Impact of surgery for stress incontinence on morbidity: cohort studyNick Black, professor of health services research,a Joanne Griffiths, research fellow,a Catherine Pope, research fellow,a Ann Bowling, senior lecturer in health services research,a Paul Abel, reader in urology b
a Health Services Research Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT, b Department of Surgery, Royal Postgraduate Medical School, London W12 0NN
Correspondence to: Professor Black firstname.lastname@example.org
Objectives: To describe the impact of surgery for stress
incontinence on the severity of symptoms, other mental and
physical symptoms, and overall health. To describe the
incidence of postoperative complications.
Surgery is recommended for women with urinary stress incontinence that has not responded to non-surgical measures and is causing social distress.1 The surgical procedures used in the United Kingdom are colposuspension, needle suspension, and anterior colporrhaphy. About 85% of women are continent one year after colposuspension and 50% to 70% after needle suspension or anterior colporrhaphy.2 Most textbooks reflect the literature and suggest that surgery is highly effective in curing stress incontinence in women.3 4 5 6
Appraisal of the 31 randomised and non-randomised prospective comparative studies reported in the literature, however, shows their methodological quality to be poor: cases are inadequately defined, outcome measures rarely consider either the patients' views or independent assessments and instead are based on surgeons' opinions, and the measures are of unknown reliability and validity. The external validity (generalisability) of the results is uncertain since most studies have been conducted in specialist centres and it is impossible to know whether the findings can be generalised to other settings.
To overcome these deficiencies we followed a large cohort of women undergoing surgery for stress incontinence in the United Kingdom from the time of or the day before surgery until 12 months after. Women were recruited from the lists of 49 surgeons working in 18 hospitals to provide a representative sample. Women were eligible for inclusion in the study if they were having surgery for stress incontinence, regardless of whether it had been confirmed by urodynamic studies. Because many believe that surgery is effective only in women with genuine stress incontinence—that is incontinence confirmed by urodynamic pressure studies—data are presented both for the whole cohort and for those with genuine stress incontinence. The evaluation of outcome was based on patients' views using validated measures.
We describe the impact of surgery on the severity of stress incontinence, other mental and physical symptoms (including urge incontinence), coexisting conditions, and global measures of health. We also describe the incidence of postoperative complications.
In 1992 all of the 137 gynaecologists and urologists who performed surgery for stress incontinence in the 33 NHS hospitals in the former North West and North East Thames regions (now known as the North Thames region) were invited to participate in the study. Sixty four agreed, 13 declined, and 60 did not reply. Of those who agreed to participate, 38 gynaecologists and 11 urologists working in 18 hospitals were selected as representative in terms of caseload, specialty, whether the surgery was done in a teaching or district general hospital, and whether the hospital served a rural, suburban, or urban population.
All women having surgery for stress incontinence between January 1993 and June 1994 were eligible for inclusion. Women were excluded if they were unable to read and understand English. Women were recruited by nursing staff during their stay in hospital. After giving consent, women completed a perioperative questionnaire, whose development has been described.7 Women provided information on sociodemographic factors, the history of urinary incontinence, extent and severity of stress incontinence (based on a newly developed symptom severity index),8 other urinary symptoms, coexisting conditions (based on an index similar to that used in other studies9 10 and described elsewhere8), mental health (based on an existing validated index),11 and their expectations of surgery. Surgeons provided information on the women's preoperative state, including results of tests and surgical management. Surgeons who did not respond received three written reminders.
All eligible women were identified from registers in each hospital. Information on non-participants was obtained from the hospital episode system. Participating and non-participating women were compared in terms of age, surgical procedure, hospital where they were treated, and length of stay.
Patients completed three postal questionnaires (at 3, 6, and 12 months after surgery). These covered the same topics as the perioperative questionnaire. Additionally, some questions were included to obtain women's perceptions of the impact of surgery, details of any postoperative complications, and length of recovery time. Non-respondents were compared with respondents to determine any response bias.
To minimise work for the surgeons and cost to the NHS, surgeons were asked to see only one third, rather than all, of the patients as outpatients at the three non-routine follow up examinations of this study at 3, 6, and 12 months after surgery. Surgeons completed a short questionnaire on the severity of any persistent stress incontinence, postoperative complications, and the outcome of surgery.
The study was descriptive. The objective was to provide preliminary estimates of variables and to generate hypotheses for testing in more specific studies using randomised trials when practical. Calculations of sample size based on specific hypotheses were therefore inappropriate.
Several statistical tests were used to detect changes in health. Student's t tests were used to find differences in mean values of continuous variables and 2 tests for differences in proportions of categorical variables. When measurements taken before and after surgery on the same patient were analysed, paired t tests and McNemar's tests for proportions were used. Repeated measures analysis of variance was used to assess change between 3 and 12 months after surgery using the generalised linear modelling procedure (SAS version 6.06).12 Non-normally distributed data were tested using the Wilcoxon matched pairs signed rank test.
Recruitment and response
A total of 367 of the 442 participants (83%) returned the three month postoperative questionnaire, 364 (82%) the six month questionnaire, and 359 (81%) the 12 month questionnaire. Although differences between respondents and non-respondents after 12 months were slight (table 1), non-respondents were more likely to have had more severe symptoms (25/76 (33%) non-respondents had severe symptoms v 78/340 (23%) respondents).
