Health Survey (SF-36).16–18
The PFDI assesses the presence and amount of bother caused by pelvic floor symptoms, and its scales include the Urinary Distress Inventory (UDI), the Pelvic Organ Prolapse Distress Inventory (POPDI), and the Colorectal-anal Distress Inventory (CRADI). Scale scores range from 0–300 (UDI and POPDI) and from 0–400 (CRADI), with higher scores indicating more symptoms, greater bother associated with symptom(s), or both. The PFIQ assesses the impact of symptoms on activities of daily living. The 3 PFIQ scales (Urinary Impact Questionnaire [UIQ], Pelvic Organ Prolapse Impact Questionnaire [POPIQ], and Colorectal-Anal Impact Questionnaire [CRAIQ]) have scores ranging from 0–400, with higher scores indicating greater functional impact. Generic health-related quality of life was assessed with the SF-36. The SF-36 has 2 summary indices, the Mental Component Summary (MCS) and the Physical Component Summary (PCS), which are derived from the weighted averages of the individual domain scores.19
Postoperative satisfaction was assessed by responses to 2 items included in the 2-year interviews: “In your opinion, has the treatment of your pelvic floor condition been: very successful, moderately successful, somewhat successful, or not at all successful?,” and “Compared to how you were doing before your recent pelvic floor operation, would you say that you are: much better, a little better, about the same, a little worse, or much worse?”
Serious adverse events (SAE) were recorded throughout the trial and defined as untoward life-threatening or fatal medical occurrences, prolonged hospitalization or readmission for the index surgery, any condition that resulted in persistent or clinically significant disability, or any other important medical condition that occurred in the intra- or postoperative period up to the 2-year visit. SAEs were reviewed by a committee of 3 investigators (who did not enroll or provide care for subjects in the CARE trial) and consensus was used to group the events by the affected organ system or as a febrile illness, wound complication, or blood transfusion.
Stress urinary incontinence was defined as an answer of “yes” to any item of the UDI stress subscale, including leakage related to “coughing, sneezing or laughing,” “physical exercise, such as walking, running, aerobics or tennis,” “lifting or bending over,” and “leakage of small amounts of urine leakage (that is, drops),” and “loss of urine during sexual activity.” Bothersome symptoms were defined as symptoms associated with “moderate” or “quite a bit” of bother. A composite “stress endpoint” was defined as a “yes” response to any of 3 questions on the UDI stress sub-scale (leakage with “coughing, sneezing or laughing,” “physical exercise,” or “lifting or bending over”); urine loss on the standardized cough stress test; or treatment for stress incontinence after the index surgery.
Urge symptoms were defined as an answer of “yes” to any UDI irritative sub-scale question, including urge incontinence, urgency, frequency, nocturia, and enuresis. In addition, a composite “urge endpoint” was defined as any bothersome urge symptom(s) or treatment for bothersome urge symptom(s) after the index surgery. As previously described, imputed values were used for the QOL outcomes in women who had treatment for stress or urge urinary symptoms after the index surgery.20
Continuous variables were summarized by mean and standard deviation (SD). Variables with nonnormal distributions were summarized by median with interquartile range (IQR). Baseline body mass index (BMI) was used to define weight groups: obese (≥ 30 kg/m2), overweight (25–29.9 kg/m2), healthy weight (18.5–24.9 kg/m2), and underweight (<18.5 kg/m2). Baseline characteristics, surgical outcomes, operative variables, and complication rates were compared between the obese and healthy-weight women, which we thought was the most clinically relevant analysis. This would also help optimize our ability to detect differences associated with BMI using our available sample size. Categorical variables were compared using the χ2 test (when unadjusted) or by fitting a logistic regression model and testing the coefficient for obesity cohort (when adjusted). Continuous/ordinal variables were compared using the Student t test or Wilcoxon rank-sum test (when unadjusted) or by fitting a general linear model either to the values of the dependent variable or to the ranks of the dependent variable and testing the coefficient of the obesity cohort (when adjusted). Comparisons of baseline symptom scores were adjusted for age, previous urinary incontinence or prolapse surgery, and baseline POP-Q stage. Analyses of postoperative variables were adjusted for age and randomization assignment (Burch or no Burch), previous urinary incontinence or prolapse surgery (only urinary incontinence outcomes), and baseline POP-Q stage (only prolapse outcomes).
Statistical significance was reported if P ≤ .05. With the sample size available for this analysis (74 obese and 125 healthy-weight women), there was 80% power to detect a difference between proportions of 0.5 and 0.7 when testing at a 5% level of significance. Also, when the dependent measure was continuous, there was 80% power to identify a change of 0.41 standard deviations (an effect size of 0.41) when comparing the 2 groups at a 5% level of significance.