Contribution of Common Medications to Lower Urinary Tract Symptoms in Men FREE

While benign prostatic hyperplasia (BPH) is the most common cause, the etiology of lower urinary tract symptoms (LUTS) is multifactorial. Clinical guidelines for BPH suggest evaluating other potential sources of LUTS (eg, concomitant medication use) before initiating pharmacotherapy or surgical intervention.¹

Through effects on detrusor muscle and urinary sphincter function, several categories of prescription drugs can worsen LUTS,^2- 3 including antidepressants, antihistamines, bronchodilators, anticholinergics, and sympathomimetics.⁴ By increasing urine volume, diuretics are also associated with LUTS.^5- 6 Because their prevalence of use rises with patient age, prescription drugs may contribute to the age-related increase in LUTS.^7- 8

Most previous research has focused on LUTS overall, without accounting for heterogeneity in etiology. If a significant proportion of LUTS can be attributed to medication use, the observed association between a suspected risk factor and LUTS could be attenuated. Thus, it is important to determine the magnitude of the association between medications and LUTS, and the degree to which LUTS may be attributed to these medications. To this end, this study assesses the cross-sectional association between current use of selected common medications and LUTS among men enrolled in the California Men's Health Study (CMHS).

A detailed description of CMHS has been previously published.⁹ Briefly, with institutional review board approval, men aged 45 to 69 years were recruited from the Kaiser Permanente California Medical Care Program (southern and northern regions). Through a 2-stage process, 84 170 participants completed questionnaires and were eligible for study inclusion using the baseline survey data and linked electronic health records between 2002 and 2003.

Men with surgery for enlarged prostate (n = 1601), prostatitis (n = 5547), or prostate cancer (n = 5487) were excluded. Other exclusion criteria included musculoskeletal conditions (n = 10 668), neurologic disorders (n = 2208), and cancers including bladder and colon (n = 1238). For the 63 579 remaining subjects, pharmacy records were assessed to determine prescription drug dispensation for antidepressants, antihistamines, bronchodilators, nonurinary anticholinergics, sympathomimetics, and diuretics during the baseline survey period with at least 1 prescription filled.

Severity of LUTS was assessed using the American Urological Association Symptom Index (AUASI), and additional survey data included age, race/ethnicity, and diagnosis of BPH.

Associations between drug exposures and prevalent LUTS were assessed using logistic regression models, expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Primary analyses considered dichotomized AUASI, with 0 to 7 (mild) as reference vs 8 to 35 (moderate and severe), while secondary analyses stratified the latter group into moderate (8-19) and severe (≥20). Each drug category was assessed individually and collectively as a single indicator. Multivariable logistic regression models adjusted for age, race/ethnicity, region, and presence of BPH. The ORs and prevalence of drug exposures were used to estimate etiologic fraction.

Baseline characteristics stratified by severity of LUTS demonstrated that older subjects displayed a progressively greater proportion of moderate to severe LUTS, with 10% of participants aged 70 to 74 years reporting severe LUTS. Asian Americans had a greater prevalence of mild LUTS, and African Americans had a greater proportion of moderate to severe LUTS.

Medication use varied by race/ethnicity (P < .001), with more Asian participants using antihistamines (16%) and African Americans using proportionately more diuretics (31%). Medication use increased with age (P < .001), particularly for bronchodilators, anticholinergics, and diuretics, with a highest rise in diuretic use from 8% to 23% in participants aged 45 to 49 years to 70 to 74 years, respectively.

In unadjusted analyses, every medication class was associated with LUTS. When adjusted for region, age, and race/ethnicity, bronchodilators and antidepressants had the strongest associations with adjusted ORs of 1.22 and 1.39, respectively. The cohort was then stratified by presence of BPH, with 14 215 participants having BPH. Significant associations were found more often among men without BPH, particularly for sympathomimetic, diuretic, and antidepressant medications (Table).

When study medications were combined into any vs none, the OR was 1.29 (95% CI, 1.25-1.34) without adjustment for BPH and 1.28 (95% CI, 1.24-1.32) with adjustment for BPH. When stratified by presence of BPH, the OR was 1.31 (95% CI, 1.27-1.36) for those without BPH and 1.16 (95% CI,1.08-1.25) for those with BPH.

Estimates of etiologic fraction suggest that antihistamines, bronchodilators, diuretics, and antidepressants account for approximately 1%, 2%, 3%, and 4% of LUTS, respectively. These percentages were slightly attenuated in men with BPH. Use of any of these medications could account for 10% of LUTS, compared with 29% of symptoms associated with BPH.

This study documents the association between specific categories of common medications and LUTS in a large, diverse population of community-dwelling men. If the observed associations represent cause and effect relationships, these medications account for 10% of LUTS, an effect approximately one-third of that for BPH. This underscores the importance of assessing medications in the differential diagnosis of LUTS.

Another implication is the importance of these medications in studies of risk factors for LUTS. An outcome based solely on the presence of LUTS would include a substantial number of men with iatrogenically caused symptoms. This could dilute the association between the risk factor of interest and LUTS, obscuring a true association and leading to an incorrect conclusion that the risk factor is not associated with LUTS. In addition, the effect of these medications could explain some portion of the well-documented age-related increase in prevalence of LUTS among women that is similar in men.

The cross-sectional study design limits the ability to establish a causal relationship, and the measures of etiologic fraction represent an upper bound of the proportion of LUTS that can be attributed to these medications. There are also inherent limitations in defining medication use, since over-the-counter medications were not considered. Moreover, LUTS assessed at one discrete time may not reflect participants' typical symptom severity.

This study's significance is based on the intrinsic importance of evaluating LUTS, affecting both the research and clinical realms. Previous studies have likely underestimated these iatrogenic symptoms, leading to misclassification in observational studies of LUTS and masking of true associations that may have been amenable to intervention. Evolving treatment guidelines should consider common medication use and modify drug treatment to reduce the burden of LUTS without additional medications or invasive therapy.