How should patients with an overactive bladder manipulate their fluid intake?


Hashim Hashim, Bristol Urological Institute, Southmead Hospital, Bristol BS10 5NB, UK.



To assess how the symptoms of overactive bladder (OAB) syndrome in adults are affected by decreasing or increasing fluid input, and the effect of fluid manipulation on quality of life.


Adults with symptoms of OAB were randomized in a two-group, prospective, cross-over trial following a 4-day screening period (to establish baseline values) using frequency/volume charts, if they had a mean of eight or more voids and one or more urgency and/or urgency incontinence episodes in 24-h. Patients were asked to either increase or decrease their fluid intake, from baseline, as follows: group 1, 4 days drinking 25% less than baseline, and 2 days normal, then 4 days at 50% less than baseline, and 2 days normal, then 4 days at 25% more than baseline, and 2 days normal, and then 4 days at 50% more than baseline. Group 2 did the reverse. The primary endpoint was the change in the frequency of unwanted events during a 24-h period.


In all, 67 patients were contacted, 40 recruited and 24 were eligible. There was a significant reduction in frequency, urgency and nocturia when patients decreased their fluid input by 25%. Increasing fluid input by 25% and 50% resulted in a worsening of daytime frequency. Overall there was no statistically significant improvement in quality of life but there was some subjective improvement in urgency and nocturia, as measured by a validated instrument, in the group that decreased their fluid input by 25%; ≈300 mL of daily fluid came from water-containing foods.


Fluid manipulation is a cheap, noninvasive and easy way to help control the symptoms of OAB. Patients have difficulty in either decreasing or increasing their fluid input by 50%. Patients can now be told to expect a significant improvement in urgency, frequency and nocturia episodes if they reduce their fluid input by 25%.


detrusor overactivity


overactive bladder


International Consultation on Incontinence


frequency-volume chart.


Overactive bladder (OAB) is a symptom syndrome consisting of urgency, with or without urgency incontinence, usually with frequency and nocturia, if there is no confirmed infection or other obvious pathology [1]. One of the main aims of this definition of OAB is that initial management can be initiated based on a clinical diagnosis rather than having to resort to filling cystometry, which is invasive and costly.

The International Consultation on Incontinence (ICI) [2] guidelines recommend that conservative treatment should be the first-line therapy for OAB. This includes lifestyle interventions, pelvic floor muscle exercises and bladder training. Lifestyle interventions includes fluid manipulation, avoiding water-containing foods such as fruits and vegetables in the evening, avoiding excessive fluid intake for 4 h before going to bed, emptying the bladder before going out and before going to bed at night, stopping smoking, and reducing weight.

Fluid manipulation has been widely practised to control urinary symptoms, as part of a behavioural approach, but there is little reported evidence of its efficacy, with very few published randomized trials [3–8]. The problems with previous studies were that there was poor compliance with study protocol, assessed cognitively impaired geriatric patients, or included only women. Most important, none of the previous studies assessed urgency, the cornerstone symptom of OAB, and concentrated mainly on incontinence. The studies also did not evaluate the effects of the various interventions on quality of life.

Essentially, there is no consensus as to what to tell patients; some physicians advise increasing fluid intake while others advise decreasing it. The other problem is that even when advising patients to increase or decrease fluids, there are no published data to tell patients when this reduction or increase should be, and by how much.

In the present study we assessed how the symptoms of OAB, in adults, are affected by decreasing or increasing fluids in a randomized-controlled cross-over trial, with the aim of trying to define a volume of fluid manipulation that patients can use. We also assessed the effects on the quality of life of patients, using a validated questionnaire, the ICIQ-OAB [9].


