In a large epidemiological study, authors investigated the effect of LUTS on quality of life among various cultures. They showed a close association between the two, and that the effect of having moderate symptoms has a similar effect on quality of life as diabetes, hypertension or cancer, and that having severe symptoms had a similar effect as a heart attack or stroke.
To investigate the effect of lower urinary tract symptoms (LUTS) on quality of life (QoL) and to determine its extent across a variety of cultures, and the confounding effects of self-reported comorbidities and demographics.
SUBJECTS AND METHODS
Data were obtained from two population-based studies in five cities: UREPIK (Boxmeer, the Netherlands; Auxerre, France; Birmingham, UK; and Seoul, Korea) and the Boston Area Community Health (BACH) study (Boston, USA). UREPIK used stratified random samples of men aged 40–79 years. BACH used a multistage stratified cluster sample to randomly select adults aged 40–79 years. QoL was assessed using a standard Medical Outcomes Study–Short Form 12 (SF-12, mental and physical health component scores); LUTS was assessed using the International Prostate Symptom Score (IPSS). The association between QoL and IPSS, associated illnesses, and lifestyle factors was investigated using weighted regression.
The UREPIK studied 4800 men aged 40–79 years; BACH recruited 1686 men aged 40–79 years. The prevalence of LUTS, defined as an IPSS of ≥8, varied by city (P < 0.001), with Auxerre reporting a prevalence (se) of 18.1 (1.2)%, Birmingham 25.6 (1.5)%, Boston 25.1 (1.6)%, Boxmeer 21.2 (1.3)%, and Seoul 19.0 (1.2)%. Overall, this was similar to the reported rate of high blood pressure. Severe LUTS, defined as an IPSS of ≥20, affected ≈ 3.3% of the age group; this was roughly similar to stroke (2.2%), cancer (4.5%), or heart attack (4.5%) and less than half as much as diabetes (8.6%). A 10-point increase in IPSS was associated with a 3.3 (0.3)-point reduction in SF-12 physical health component score, with the same effect in all cities (P = 0.682 for the interaction test). This was more than the physical health component score reduction caused by cancer, diabetes, or high blood pressure (2 points each), but less than stroke or heart attack (6 points). The comorbidities had no significant impact on SF-12 mental health component score (other than a heart attack, that had a 1.8-point reduction). A 10-point increase in IPSS was associated with a 3.4 (0.6)-point reduction of the mental health component score in the four western cities and a 1.4 (0.3)-point reduction in Seoul.
Increasingly severe LUTS is associated with a lower QoL. The effect of moderate LUTS on QoL physical health component score is similar to that of having diabetes, high blood pressure or cancer, while the effect of severe LUTS is similar to a heart attack or stroke. These changes were consistent across cultures. This analysis shows the magnitude and consistency of the effects of LUTS on QoL. While these patients might be seen by several types of practitioners, it is likely that urologists will be in the best position to recognize the true impact of LUTS on a patient’s QoL, to be aware of the effects of therapies for LUTS on QoL, and to ensure that colleagues in other disciplines recognize the importance of these symptoms and their treatment.
quality of life
body mass index
Boston Area Community Health (study)
Although LUTS and other comorbidities are associated with a reduced quality of life (QoL) in existing cross-sectional studies [1–3], the need for population-based epidemiological data is well-recognized . Previous studies reported that the decrease in QoL caused by LUTS is consistent across a wide variety of cultures [1,5,6]. We extend this work by comparing several communities using the same questionnaire in all cities and investigating if it is the severity of symptoms that is important, or if there are there cultural differences in the effect of LUTS on QoL.
There is ample evidence of the detrimental effect of long-term illness on QoL, particularly among the elderly who often have multiple concomitant illnesses . However, relatively few studies have assessed the joint effect of LUTS and concomitant illnesses on QoL. A study in Finland  established that the reduction in QoL associated with urinary symptoms was still present after adjusting for concomitant illnesses. Many concomitant illnesses are positively associated with LUTS, although only fecal incontinence, neurological disease, constipation and arthritis had statistically significant associations . In the present study we pooled data from an international collaborative epidemiological study in Europe and Asia with a community health survey in the USA, to help to obtain a better understanding of the joint association between LUTS and potentially life-threatening comorbidities and the health status of men.
