The Epidemiology of Acute Urinary Retention in Benign Prostatic Hyperplasia

Abstract

One of the most important events in the natural history of BPH is acute urinary retention (AUR). Better estimates of AUR incidence are now available from both population-based studies and placebo control groups (patients diagnosed with BPH). Several strong risk factors for AUR have been identified by analytical epidemiology, the most important being serum PSA levels. When counseling patients with BPH who are considering watchful waiting, clinicians should also take into account prostate volume, maximum flow rate, and symptom severity.

For a variety of reasons, acute urinary retention (AUR) is one of the most significant complications of long-term benign prostatic hyperplasia (BPH). In the past it has represented an immediate indication for surgery. Between 25% and 30% of men who underwent transurethral prostatectomy (TURP) had AUR as their main indication in older series,¹ and today most patients failing to void after attempted catheter removal still undergo surgery. For this reason alone, AUR is an important and feared event, from an economic standpoint as well as from the viewpoint of the patient. The patient originally has inability to urinate, with increasing pain, and eventually a visit to the emergency room, catheterization, follow-up visits to the physicians, an attempt at catheter removal, and eventual recovery or surgery, which is both painful and time consuming. In the older literature, the risk of recurrent AUR was cited as 56% to 64% within 1 week of the first episode and 76% to 83% in men with diagnosed BPH.^2–4

Etiology of AUR

The etiology of AUR is poorly understood. Prostate infection, bladder overdistension,⁵ excessive fluid intake, alcohol consumption, sexual activity, debility, and bed rest have all been mentioned.⁶

Prostatic infarction has been suggested as an underlying etiologic event.⁷ Spiro and colleagues⁸ found evidence for infarction in 85% of prostates removed for AUR versus 3% in prostates of men having surgery for symptoms only. In contrast, there was no evidence of infarction in six prostatectomy specimens removed from men who had surgery for AUR.⁹ Anjum and colleagues¹⁰ found fundamentally similar rates of infarction in 35 men who were in AUR versus no AUR (1.9% versus 3.0%).

From a clinical and prognostic point of view, spontaneous AUR should be separated from precipitated AUR, although this is by no means consistently done in the literature. Precipitated AUR refers to the inability to urinate following a triggering event such as non-prostate-related surgery, catheterization, anesthesia, ingestion of medications with sympathomimetic or anticholinergic effects, ingestion of antihistamines, or other events. All other AUR episodes are classified as spontaneous.^11,12 The importance of differentiating the two types of AUR becomes clear when evaluating ultimate patient outcomes. Following spontaneous AUR, 15% of patients had another episode of spontaneous AUR, and 75% underwent surgery; after precipitated AUR, only 9% had an episode of spontaneous AUR, and 26% underwent surgery.¹¹

Epidemiology of AUR in BPH

Definition. The study of epidemiology involves the distribution and determinants of diseases in humans. Epidemiology can be descriptive (description of disease incidence, mortality, and prevalence by person, place, and time) or analytical (a search for determinants of disease risk that may serve to increase prospects for prevention).¹³ Epidemiologists assess and compare rates of diseases within one population stratified by sex, age, and other demographic and socioeconomic parameters, and between populations of different culture, ethnicity, lifestyles, and diet.

Methods of studying the natural history of BPH. AUR is an event that takes place in the course of the natural history of the disease process. The natural history of BPH can theoretically be evaluated by studies using a variety of designs (see box).

Study Designs for Evaluation of BPH

1. Longitudinal studies of untreated cohorts of men diagnosed with lower urinary tract symptoms (LUTS) and clinical BPH by any definition (watchful waiting cohorts).

2. Studies of the behavior of men diagnosed with LUTS and BPH and enrolled in controlled studies of LUTS and BPH (control groups) and receiving either

1. no treatment (compared with active intervention)

2. placebo treatment (compared with medical treatment)

3. sham treatment (compared with device or surgical treatment)

4. Longitudinal studies of unselected (ie, undiagnosed) men living in the community who are less likely to progress and request or require therapy (longitudinal population-based studies).

There are problems associated with all these approaches. Concerning the watchful waiting cohorts, the first question to be resolved is whether or not it is ethical (or feasible) to enroll symptomatic men in such a study even if the disease studied is not fatal. Second, the very fact that the patients had an initial contact (and presumably subsequent contacts) with health care providers in the course of the study will bias them, leading to changes in outcome parameters of interest presumably different from those observed in an age-matched cohort of men who are similar in all parameters at baseline but who choose to participate (a cohort one might call “wild-type,” in analogy to genetic language). Furthermore, in the course of such a natural history study, many diagnosed men will become more and more symptomatic and will desire and receive treatment, making them ineligible for further study participation and thus reducing the number of men available for analyses.

Concerning the incidence of AUR, data are available from population-based studies as well as from placebo control groups of BPH studies. These data can be discussed in terms of descriptive and analytical epidemiologic approaches.