Adult men and women (≥18 years old) with symptoms of OAB were randomized in a two-group, prospective, cross-over trial after a 4-day screening period (to establish baseline values) using frequency/volume charts (FVC), completed daily, if they had a mean of eight or more voids and one of more urgency and/or urgency incontinence episodes in 24 h. Patients were identified from a database that included those who were contacted and/or participated in previous trials. They were initially screened by telephone to see if they agreed to take part in the trial and were eligible for it. If they were eligible, all the trial information and paperwork, including the baseline FVC and consent forms, were sent to them by mail, and they were then reviewed in the clinic. The information recorded on the FVC included the volume and type of fluid drunk over 24 h, the time and volume of urine produced, the time of urgency or leakage episodes, and the number of pads used. Once the FVC was completed, the patients attended for a visit to see if they were eligible, based on the inclusion and exclusion criteria. Patients completed a 4-day FVC in each part of the study, and the ICIQ-OAB at the end of each 4-day period, to assess their quality of life for that period. The study was approved by the local research ethics committee.

Patients were excluded from any part of the study if their increase/decrease resulted in them drinking >3 L or <1 L of fluid, as it was thought, based on a previous study [10], that drinking >3 L would be excessive and drinking <1 L would cause symptoms of dehydration, e.g. headaches and constipation. Patients were also excluded if they were pregnant or breast-feeding; had haematuria, bacteriuria, pyuria, proteinuria, glucosuria or ketosuria on urine dipstick testing; had a residual volume of >150 mL, as assessed by a bladder scan; had uncontrolled hypertension, suspicion or evidence of clinically relevant cardiac failure, renal disease or hepatic disease; were diagnosed or suspected of diabetes insipidus/primary polydipsia or diabetes mellitus; had neurogenic dysfunction of the lower urinary tract; were known alcohol or drug abusers; were scheduled to be admitted to hospital for inpatient surgery during the trial; had any history of clinically relevant psychiatric disorders within the last 24 months preceding enrolment in the trial; had a history of not complying with medical regimens or were not compliant with protocol requirements or unable to keep a diary or perform the required volume measurements on their own; had significant pelvic organ prolapse (Stage III or IV) or had significant stress urinary incontinence.

Patients were initially randomized into one of two groups and asked to either increase or decrease their fluid intake first, from baseline (Fig. 1). The primary endpoint was the change in the frequency of unwanted events (day and night voids, urgency and urgency incontinence episodes) during a 24-h period. The secondary endpoint was to assess the change in the quality of life of patients with OAB, as assessed by the ICIQ-OAB.

Figure 1.

The study design.

The calculation of a statistically valid sample size was based on a previous published study [10]. Interpolation was used for each observation in the previous study to estimate the voiding frequency for a 25% increase/decrease in fluid intake, and the sd of the paired differences; assuming a linear relationship between change in fluid intake and change in voiding frequency. In the period with a 25% increase, a sample size of 11 would have 80% power to detect a difference in mean voiding frequency of 1.2, assuming a sd of differences of 1.24, using a paired t-test with a 0.05 two-sided significance level. In the period with a 25% decrease, a sample size of 24 would have 80% power to detect a difference in mean voiding frequency of 0.8, assuming a sd of differences of 1.33, using a paired t-test with a 0.05 two-sided significance level. For the 50% increase and decrease periods, the number of patients required to detect a difference was assumed to be smaller and therefore no formal calculation was made. It was assumed that 30% of patients contacted would not be eligible for the trial on initial screening, and then another 30% would also not be eligible after the second screening period, and 10% would withdraw from the study after randomization. The sign test was used to analyse the data. The data from patients in each of the four fluid manipulation periods were compared with the data from the baseline period.


In all, 67 patients were contacted, 40 recruited and 24 fulfilled the inclusion criteria and were eligible for the trial (Table 1; Fig. 2). All 24 patients participated in the period with a 25% reduction but not all participated in the other parts of the study because they did not fulfil the criteria, i.e. their input was <1 L or >3 L when the fluid intake was manipulated (Table 2). The mean fluid output was ≈289 mL higher than the fluid intake.