We compared data from five cities: Auxerre, France; Birmingham, UK; Boston, USA; Boxmeer, the Netherlands; and Seoul, South Korea; from two studies, using the same symptom and QoL questionnaires. The presence and severity of LUTS was determined using the first seven questions of the IPSS, which is equivalent to the AUA Symptom Index. We determined the effect of LUTS on QoL, as measured by the standard Medical Outcomes Study–Short Form 12, and then established if the effect was common over all five cities. Furthermore, we considered the effect of comorbidities and demographics on QoL.
SUBJECTS AND METHODS
The UREPIK study is a population-based, cross-sectional survey completed in Boxmeer (the Netherlands), Auxerre (France), Birmingham (UK) and Seoul (Korea) . The study sample was based on an index male, and a random sample of men aged 40–79 years was recruited. Standard questionnaires, supplemented with extra questions, were used. Face-to-face interviews were held in Seoul; self-completed questionnaires were sent by mail in the other cities. The surveys were carried out from June 1998 to June 1999.
For UREPIK, a sample size of ≈ 1300 was required in each city to estimate the overall prevalence of LUTS in the cities to within 3%. This was calculated using a normal approximation, a 95% CI, specifying that its width must be 3% either side of the true value, and letting the true value range over 40–60%. In each city, stratified random samples of index men were selected. Local authority administrative registers were used in the Netherlands and Korea, GP registers in the UK, and electoral registers in France. Stratification by age was used in the Netherlands, France and Korea, with a target of 200 men in each 5-year age group from 50–54 to 70–74 years and 100 in age groups 40–44, 45–49 and 75–79 years. Stratification was by Townsend score , an index of social deprivation, of the GP practice in the UK. In Seoul, only couples were included in the study. No other inclusion or exclusion criteria were used.
Sampling weights were based upon the ratio of the number of men in each age group in the samples to the number of men in that age group in the population. These sampling weights also have the effect of countering response bias, as response rates for younger men were lower than for older men . All results presented are based on the weighted data. This is necessary because stratified random samples were used and the number selected in each age group was not proportional to the number of men in the population. As discussed below, as the population distribution by age group varied by city, we chose weights that sum to the city sample size and standardized the population to a standard ‘pseudo’ population.
In all, 4979 index men responded. In Birmingham, 179 men completed a short version of the questionnaire without full information on QoL, leaving a total of 4800 men. For the index man, the response rate was 72% in Boxmeer, 28% in Auxerre, 42% in Birmingham and 68% in Seoul. Adjusting the Birmingham response rate to the age distribution in the other cities gives an age-adjusted response rate of 60%.
The Boston Area Community Health (BACH) study was a population-based, random sample epidemiological survey of a broad range of urological symptoms. The research design had as its goal equal numbers of subjects in each of 24 cells, defined by age (30–39, 40–49, 50–59, 60–79 years), gender, and race/ethnicity (Black, Hispanic and White). The BACH multistage, stratified cluster sample (5506 subjects) was recruited from April 2002 to June 2005.
The city of Boston was stratified into 12 strata: four geographical areas by three levels of minority density. The geographical areas were formed by grouping Boston’s planning districts. The levels of minority density were: low-density minority (primarily White), high-density Black (≥25% of the residents were Black), and high-density Hispanic (≥30% Hispanic). Census blocks were randomly sampled from 4266 blocks in the city by stratum such that ≈ 10% of the low-density minority blocks, 15% of the high-density African-American blocks, and 75% of the high-density Hispanic blocks were selected.
Sampling proceeded in five batches, each a random subsample (or ‘mini-version’) of the overall BACH study . Households from selected census blocks were identified using a current Boston Resident List, geo-coded with census tract and block information for each individual. Telephone numbers were obtained from a telephone matching service for about half of the selected individuals. One individual per household was designated as the primary contact, with preference given to a person with a telephone number. Introductory letters were mailed to the selected households; these requested a contact telephone number if not already available (47.5% of the households). Households were screened either by telephone or by a field visit (in the absence of telephone number or if unable to reach by telephone). Screeners were completed for 36.0% of the selected households, 30.0% of the households refused screening, and 34.0% could not be contacted after ≥16 attempts to reach them by mail, telephone, or field visit.
