Women with symptoms of uncomplicated urinary tract infection are often willing to delay antibiotic treatment: a prospective cohort study

Acute uncomplicated urinary tract infections (UTIs) are infections of the lower urinary tract in otherwise healthy, non-pregnant, adult women without known anatomical or functional abnormalities of the urinary tract. The symptoms are bothersome and have a negative impact on quality of life [1‐4]. Although empiric antibiotic treatment of all women with urinary symptoms has been reported to be cost-effective [5, 6], bacterial resistance is rising [7‐10] and strategies to reduce antibiotic use are needed.

Placebo arms of randomized trials have shown that 25 to 50% of women presenting with UTI symptoms will have recovered in one week without using antibiotics [11‐13]. Moreover, qualitative research has suggested that these women often want to avoid taking antibiotics and may prefer delayed antibiotic treatment [14]. Therefore, antibiotic use might be reduced if all women with UTI were asked to delay treatment.

We investigated how many women with UTI symptoms were willing to delay antibiotic treatment when asked by their general practitioner (GP). In addition, we explored how many of these women reported not to have used antibiotics after one week and whether their symptoms had improved.

In total 205 women were eligible, of whom 29 were excluded (Figure 1). GPs participating in one of the 20 GP practices did not ask any of their 25 patients to delay antibiotic treatment, because they disagreed in principle with this approach. No differences in main patient characteristics (age, socioeconomic status, number of diagnosed UTIs in the past) were present between patients of this non-participating practice and those of the other practices. For four patients, GPs did not report whether they had been asked to delay.

Of the remaining 176 patients, 137 were asked by their GP to delay antibiotic treatment and 39 were not. For these two groups, median numbers of recruited patients per practice were 5 (range 0–33) and 1 (range 0–9), respectively. Table 1 shows the main characteristics of the two groups. Overall UTI prevalence was 59% (103/176). Women who reported at least considerable pain and women who thought they had a UTI were more likely to be asked to delay antibiotic treatment. Of those asked, 37% (51/137) were willing to delay (Figure 1). To investigate the variation between practices with respect to the proportions of women willing to delay, we calculated a weighted mean proportion across practices (random-effects meta-analysis), which yielded a similar result (39% (95% CI 27-52%)). Severity of symptoms did not differ significantly between women willing to delay and women not willing to delay (Table 2). The median number of recruited patients per practice was 2 for women willing to delay (range 0–18) and 4 for women not willing to delay (range 0–15).

After one week, 55% (28/51) of women who were willing to delay reported not to have used antibiotics and 71% (20/28) of these women reported clinical improvement or cure (Table 3). Of these patients, 35% (7/20) had a positive baseline culture. Of the eight women not reporting clinical improvement or cure after one week, two turned out to have had a positive culture at baseline.

Of the 51 women willing to delay, 15 (29%) had used an antibiotic after one week, of whom all reported clinical improvement or cure and 13 had a positive baseline culture. Eight of the 51 women (16%) did not report on their antibiotic use.

Of women who had been asked by their GP to delay treatment, 25% (34/137) reported not to have used antibiotics after one week, against 15% (6/39) of women who had not been asked, yielding a difference of 10% (95% CI −6 to 21%). Of the six women not being asked to delay, five had cure or improvement of symptoms, of whom one had a positive baseline culture. Of the 137 women who were asked to delay, 30 did not report on antibiotic use after one week (Figure 1), 22 of whom belonged to the group not willing to delay at baseline. Of the 39 women who were not asked to delay, one did not report on antibiotic use.

The result of the baseline culture was known only after the follow-up period of one week. It turned out to be positive for 51% (26/51) of the delaying women and for 67% (58/86) of the non-delaying women (risk difference 16% (95% CI 0-33%)) (Table 2).

No patients developed pyelonephritis during the follow-up period of one week, implying a 95% CI of 0-12% for the 28 delaying women who had not used antibiotics after one week.