Table 1. 
The characteristics of the patients
No. of patients40 24
Mean (median, range):
 Age, years60.5 (61.0, 41–80) 62.7 (62.5, 42–80)
n (%):
 Male19 (48)  11 (46)
 Female21 (52) 13 (54)
Duration of symptoms, months
 Mean 144.5
 Median  60.0
Figure 2.

The patient disposition during the study.

Table 2.  The median fluid input and fluid output, and median (range) episodes of urgency, urgency incontinence, daytime frequency and nocturia at baseline (24 patients), and during each 4-day fluid manipulation period. The baseline for each of the manipulation periods is also shown for reference, as the number of patients differed in each period. Statistical analysis (P) compared data from each period of fluid manipulation with the whole group baseline (24) shown in the first row
Drinking pattern (n)Input volume, mLOutput volume, mLUrgencyUrgency incontinenceDaytime frequencyNocturia
Baseline (24)185421653.8 (1–14.5)0 (0–4.8)9.5 (6.0–16.5)1.4 (0.3–4.5)
25% less (24)139618252.5 (0–13)0 (0–5.5)7.3 (4.8–13.0)1.3 (0–4.0)
P  0.0041.0<0.0001<0.0001
Baseline for 50% less (12)237626563.
50% less (12)124817923.1 (0–11.8)0 (0–4.5)7.3 (4.3–11.5)1.8 (0.5–3.8)
P  0.040.690.0010.04
Baseline for 25% more (21)167818103.809.51.5
25% more (21)217122903.3 (0–16.3)0 (0–10.3)10.8 (6.0–15.5)1.3 (0.25–5.0)
Baseline for 50% more (14)160617354.109.51
50% more (14)228824194.6 (0–13.5)0 (0–12.8) 11.4 (5.5–16.0)1.0 (0.3–5.0)
P  1.000.690.0021.00

When patients were asked to drink 25% more than their normal fluid input they only managed to drink 17% more, and when asked to drink 50% more they managed 23% more, when comparing the respective groups with baseline. When asked to drink 50% less, patients managed to drink 32% less than the baseline. The 25% reduction was adhered to quite well.

When patients reduced their fluid input by 25% they had a significant reduction in daytime frequency (23%), urgency (34%) and nocturia (7%). There was also a significant reduction in symptoms when they were asked to reduce fluid intake by 50%. There was a slight increase in nocturia when patients decreased their fluid input by 50%, which although statistically significant, tended to insignificance and might be a result of having too few patients in that group.

When patients were asked to increase their fluid intake by 25% and 50%, there was a statistically significant increase in daytime frequency with no increase in urgency or nocturia. In none of these fluid manipulations did urgency incontinence worsen.

There was statistical improvement in only two questions of the ICIQ-OAB, both in the 25% decrease regimen; the response to ‘Do you have a sudden need to rush to the toilet to urinate?’ improved from ‘sometimes’ to ‘occasionally’ (P = 0.04); and ‘During the night, how many times do you have to get up to urinate, on average? improved from ‘two’ to ‘1.5’ (P = 0.01).

At the end of the trial patients were asked which part of the fluid manipulation they found most helpful in terms of symptom improvement, and 83% said that the period of 25% reduction was the most helpful.

Adverse events were mild and tolerable. Four patients felt thirsty and two had headaches, constipation or concentrated urine when on the 50% decrease, and one had headache on the 25% decrease.


Conservative treatment and life-style interventions are first-line treatments in patients with OAB symptoms, but there is little published evidence of the effect of fluid manipulation on these symptoms. Dowd et al.[3] conducted a 5-week study on 58 women to determine the effects of hydration on the number of urinary incontinence episodes in those who were randomly assigned to one of three groups. There was poor adherence to the protocol in all groups, and the results were inconclusive and inconsistent in both the ‘maintain’ and ‘increase’ groups. Urinary incontinence episodes decreased most in the ‘decrease’ group. Qualitative results from this study indicated that increasing fluid intake was helpful in managing urinary incontinence. There was no relation between incontinence and caffeine intake in that study.