Individuals from the selected census blocks were chosen according to eligibility rules to achieve approximately equal numbers of Black, White, and Hispanic respondents in four age categories: 30–39, 40–49, 50–59, and 60–79 years by gender. Eligibility rules varied by batch and were randomly assigned to selected households based on household demographics at the start of each batch. BACH eligibility criteria included: screened eligible from selected household, competent to sign informed consent, and able to speak English or Spanish well enough to complete the survey. In all, 2301 men and 3205 women, 1770 Blacks, 1877 Hispanics, and 1859 Whites were recruited. Interviews were completed with 63.3% of the screener-identified eligible individuals from the selected households.
Because of design requirements, the BACH subjects had unequal probabilities of selection into the study. For the analyses to be representative of the city of Boston, it was necessary to weight observations inversely proportional to their probability of selection into the study [13,14]. Weights were further post-stratified to the population of Boston according to the 2000 Census.
Data were obtained in a 2-h, in-person interview by a well-trained, bilingual phlebotomist/interviewer, generally in the subject’s home . After obtaining written-informed consent (all protocols and informed consent procedures were approved by the New England Research Institutes, NERI, Institutional Review Board), a venous blood sample (20 mL) and anthropometric measurements (blood pressure, height and weight) were taken, along with information on medical and reproductive history, major comorbidities, prescription and over-the-counter medications, lifestyles, psychosocial factors, medical care use and detailed self-reported major symptoms of seven different urogynaecological conditions (urinary incontinence, BPH, interstitial cystitis, chronic pelvic pain, prostatitis, erectile dysfunction, and female sexual dysfunction). Where possible, the questions and scales used on BACH were selected from published instruments with documented metric properties. To ensure the highest quality data, all staff were trained, certified, monitored and regularly retrained in all procedures and protocols. A minimum of 10% double data entry ensured accurate data computerization. Regular reports from NERI’s electronic data capture ADEPT software closely monitored all aspects of data completeness and quality.
A combined analysis of both surveys was carried out. As the population distribution by age group varied over the five cities, we chose weights that sum to the sample size in each city and standardized the population distribution in each city to a standard ‘pseudo’ population (Table 1). The relationship between the health status and age and other factors, e.g. disease severity and comorbidities, was estimated using linear regression. Statistical calculations were carried out using SAS  and SUDAAN  software.
Our main aim was to estimate the effect of LUTS and comorbidities on QoL. For each of the regression models, the following strategy was adopted. Interactions between LUTS and comorbidities and age and city were only included if they were significant at the 1% level. LUTS and age were both included as linear effects as using categories suggested that the linear effects were reasonable. Body mass index (BMI) was included as a categorical variable as there was evidence of nonlinearity in its effect. Confounding variables such as education and marital status were included if they were significant at the 5% level.
The IPSS was used to measure the severity of LUTS . Culturally and linguistically validated versions of this questionnaire were administered . Information on comorbid illnesses including diabetes, high blood pressure, stroke, heart attack, high cholesterol, and cancer was obtained through self-reported questions. QoL was assessed using the mental and physical component scores of the SF-12 . In the USA, the physical and mental health component scores in the adult population have a mean (sd) of 50 (10). In the analysis, adjustment was also made for age, BMI (self-reported in UREPIK and measured in BACH), living with a partner, education, and working status.
The analysis was based on 4800 men from the UREPIK study and 1686 from BACH, all aged 40–79 years. The age adjusted prevalence of moderate or severe LUTS (IPSS of ≥8) ranged from a low (se) of 18.1 (1.2)% in Auxerre to a high of 25.6 (1.5)% in Birmingham (Fig. 1); severe LUTS was more commonly reported in Birmingham and Seoul than in Boston. The trend of increasing prevalence with increasing age was similar in all cities, except in Boston, where the prevalence among men aged 70–79 years was lower than in men aged 60–69 years (Fig. 2).