Our results suggest that more than a third of women with UTI symptoms are willing to delay antibiotic treatment when asked by their GP. More than half of these women (28/51) will not have used antibiotics after one week, of whom more than 70% (20/28) will have improvement of their symptoms.

The proportion of women reporting the use of antibiotics was 10% lower in the group that had been asked to delay as compared to those that had not been asked. The true reduction may be somewhat higher, since in the group asked to delay 30 patients were analyzed as having used antibiotics because no follow-up data were available. However, it should be taken into account that not all baseline characteristics were similar between these two groups (in the group asked to delay, more women reported at least considerable pain and/or thought they had a UTI).

As far as we know, this is the first study that describes the proportion of women with UTI symptoms that are willing to delay antibiotic treatment. Our findings are consistent with the results from a qualitative study by Leydon et al., which revealed that patients do not always want to use antibiotics, although clinicians often assume that they do [14]. This misinterpretation by clinicians might be illustrated by the fact that not all eligible patients in our study were asked to delay antibiotic treatment by their GP.

In addition to the 28 women willing to delay treatment, six women not willing to delay had not used antibiotics after one week. Although being asked to delay might have stimulated these six women in changing their minds to not using antibiotics, we did not obtain information on their motives. For five of these women, symptoms were cured or improved, one having had a positive baseline culture.

We had no further information on reasons for GPs not to enquire after their patients’ willingness to delay, which may be seen as the main limitation of our study. However, patients who were excluded by their GP did not have worse baseline characteristics than those who were included. On the contrary: women who reported at least considerable pain and women who thought they had a UTI were more likely to be included. This suggests that the GPs’ decisions whether to ask patients to delay or not may be based more on their personal attitude towards antibiotic prescriptions than on patient characteristics, which is in line with the previously mentioned findings of the study by Leydon et al [14]. In addition, patient attitudes (e.g. demanding personality) and previous experiences (e.g. problematic UTI history) might influence the GPs’ decisions.

Another limitation is the large number of missing data in the group of patients who were not willing to delay antibiotic treatment. More than a third of these patients did not report on antibiotic use and symptom improvement after one week. We consider it likely that most of these patients were cured and therefore did not feel inclined to report on follow-up results.

Due to cultural differences, our results might not be completely generalisable to other countries, where women may be less willing to delay antibiotic treatment than in the Netherlands. However, similar patient attitudes towards antibiotic prescriptions have been reported for the United Kingdom and Hong Kong [14, 21, 22].

Whereas we evaluated improvement of symptoms after a week, we did not obtain follow-up urine samples. Therefore we can not draw any conclusions on microbiological cure.

An uncommon but severe complication of UTIs is pyelonephritis, which may be a reason to treat all women with a suspected UTI. However, placebo arms of randomized trials suggest that cystitis seldom progresses to pyelonephritis [11‐13]. Similarly, no women developed pyelonephritis in our study population.

Since in clinical practice the result of the urine culture is not available at the moment that a treatment decision is taken, we included all patients with symptoms, independent of the result of their urine culture. We consider this a more pragmatic approach than including only patients with microbiologically proven UTI.

Our findings imply that antibiotic use might be considerably reduced by simply asking women with UTI symptoms whether they are willing to delay antibiotic treatment. Besides, most of delaying women will have symptom improvement after one week. This is in line with results from a randomized trial by Little et al., in which a treatment strategy of delayed prescription reduced antibiotic use by 20% while yielding the same symptom control when compared to a strategy of immediate antibiotics [23]. A potentially promising strategy being currently under investigation is initial treatment with pain medication instead of immediate antibiotics [24, 25].

Our results support previous findings that women with UTI symptoms may be more receptive to delayed antibiotic prescriptions than is assumed by many clinicians. If all of these women were asked to delay antibiotic treatment, antibiotic use might be substantially reduced without negatively impacting clinical recovery. As a safe alternative to antibiotics (without the risk of bacterial resistance), symptomatic treatment may be offered.