Wyman et al.[4] retrospectively investigated 126 women with urethral sphincteric weakness and/or detrusor overactivity (DO). They examined the relationship between the volume of oral fluid intake and the frequency of voluntary voids and urinary incontinence episodes. Overall, a higher fluid intake resulted in higher daytime and night-time voiding frequencies, and urinary incontinence episodes, but this was mainly due to a stronger and statistically significant relationship in the sphincteric incompetence group. There was no significant relationship between fluid intake and diurnal or nocturnal voiding, or incontinence episodes in patients with DO.

Griffiths et al.[5] investigated 76 women and 52 men with established incontinence. The study showed that diurnal and nocturnal voided volumes and urine loss were related to fluid intake, but there was a weaker relationship with voiding frequency.

Tomlinson et al.[6] showed that in 41 women, the increase in the average amount of fluid intake was significantly related to an increase in the average voluntary volume voided, but not to involuntary urine loss. Dallosso et al.[8] showed that there was no association between total fluid intake, including caffeine and alcohol, and OAB in women aged >40 years.

Swithinbank et al.[10] conducted a 4-week prospective, randomized crossover study in 30 women with urodynamically confirmed idiopathic DO to determine the effect of caffeine restriction, and of increasing and decreasing fluid intake, on urinary symptoms. There was no significant difference between the baseline week and the caffeine-free week for any outcome measures episodes, providing some evidence that caffeine restriction has no effect on OAB symptoms. Increasing fluid intake significantly increased voiding frequency and urgency episodes (each P < 0.003). The present study also showed an increase in frequency but not urgency, and this might be because in the previous study [10] the patients increased their fluid input to 3 L, while in the present they only managed to increase it to ≈2.3 L.

Decreasing fluid intake significantly decreased the voiding frequency (P < 0.003) and urgency episodes in the study by Swithinbank et al.[10]. This was similar to the results in the present study, but there was no quantification of how much patients should increase or decrease their fluids in the previous study; patients were instructed to drink fixed amounts, and from a practical perspective this is important, as patients often ask how much fluid they should drink. Decreasing the fluid intake was associated with a significant improvement in quality of life compared with the baseline week in women with idiopathic DO (P < 0.003). However, there was no difference in the effect on quality of life among any of the other weeks.

One drawback of the previous study [10] was that it only included women with a urodynamic diagnosis, while initial treatment of OAB is based on a clinical diagnosis. From epidemiological studies, the prevalence of OAB is almost equal in men and women, and therefore the present study included both. Also, the treatment of OAB is based on a clinical diagnosis rather than a urodynamic one, and therefore the present study aimed to be practical and recruited patients based on a clinical diagnosis rather than a urodynamic diagnosis.

We showed that reducing the amount of fluid intake by 25% from baseline improved daytime urinary frequency, urgency and nocturia; however, increasing the fluid intake made daytime frequency worse, but had no effect on nocturia or urgency.

In this type of study the FVC is an essential measurement instrument. It is thought that there is a bladder-training effect while completing urinary diaries, and therefore the study design randomized the order of fluid increase and decrease in an effort to minimize this effect. The 4-day periods were used because it was previously shown that a 3-day FVC gives the same amount of information as a 7-day chart in men [11] and a 4-day chart in women [12], and therefore to improve compliance with the trial a 4-day period was chosen.

The difference of ≈300 mL between fluid input and output is mainly attributable to the amount of water contained in food, which we did not measure in the study, but patients must be advised about this, as they often forget that fruits and vegetables are mainly water. Unfortunately, not all patients recorded their output during the 4 weeks, but they all recorded their output during the week with a 25% decrease, and thus it might be a truer reflection of the actual water content of food, i.e. ≈500 mL.