There was variation among the cities in the reported prevalence of the various comorbidities: diabetes, high blood pressure, heart attack, high cholesterol, and cancer were all higher in Boston, sometimes considerably higher (Table 2). Frequently the lowest prevalences were in Seoul; the highest prevalence of stroke was in Birmingham; the BMI was much higher in Boston, with 34% of men reporting a BMI of >30 kg/m2, compared to <14% in the European cities and 4% in Seoul. Only 55% of the men in Boston currently lived with a partner, compared to 80–90% in Europe (by design, the Seoul sample contained only married men). Tertiary education was much higher in Boston, with 65% of men reporting this compared to 22% to 31% in the other four cities. About 60% of men were currently working in Boston and the European cities, compared to 78% in Seoul.
|Diabetes||8.6 (0.5)||9.0 (0.9)||6.1 (0.8)||13.8 (1.4)||4.5 (0.7)||7.3 (0.8)||<0.001*|
|High blood pressure||25.1 (0.7)||24.8 (1.5)||25.5 (1.5)||36.9 (1.8)||17.8 (1.3)||16.9 (1.2)||<0.001*|
|Stroke||2.2 (0.2)||1.8 (0.4)||3.6 (0.6)||2.6 (0.5)||1.9 (0.4)||1.2 (0.2)||0.002*|
|Heart attack||4.5 (0.3)||2.6 (0.4)||5.5 (0.8)||6.8 (0.9)||6.3 (0.6)||1.0 (0.3)||<0.001*|
|High cholesterol||24.5 (0.8)||33.4 (1.7)||18.7 (1.4)||37.7 (1.7)||16.9 (1.3)||11.4 (1.1)||<0.001*|
|UTIs||5.9 (0.4)||1.5 (0.4)||9.6 (1.0)||10.3 (1.0)||6.1 (0.8)||1.7 (0.4)||<0.001*|
|Cancer||4.5 (0.4)||3.7 (0.6)||3.0 (0.6)||10.1 (1.1)||2.8 (0.5)||1.0 (0.2)||<0.001*|
|0–7||78.2 (0.6)||81.9 (1.2)||74.4 (1.5)||74.9 (1.6)||78.8 (1.3)||81.0 (1.2)|
|8–19||18.5 (0.6)||15.6 (1.2)||20.8 (1.4)||23.0 (1.5)||17.8 (1.2)||14.6 (1.1)|
|20–35||3.3 (0.2)||2.6 (0.5)||4.8 (0.8)||2.1 (0.5)||3.4 (0.5)||4.4 (0.6)|
|≥8||21.8 (0.6)||18.1 (1.2)||25.6 (1.5)||25.1 (1.6)||21.2 (1.3)||19.0 (1.2)||<0.001*|
|<20||3.2 (0.3)||1.8 (0.5)||2.8 (0.6)||1.5 (0.3)||1.5 (0.4)||8.5 (0.9)|
|20–25||42.3 (0.8)||41.6 (1.8)||39.3 (1.8)||22.1 (1.5)||41.7 (1.7)||70.8 (1.5)|
|25–30||39.4 (0.8)||45.2 (1.8)||43.7 (1.8)||42.5 (1.7)||47.7 (1.8)||19.9 (1.4)|
|30–35||11.4 (0.5)||9.8 (1.0)||11.7 (1.2)||23.7 (1.6)||7.6 (0.9)||4.0 (0.8)|
|35+||3.6 (0.4)||1.7 (0.5)||2.6 (0.6)||10.2 (1.2)||1.4 (0.5)||0.00 (–)|
|Married or live with partner||80.2 (0.9)||85.6 (1.4)||78.2 (1.5)||55.1 (2.3)||88.7 (1.1)||100.0 (–)||<0.001*|
|> Secondary education||36.5 (1.0)||24.0 (1.67)||28.9 (1.7)||65.2 (1.9)||21.8 (1.5)||31.1 (1.6)||<0.001*|
|Currently employed||63.5 (0.8)||61.2 (1.6)||60.5 (1.7)||57.2 (2.3)||60.8 (1.6)||77.8 (1.1)||<0.001*|
|never||27.8 (0.8)||26.8 (1.7)||30.7 (1.7)||39.2 (1.8)||19.2 (1.4)||20.4 (1.3)|
|former||40.3 (0.8)||47.4 (1.8)||43.7 (1.8)||36.0 (1.8)||52.4 (1.8)||25.8 (1.4)|
|current||31.8 (0.8)||25.8 (1.7)||25.7 (1.6)||24.8 (1.8)||28.4 (1.6)||53.7 (1.7)|
|Continuous variables, mean (sem)|
|SF-12 component score:|
|physical health||50.3 (0.2)||50.4 (0.3)||49.1 (0.4)||48.7 (0.4)||50.1 (0.3)||53.2 (0.2)||<0.001†|
|mental health||49.0 (0.2)||46.8 (0.4)||49.6 (0.4)||50.5 (0.4)||52.7 (0.3)||45.3 (0.2)||<0.001†|
|IPSS||4.8 (0.1)||4.2 (0.2)||5.7 (0.2)||5.0 (0.2)||4.8 (0.2)||4.2 (0.2)||<0.001†|
|BMI, kg/m2||26.1 (0.1)||26.0 (0.2)||26.2 (0.2)||28.7 (0.2)||25.9 (0.1)||23.2 (0.1)||<0.001†|