Surgeons provided postoperative data on 278 of the 442 women (63%). These women were similar to the 164 for whom such data were not available (difference in mean age 0.5 years (95% confidence interval -2.4 to 1.9); distribution of preoperative symptom severity 2 0.8, df=4, P=0.9). Respondents were more likely to have had a colposuspension (149/262 (57%) v 83/180 (46%)) and less likely to have had a needle suspension (18/262 (7%) v 47/180 (26%)) (P<0.001).
Severity of stress incontinence
Improvements reported three months after surgery persisted throughout the first year after surgery (table 2). Repeated measures analysis of variance of the distribution of symptom severity scores over the 3 to 12 month period gave a significant result (P=0.003). The data suggest that there was a slight deterioration at 12 months (difference between mean preoperative and mean score after three months of follow up was 7.0 compared with 6.6 at 12 months).
The amount of improvement depended on severity of stress incontinence before surgery (fig 2). Women with severe symptoms (scores 13) obtained a greater reduction in severity scores than women with moderate or mild symptoms (scores 12) The more severe a woman's symptoms were before surgery the more severe her symptoms after surgery. Thus, women with preoperative scores of 1-8 had scores of about 3 after surgery, whereas women with preoperative scores of 13-20 had mean scores of about 7 after surgery.
Data on the extent of changes in the severity of symptoms were available from women and surgeons for 193 of the 278 (69%) women whose surgeons provided information. For 108 (56%) the two assessments were similar, for 28 (15%) the surgeons underestimated the benefits of surgery, and for 57 (30%) the benefits were overestimated.
Many believe that women with genuine stress incontinence respond better to surgery than women who have not had urodynamic pressure studies performed. Of the 285 women for whom information was available from their surgeons, 164 had undergone pressure studies and 103 had not (there were no data for 18 women). Comparison of these groups showed a similar difference in mean preoperative symptom severity score: 12.2 for women with genuine stress incontinence v 14.6 for women whose stress incontinence was not diagnosed urodynamically (difference 2.4 (1.4 to 3.4)). One year after surgery the difference was even less: 5.7 v 6.0 (difference 0.3 (-1.2 to 1.8)). The proportions of women cured (continent) were also similar: 24% v 26% (P=0.8).
Impact on other symptoms and comorbid
Surgery had some effect on bowel function. A year after surgery, fewer women described their function as "about right" (167/350 (48%) after surgery v 269/433 (62%) before surgery; difference 14% (9% to 20%), P<0.001) and more described it as "variable." There was little change in the proportions who saw themselves as "constipated" (56/350 (16%) v 79/433 (18%)), or as having "diarrhoea" (14/350 (4%) v 13/433 (3%)).
When the severity distribution of comorbidity was compared before surgery and one year later it was largely unchanged (table 4). Similarly, there was no significant change in the distribution of women's perceptions of their general health (table 4).
After discharge, women continued to report complications. During the first three months after surgery 31 (9%) reported difficulty urinating, 25 of whom required catheterisation. This was a rare event after three months. Although other complications were reported, these rarely resulted in a visit to a general practitioner or hospital specialist.
About 65% to 70% of patients rated the outcome as, or better than, expected. This view remained largely unchanged during the first year after surgery. A similar proportion felt their recovery had been as, or faster than, expected. In contrast to the expected outcome, patients tended to be less critical of the speed of their recovery as time progressed (one year later 95 (27%) felt it was slower than expected compared with 133 (37%) after three months). These assessments were consistent with the finding that a year after surgery only 239 (68%) would recommend surgery to a friend (table 6).
Surgeons tended to view the outcome of surgery more optimistically. They felt the outcome was as expected or better than expected for 176 out of 207 women (85%) and worse in only 31 (15%) cases. Surgeons would have changed their clinical management in only 13 (6%) women.
Outcomes of surgery
The benefits of surgery were not confined to improvements in stress incontinence. The prevalence and severity of several other urinary symptoms also decreased. Unexpectedly, these included urgency, a symptom believed to worsen after surgery and often considered to be a contraindication.3 4 5 6 The prevalence of women with severe urgency fell from 41% before surgery to 24% after.
The mental health of many (63%) women improved with the reduction in the severity of their stress incontinence. The finding that 25% of women reported worse mental health after surgery is of concern. It may reflect disappointment in the outcome of a treatment that many women regard as their last hope.
Postoperative complications were more common than previous studies have suggested. This may be because our data came from patients rather than surgeons and because previous studies have been done in specialist centres and involved surgeons with a particular interest in the subject. Given that 1 in 6 women reported difficulty urinating for up to three months after surgery, patients should be warned of this potential problem. The high prevalence of inadequately controlled pain after surgery indicates the need for better pain management. Better preoperative information may reduce the need for postoperative analgesia.13
Responses to the global questions on women's views of their treatment were remarkably consistent. About 70% were satisfied with the outcome. While this is reassuring for potential patients and for surgeons, it is not as optimistic as the views of the surgeons in this study (or reported in previous studies).2
Implications for clinical practice
We thank all the women who completed questionnaires, the doctors and nurses who recruited patients and supplied clinical information, and Jenny Stanley for help with administering the study.
Funding: The Health Services Research and Public Health Board of the Medical Research Council provided funding for the study.
Conflict of interest: None.
This article has been cited by other articles: (Search Google Scholar for Other Citing Articles)
Read all Rapid Responses
|Home||Help||Search||Archive||Feedback||Table of Contents|