Impact of surgery for stress incontinence on morbidity: cohort study

Recruitment and response

During the study period 650 women had surgery for stress incontinence, of whom 631 (97%) were invited to participate. Questionnaires were returned by 521 (83%) women. However, 62 of the 521 (12%) questionnaires were received late (and we could not guarantee that they had been completed before the women were aware of the short term outcome of their surgery) and 17 of the 521 (3%) were not completed. Thus, 442 out of 650 (68%) women were successfully recruited. Non-participants were similar to participants in age (mean age 52 years in both cases), length of hospital stay (mean 8.3 v 8.1 days), and type of surgical procedure (colposuspension 116/197 (59%) v 223/425 (53%); anterior colporrhaphy 46/197 (23%) v 129/425 (30%); needle suspension 28/197 (14%) v 53/425 (13%); other 7/197 (4%) v 20/425 (5%)). Surgeons completed perioperative questionnaires for 285 of the 442 (64%) women recruited into the study. There was no significant response bias: there was no difference in the mean age of patients (52 years) or in the frequency distributions of scores on the symptom severity index (P=0.09).

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 χ² 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

Before surgery 341 out of 416 women (82%) reported moderate or severe symptoms (symptom severity index ≥9) (Table 2). Three months after surgery 288 out of 348 women reported an improvement (83%), 17 experienced no change (5%), and 27 (8%) reported that their condition worsened (fig 1). Altogether 92 out of 333 (28%) women achieved continence and an additional 59 out of 333 (18%) suffered from incontinence only once a month or less. The amount leaked had also reduced: before surgery 135 out of 442 (31%) described their usual loss as a cupful or more, but after surgery only 31 out of 360 (9%) still experienced such losses. This was reflected in a change in pad use: 105 out of 357 (29%) were using pads after surgery, compared with 333 out of 431 (77%) before.

Fig 1

Comparison of distribution of symptom severity scores before and 3 months after surgery. A score of 0 indicates that patient is considered cured

• Download figure
• Open in new tab
• Download powerpoint

Fig 2

Mean scores on the symptom severity index before and 3, 6, and 12 months after surgery

• Download figure
• Open in new tab
• Download powerpoint

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 ≤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 conditions

While surgery had no impact on the prevalence of some urinary symptoms (cystitis, straining to start, hesitancy), it led to an improvement in others which lasted for at least 12 months. There were improvements in frequency (micturition every hour or less 42% to 18%; difference 24% (19% to 29%), P<0.001); nocturia (micturition twice or more a night 52% to 27%; difference 25% (21% to 30%), P<0.001); urgency (not able to wait more than 2 minutes 41% to 24%; difference 17% (13% to 21%), P<0.001); postvoid fullness (ever 73% to 54%; difference 18% (14% to 22%), P<0.001); and dysuria (ever 53% to 33%; difference 20% (16% to 24%), P<0.001) (Table 3).

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).

Mental health

Before surgery, 265 out of 424 women (63%) reported some mental health problems (index scores of ≥5) (Table 4). This fell to about 134 out of 343 (39%) three months later (McNemar's test, P<0.0001) and was sustained over the following nine months. Overall, 194 out of 309 (63%) reported an improvement in mental health, 36/309 (12%) reported no change, and 79/309 (26%) reported worsening mental health.

Postoperative complications

Altogether 67 out of 366 women (18%) reported “a great deal” of pain during the first 24 hours after surgery despite receiving postoperative analgesia. A further 130 (36%) reported “a fair amount.” Only 37 (10%) reported fully successful pain control. Overall, 119 out of 360 women (33%) reported one or more complications during their hospital stay, the commonest being urinary retention (76/360 (21%)) and vaginal bleeding (47/360 (13%)) (Table 5).

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.

Overall impact

About 70-75% of patients felt better three months after surgery (Table 6). This perceived benefit persisted throughout the first year after surgery. About 17% of the others reported no benefit from the operation and, for the first six months about 7% felt their health had deteriorated. The proportion who felt worse rose during the second six months so that one year after surgery 41 (11%) reported feeling worse than before surgery.

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

While most women benefited from surgical treatment for stress incontinence, surgery was not as successful as claimed in most textbooks and reported by other studies. Although 87% reported some improvement in their condition, only 28% achieved continence (compared to previous claims of up to 85%).2 This difference probably occurred because we used women's reports of their symptoms, unlike previous studies which have used the operating surgeon's views.

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

Methodological limitations

Rates of both recruitment (68%) and patient response (81% at 12 months) were high. If any bias was introduced it was that the women studied were less likely to be severely incontinent (55% of participants severely incontinent v 70% of non-participants). This bias would reinforce the finding that poor outcomes are more common than previously believed. There was also concern over the relatively low response rates of surgeons to the perioperative (64%) and postoperative (63%) questionnaires. There was a greater likelihood of surgeons reporting on women who had undergone a colposuspension rather than needle suspension. The effect this may have had on the outcomes reported by surgeons is unclear. One further limitation was the relatively short outcome period. There is considerable evidence that the benefits of surgery deteriorate over time so it would be desirable to reassess the study participants again after a few years.

Implications for clinical practice

The findings of this study have six implications for clinical practice. Firstly, they show that it is possible to collect standardised data from patients on the severity of their stress incontinence rather than relying on surgeons' impressions. Secondly, there is a need to reassess the information given to women considering surgery. Thirdly, there is clearly a need for better postoperative pain control; many women are suffering unnecessarily. Fourthly, we found no evidence that urgency and urge incontinence are contraindications to surgery; indeed, surgery was associated with a reduction in the prevalence of this problem. Fifthly, the role of urodynamic testing needs reappraisal. As patients were not randomised it is not possible to conclude with certainty that urodynamic testing has little or no prognostic value. The only alternative explanation for our findings is that the 39% of women who had surgery without urodynamic confirmation of genuine stress incontinence were carefully (and accurately) selected by surgeons on the basis of their medical history and clinical examination. And finally, the consistent reports of outcome between 3 and 12 months after surgery suggest that care of these patients can be audited at any time during this period. This allows follow up of patients in batches rather than necessitating the organisation of a continuous system in which all women are followed up at the same point in time after their operation.