Hypothesis / aims of study
There is little long-term data on the benefits of pelvic floor muscle training for women with prolapse. The aim of this study was to determine if women in the intervention group of a trial of pelvic floor muscle training (PFMT) for prolapse were less likely to receive hospital treatment related to pelvic floor dysfunction in the long-term, than women who were in the control group and received only a lifestyle advice leaflet. In the original study 447 participants were randomised, between 2007 and 2010, either to the intervention group (n=225) or the control group (n=222) with 66% completing the questionnaires at the 12 month follow-up. At this time-point women in the intervention group reported significantly fewer symptoms in the pelvic organ prolapse symptom score (POP-SS) at 12 months than those in the control group (mean reduction in POP-SS from baseline 3·77 [SD 5·62] vs 2·09 [5·39]; adjusted difference 1·52, 95% CI 0·46–2·59; p=0·0053)(1).
Study design, materials and methods
This study is a data linkage follow-up of participants from 11 out of 23 centres of the multicentre trial(1). Participants were followed up for between 8 and 10 years by linking to outpatient and inpatient hospital data relating to activity between the date of randomisation and the end of the linkage period in May 2018. All participants had a Community Health Index (CHI) number which was linked to routinely collected NHS datasets.
Primary Outcome: Any treatment relating to urinary incontinence (UI), prolapse and faecal incontinence (FI) received during follow-up period identified from participants’ outpatient and inpatient data.
Secondary Outcomes included treatment specifically for related surgery, conservative treatment (pessary or neuro stimulation), or time to first treatment.
Analysis: Mixed effect logistic regression was performed with a dependent variables for hospital treatment during follow-up, with intervention group as a binary independent variable, treatment centre as a random effect, and adjustment for baseline symptom severity (POP-SS), prolapse stage (POP-Q) and motivation for prolapse surgery. Cox regression stratified by preference for surgery was used to analyse time to first treatment. Women who died were still included in the analysis, with their data censored in the survival analysis. Missing POP-SS scores were imputed at the median.
Interpretation of results
There was a significant difference between the trial groups in terms of whether or not participants had received hospital treatment for UI, prolapse or FI during the follow-up period (inpatient or outpatient, irrespective of whether they had surgery): a lower proportion of the intervention group (43.6%) received treatment than the control group (52.8%). This difference was statistically significant in a mixed effect model (aOR 0.61, 95% CI 0.37 to 0.99 95% p=0.047) and on multiple imputation (aOR 0.60, 95% CI 0.36 to 0.98, p=0.042).
The median time to first treatment or censoring in the intervention group was 3008 days (IQR=589-3396) and in the control group 2242 days (IQR 628.5-3279). In a cox proportional hazards model stratified by motivation for surgery and adjusted for prolapse symptom severity, stage of prolapse, age and BMI there was a significant hazard ratio in favour of the intervention group of 0.65 (95% CI 0.46-0.94, p=0.020).
There was no significant difference between groups in the use of pessary or neuro-stimulation (OR 0.66,95% CI 0.33 to 1.33, p=0.244) or undergoing any type of related surgery during follow up (OR 0.78,95% CI 0.47 to 1.32, p=0.356).