Interestingly, most patients managed to adhere to the reduction in their fluid intake by 25%, with a reduction of 458 mL, but almost all patients failed to reduce their intake by 50%; their reduction from baseline was 606 mL, and this might be because patients had already restricted fluid themselves, making it difficult to reduce fluid input further. This probably explains why there were no statistical differences between the groups reducing intake by 25% and the 50%. When asked to increase by 25% the patients managed to increase by 17%, and when asked to increase by 50% they increased intake by 23%, indicating that patients have difficulty increasing their fluid input. This probably reflects why there was little worsening in symptoms of OAB when fluid input was increased, except for increased frequency, and it might also reflect that patients associate increased fluid intake with worsening OAB symptoms.

This is the first study of fluid manipulation to assess the effects on nocturia, and although the difference was statistically significant, it was probably not clinically significant because the reduction was small (by 0.7 nocturia episodes/week), suggesting that fluid manipulation mainly affects daytime frequency. When fluid was decreased by 50%, nocturia increased from 1.4 to 1.8. This is probably a result of the statistical analysis, comparing the 12 patients in the 50% decrease group to the 24 at baseline. If the data were only taken from the same 12 patients and compared, then the baseline nocturia for this group would be 2.3, suggesting that nocturia actually decreased rather than increased, which is what would be expected. Nonetheless, the changes in nocturia do not seem to be clinically significant and this is probably because patients with OAB tend to manipulate their fluid input mainly during the daytime, and most attempt not to drink before going to bed, to help reduce nocturia. We did not assess when patients drank their fluid and did not restrict the time of fluid manipulation, because we tried to keep the regimen as practical as possible and to examine the concept of fluid manipulation rather than assess nocturia alone, which has a multifactorial aetiology.

Another factor that was not included in the evaluation was the ambient outside temperature, which might have effects on how much fluid patients drink, and on how much exercise they take. Both of these might explain why the changes in frequency, urgency and nocturia were not proportional to the changes in volume, i.e. a 25% or 50% manipulation in fluid input did not lead to a 25% or 50% proportional change in OAB symptoms.

We found no effect on urgency incontinence, possibly because only seven patients had one or more urgency incontinence episodes at baseline, and thus there were too few to derive any conclusions.

Interestingly, there was no statistically significant difference in subjective symptoms and quality of life using the ICIQ-OAB. Apparently the symptoms of urgency and urgency incontinence are more bothersome (both scoring a median of 7 on the visual analogue scale; possible score 0–10) than frequency (score 5). On closer analysis there seemed to be some improvement in quality of life for both urgency and urgency incontinence, ranging from 0.5 points to 2 points when fluid intake was reduced, and worsening of the score by 2 points on increasing fluid intake. As there are no studies available to quantify ‘minimally important differences’ when using the ICIQ-OAB, possibly a 2-point reduction is clinically significant and therefore important to patients. Studies assessing this aspect of the ICIQ-OAB are urgently needed to accurately interpret quality-of-life data in clinical studies. Another important factor is that it is difficult to measure urgency and there is still no universally accepted method to grade it, which will have consequences on how patients perceive or interpret the definition of urgency, and therefore on statistical significance. There were only seven patients with urgency incontinence at baseline and this would also limit any statistical analysis when assessing the effects on quality of life. As urgency and urgency incontinence seem to be the most bothersome to patients, and the driving force of OAB, the development of validated scales to assess urgency is urgently needed, so that they can be used in clinical trials to help assess the effects on quality of life of patients with OAB. Until such scales are developed, frequency remains to be the easiest and most direct objective measure for this ‘subjective’ disease.

In conclusion, fluid manipulation, as part of conservative treatment, is the first-line therapy for patients with OAB; it is cheap, noninvasive, easy to practise, provides almost immediate results, and has minimal side-effects. Patients can be advised to reduce their fluid input by 25% to help control all OAB symptoms, providing they do not drink <1 L/day, remembering that 300–500 mL of fluid is provided by food. Doctors and nurses should also avoid asking patients with OAB to drink more fluid, as this can worsen daytime frequency.


Sources of funding: Bristol Urological Institute PA Research Fund. We thank Nikki Gardener and Julie Ellis-Jones for their nursing support and Hazel Taylor for her statistical analysis of the data.


None declared.