There was evidence of interactions on the SF-12 physical health component score between ‘currently working’ and ‘city’ (Table 3). Adjusting for this interaction, there was no evidence of any effect of age or any interaction between age and BMI, age and city, IPSS and city, or IPSS and age. As Table 3 suggests, the SF-12 physical health component score was reduced if the man reported having diabetes (2.0-point reduction), high blood pressure (2.0), stroke (6.1), heart attack (5.9) and cancer (2.4; all P < 0.001). A 10-point increase in the IPSS was associated with a 3.3-point reduction in the SF-12 physical health component score. A very high or a very low BMI were both associated with a lower SF-12 physical health component scores, with a 1.9 and 3.3-point reduction, respectively. The physical health component of QoL was higher among those currently working, among those men with a tertiary education, and among those who never smoked.
|High blood pressure||−2.0 (0.3)||<0.001|
|Heart attack||−5.9 (0.8)||<0.001|
|> Secondary education||1.6 (0.3)||<0.001|
|former||− 0.6 (0.3)|
|current||− 1.2 (0.3)|
|IPSS (per 10 units)||− 3.3 (0.3)||<0.001|
|Currently employed, by city||<0.001|
The coefficients for the regression model for SF-12 mental health component score are shown in Table 4. There were significant interactions between city and age, and between city and IPSS (both P < 0.001). There were also differences in mean QoL across the cities. For a man aged 50 years with a score of zero on the IPSS, the mean mental component of quality of life was higher in Boxmeer and Boston, followed by Birmingham, than Auxerre and Seoul. The increase in the SF-12 mental health component score with increasing age was much lower in Seoul than in the other four cities. In all five cities, there is evidence that the mental health component score decreased as the IPSS increased. The association is weakest in Seoul, where a 10-point increase in IPSS is associated with a 1.4-point reduction in SF-12 mental health component score, perhaps because the baseline mental health component score is lowest there. Boston and Auxerre have the greatest effects, with a 3.9-point reduction in Boston and a 3.6-point reduction in Auxerre. By contrast, only heart attack had a significant (P < 0.05) effect on the SF-12 mental component (a reduction of 1.8 points). Currently working and currently living with a partner were both associated with an increase in the SF-12 mental health component of QoL.
|High blood pressure||–||–|
|Heart attack||−1.8 (0.7)||0.012||–|
|Married or living with a partner||1.8 (0.5)||<0.001|
|Currently employed||2.9 (0.4)||<0.001|
|Age – 50, by city (per decade)||<0.001|
|IPSS (per 10 units), by city||<0.001|
The present study reports the epidemiology of LUTS, confirms the impact of LUTS on QoL and its constancy across cultures, and shows the confounding effects of comorbidities and demographics.
This study shows that, consistent with other studies, moderate-to-severe LUTS (IPSS of ≥8) affects ≈ 20% of the population of men aged >40 years and that severe LUTS (IPSS of ≥20) affects ≈ 3.5% of them . The differences among the cities are statistically significant, in view of the large sample sizes, but they are relatively small. In broad terms, moderate-to-severe LUTS has a similar prevalence to that of high blood pressure, whereas the prevalence of severe LUTS is about the same level as stroke (2.2%), cancer (4.5%), or heart attack (4.5%).
There is significant variation in the self-reported prevalence of the comorbidities over the cities, with men from Boston reporting the greatest prevalence for virtually all of the comorbidities. The difference in comorbidity prevalence is greater than the differences in LUTS. Many of the comorbidities are positively associated with BMI, and this is much higher among men in Boston than in the other four cities, especially Seoul, where there was a very low incidence of a BMI of >30 kg/m2.