Descriptive epidemiology. Older estimates of occurrence of AUR range from 4 to 15 to as high as 130 per 1000 person-years (calculated by Jacobsen and colleagues¹⁴ based on studies by Birkhoff and colleagues,¹⁵ Ball and colleagues,¹⁶ and Craigen and colleagues¹⁷); such estimates lead to 10-year cumulative incidence rates ranging from 4% to 73%. The self-reported rate of AUR in a crosssectional study in 2002 Spanish men was 5.1%.¹⁸

More recent data from carefully controlled studies in better defined populations shed additional light on the incidence rates in community dwelling men and clinical BPH populations (Table 1). AUR occurred in the VA Cooperative Study over 3 years in 1 man after TURP and in 8 of 276 men in the watchful waiting arm, for an incidence rate of 9.6/1,000 person-years.¹⁹ Barry and colleagues²⁰ reported outcomes of 500 men diagnosed by urologists with BPH, who were candidates for prostatectomy by established criteria but elected to be followed conservatively. In 1574 person-years, 40 episodes of AUR occurred at a constant rate throughout the 4 years of follow-up, for an incidence rate of 25/1000 person-years.

Table 1.

Descriptive Studies on the Incidence of Acute Urinary Retention (AUR)

Author/Source	Description	No. of	Cohort	Years of	% Overall	% Year	IR/1000	95% CI
	of Cohort	Cases		Follow-up			patient years
Ball et al,	Watchful	2	107	5	1.9	0.37	3.7
198116	waiting study
Craigen et al,	Watchful						15.0
196917	waiting study
Birkhoff et al,	Watchful	10	26	3	39	13	130
197615	waiting study
Wasson et al,	TURP vs. watchful	8	276	3	2.8	0.9	9.6
199519	waiting VA COOP
Hunter et al,	Self-reported prior	102	2002	?	5.1		50.9
199718	events in Spanish men
Barry et al,	Prostatectomy	40	500	4	8	2.5	25
199720	candidates
Meigs et al,	Physicians Health	82	6100	3	1.3		4.5	3.1–6.2
199921	Study, self-reported
Olmsted	Community cohort	57	2115	4			6.8	5.2–8.9
County,	40–49 years old
199722
McConnell et al,	Placebo group	99	1376	4	7.2	1.8	18
199828	of Pless Study
Andersen et al,	Placebo groups of	57	2109	2	2.7	1.35	13.5
199724	2-year BPH studies

During 15,851 person-years of follow-up in the Physicians Health Study, 82 men reported an episode of AUR, for an incidence rate of 4.5/1000 person-years (95% CI 3.1–6.2).²¹ Of the 2115 men aged 40 to 79 years in the Olmsted County Study, 57 had a first episode of acute urinary retention during 8344 person-years of follow-up (incidence 6.8/1000 person-years, 95% CI 5.2–8.9).²²

The best data from men diagnosed with BPH stem from the Proscar Long Term Efficacy and Safety Study (PLESS).²³ In PLESS, 1376 placebo-treated men with enlarged prostates and moderate symptoms had complete follow-up over 4 years; 99 of them experienced an episode of AUR, for a calculated incidence rate of 18/1000 person-years. The placebo treatment groups from three 2-year studies with a similar patient population were meta-analyzed by Boyle and colleagues.²⁴ Of 2109 patients, 57 experienced AUR over the 2 years with a constant hazard, for an incidence rate of 14/1000 person-years (Table 1).

Analytical epidemiology. Several well-controlled studies have provided considerable insight into the risk factors for AUR. In the Physicians Health Study, rates increased with age and baseline symptom severity retention.²¹ In men with mild symptoms, AUR incidence increased from 0.4/1000 person-years for those 45 to 49 years old to 7.9/1000 person-years for those 70 to 83 years old. In men with symptom scores of 8 to 35, rates increased from 3.3/1000 person-years for those 45 to 49 years old to 11.3/1000 person-years for those 70 to 83 years old. Men with a clinical diagnosis of BPH and a symptom score of 8 or greater had the highest rates (age-adjusted incidence 13.7/1000 person-years). All seven lower urinary tract symptoms comprising the American Urological Association symptom index individually predicted AUR. The sensation of incomplete bladder emptying, having to void again after less than 2 hours, and a weak urinary stream were the best independent symptom predictors. Use of medications with adrenergic or anticholinergic side effects also predicted acute urinary retention.