The present study shows that the association between LUTS and the SF-12 physical health component score is the same in all five cities. Also, the effects of the comorbidities are the same despite the prevalence varying across cities. Thus the effect of a heart attack on QoL is the same in Seoul as in Auxerre. The effect of an increase in LUTS is also the same in Seoul as in Auxerre. A 10-point increase in the IPSS is associated with a 3.3-point reduction in the SF-12 physical health component score. Cancer, diabetes or high blood pressure have a reduction in SF-12 physical of about 2 points each; stroke or heart attack have a 5-point reduction. An 8-point increase in the IPSS will move a man from the middle of the moderate symptom range (8–19) into the severe range (20 +), or can take a man from having no symptoms to having moderate symptoms. Thus moderate LUTS has much the same impact on QoL in all five cities as the potentially life-threatening conditions of cancer, diabetes or high blood pressure. Going from no LUTS to severe LUTS is similar to the effect of a stroke or a heart attack on the SF-12 physical health component score of the QoL.
The comorbidities have no significant impact on the SF-12 mental health component score of QoL (other than a heart attack, which reduced it by 1.8 points). A 10-point increase in the IPSS was associated with a 3.4-point reduction in the four western cities and a 1.4-point reduction in Seoul, showing the influence of cultural differences in the effect of LUTS. This is a quantitative interaction, displaying the same general trend but with a lesser effect in Seoul, as noticed in a previous analysis . A similar pattern among Asian populations relative to European and USA populations has also been reported , where the psychological well-being gradient in Japan was not as steep as in Stirling and Olmsted county, but the ‘activity interference’ gradient was similar in Europe, the USA and Japan.
The decrease in QoL with increasing LUTS was studied previously in general populations [1,6], and in patients , but not always in conjunction with non-urological comorbidities. We found the same effect in the present study, and the magnitudes of the effect were similar. However, those were all cross-sectional studies, and a longitudinal study  showed that changes in LUTS symptoms have no great influence on changes in general QoL, while another  showed that changes in symptoms were correlated with changes in LUTS-related bother. It was suggested  that the differences between the cross-sectional studies and the longitudinal studies might be due to the presence of comorbidities, which also have an adverse effect on QoL. However, our analyses adjusted for these, and we found an effect of LUTS on QoL over and above those of the comorbid illnesses.
Previous studies that considered the effect of comorbidities and LUTS on QoL reported that the effect of LUTS was similar to that of cardiac symptoms , that age and concomitant diseases explained about half of the association between LUTS and a reduced QoL; this implies that LUTS had an independent effect on QoL over and above the effect of potentially more serious comorbidities . In another study , severe LUTS had a greater impact on the role physical, role emotional, vitality and mental health scales of the SF-36 than gout, angina, hypertension and diabetes.
In summary, the present study shows that LUTS occurs in ≈ 20% of men aged >40 years, that an increase in LUTS is associated with significant decreases in both the mental and physical health components of QoL, and that these effects are consistent over several different cultural settings. These are important findings; although not life-threatening, the effect of LUTS on the daily living of men aged 40–79 years is as influential as some of the more obvious, potentially life-threatening concomitant diseases. Furthermore, these affect the mental component of QoL whereas the concomitant diseases generally do not. As the number of men aged 40–79 years continues to increase, even what used to be thought of as ‘mild’ disease states, such as LUTS, will become increasingly important. Only with fully realising the impact of LUTS on QoL, and the impact of treatment on QoL [27,28], will we be able to make informed choices about appropriate priorities for potentially limited healthcare resources. While these patients might be seen by several types of practitioners, it is likely that urologists will be in the best position to recognize the true impact of LUTS on a patient’s QoL, and to ensure that colleagues in other disciplines also recognize the importance of these symptoms.
Source of Funding: UREPIK – Unrestricted Research Grant from GlaxoSmithKline. BACH – NIDDK (NIH) funding source (U01 DK 56842). The UREPIK Study Group is Peter Boyle, Lyon; Chris Robertson, Glasgow; Chiara Mazzetta, Cambridge; Martin Keech, Greenford; Richard Hobbs, Birmingham; Richard Fourcade, Auxerre; Lambertus Kiemeney, Nijmegen; Chongwook Lee, Seoul
CONFLICT OF INTEREST