In the Olmsted County Study, analyses focused on age, symptom severity, maximum flow rate, and prostate volume²² (Figures 1 and 2). Incidence rates per 1000 person-years increased from 2.6 for men in their 40s to 9.3 for men in their their 70s if they had mild symptoms, and from 3.0 to 34.7, respectively, if they had more than mild symptoms (Figure 1). The relative risk increased for older men, those with moderate to severe symptoms (3.2 times), those with a flow rate under 12 mL/sec (3.9 times), and those with a prostate volume > 30 mL by TRUS (3.0 times), all compared with a baseline risk of 1.0 times for the corresponding groups (Figure 2). The highest relative risk by proportional hazard models exists for 60- to 69-year-old men with more than mild symptoms and a flow rate < 12 mL/sec (10.3 times), and for 70- to 79-year-old men except if they had mild symptoms and a flow rate > 12 mL/sec. All other stratifications of men over 70 years had a relative risk (RR) ranging from 12.9 to 14.8 times (all compared with men 40 to 49 years old with mild symptoms and a flow rate > 12 mL/sec, for which the base risk is 1.0 times).

Figure 1.

Incidence rates of AUR in Olmsted County Study by age and symptom severity. Data from Jacobsen et al.²²

Figure 2.

Relative risk of AUR in Olmsted County Study by age, symptom severity, peak flow rate, and prostate volume; the red columns represent the baseline and a relative risk (RR) of 1.0; the vertical lines represent the 95% CI. Data from Jacobsen et al.²²

Although age in community dwelling men is an important risk factor, in a BPH trial population of men already diagnosed with BPH, other factors can be analyzed. In the placebo groups of three 2-year studies²⁵ and a 4-year study (PLESS),^12,26–28 prostate volume, serum prostate-specific antigen (PSA) levels, and symptom severity were all predictors of AUR episodes.

The incidence increased from 5.6% to 7.7% in men with a serum PSA of under 1.4 ng/mL and mild to severe symptoms, and from 7.8% to 10.2% for those with a serum PSA > 1.4 ng/mL over 4 years in PLESS.²⁶ In the 2-year studies, the rate of AUR was eightfold higher in those with a serum PSA > 1.4 ng/mL (0.4% versus 3.9%), and threefold higher for a prostate volume > 40 mL (1.6% versus 4.2%).²⁵ A detailed analysis showed a near linear increase in risk for AUR with increasing thresholds of serum PSA (Figure 3) in PLESS, an observation that applied to both spontaneous and precipitated AUR.¹² The risk for both types of AUR increases with increasing serum PSA as well as prostate volume stratified by tertiles (Figure 4). An analysis of over 100 possible outcome predictors alone or in combination revealed that a combination of serum PSA, urination more than every 2 hours, symptom problem index, maximum urinary flow rate, and hesitancy were only slightly superior to PSA alone in predicting AUR episodes.²⁹

Figure 3.

Incidence of spontaneous or precipitated AUR in the PLESS study over 4 years stratified by increasing thresholds of serum prostate-specific antigen (PSA) at baseline.

Figure 4.

Spontaneous, precipitated, or combined AUR incidence over 4 years in the PLESS study stratified by tertiles of serum PSA or prostate volume at baseline.

Conclusions

AUR is one of the most significant events in the course of the natural history of BPH. The concept that this disease is in fact progressive in nature is slowly being accepted. Although in the past uncertain estimates regarding the incidence rates of AUR existed, and almost one-third of patients presented in AUR for TURP, nowadays better estimates are available from population-based studies of community dwelling men, as well as from patients diagnosed with BPH (placebo control groups). Descriptive and analytical epidemiologic data have shown that the incidence rate per 1000 person-years is less variable in the community than previously assumed. The best controlled studies allow for an estimate between 5 and 25 per 1000 person-years, or 0.5% to 2.5% per year. However, this risk is cumulative and increases with advancing age. Even at these seemingly low rates, the cumulative risk for a man in his 50s with more than mild symptoms of experiencing AUR if he lives to be 80 is about 20%; for a man in his 60s who lives another 20 years it is about 23%; and for a man in his 70s who lives another 10 years it is 30%. The risk in patients diagnosed with BPH is naturally higher, and analytical epidemiology has identified several strong risk factors for AUR, the most important being elevated serum PSA levels. In addition, prostate volume, maximum flow rate, and symptom severity should be considered when counseling patients presenting with LUTS and clinical BPH who are considering a course of watchful waiting.

Main Points.

• Acute urinary retention (AUR) is one of the most important complications of benign prostatic hyperplasia (BPH).

• For evaluating outcomes, it is important to distinguish between precipitating AUR (which follows a triggering event) and spontaneous AUR.

• BPH may be evaluated by controlled studies or by longitudinal (either watchful waiting or population-based) studies.

• Recent data from controlled studies have given better estimates of AUR incidence rates: 5 to 25 per 1,000 person-years; the risk is cumulative and increases with age.

• Analytical epidemiology studies show that incomplete bladder emptying, having to void again after less than 2 hours, and a weak urinary stream are the best risk predictors for AUR.

• In the important PLESS study, predictors of AUR were prostate volume, serum prostate-specific antigen (PSA) levels, and symptom severity.