SEARCH

SEARCH BY CITATION

Introduction

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

There is evidence of variation in the management of men with insidious changes in urinary frequency, flow, dribbling, urgency and hesitancy [1–6]. Given this confusion among GPs and urologists alike, several countries have initiated the development of clinical practice guidelines. For example, the USA Agency for Health Care Policy and Research (AHCPR) and the WHO published guidelines in 1994 [7,8]. Guidelines were published soon after in the UK [9] and by Australia's peak medical authority, the National Health and Medical Research Council (NHMRC) [10]. By then, previously accepted dogma of a causal association between the presence of such symptoms and prostatic enlargement (so-called ‘BPH’) had largely been rejected. International authorities instead referred to a symptom complex known as ‘LUTS’[11]. The relationship between BPH and LUTS is not straightforward. The presence of an enlarged prostate is not predictive of LUTS, nor is the presence of LUTS always accompanied by the presence of an enlarged prostate [11].

Given the controversy continuing to surround the value of prostate cancer screening [12], one of the more problematic aspects of the management of LUTS arises from the relationship of these symptoms to the presence of early prostate cancer. For example, the Australian guidelines [10] state ‘It is not recommended to estimate serum PSA as part of the normal evaluation of a man with LUTS’ (page xiii). After excluding clinical signs suspicious of malignancy such as haematuria, GPs and urologists are advised in these guidelines that the presence of LUTS is insufficient reason alone to screen for early prostate cancer, consistent with national policy recommending against screening of asymptomatic men [13].

However, a survey of randomly selected Australian GPs one year after publication of the NHMRC guidelines suggested resistance to this recommendation [14]. Specifically, 75% of GPs agreed that patients with urinary symptoms expect to be tested for prostate cancer, significantly more than the proportion agreeing that asymptomatic men expect to be tested (39%). More than a third (37%) agreed there was an established link between urinary symptoms and early prostate cancer [14]. In another study in which randomly selected GPs reported their typical management of a man with bothersome LUTS, 66% of clinicians agreed they would screen using PSA [15]. As in the other study, an identical proportion (37%) perceived an established relationship between LUTS and early prostate cancer [15]. In 1999, half of randomly selected GPs participating in a local trial to implement preventive care guidelines supported PSA screening of men with LUTS (Young & Ward, unpublished data).

Community surveys also show that the presence of LUTS independently predicts PSA testing [16–18]. While some PSA tests appear to be ordered in response to patient demand [19], there is evidence that GPs recommend the test solely because patients have LUTS and no other accepted indication [20].

To date, this dilemma has not benefited from a rigorously conducted systematic review of the evidence for the relationship of LUTS and early prostate cancer. As health authorities endorse evidence-based policies about prostate cancer screening [21], any clinical recommendation that PSA testing be included in the diagnostic evaluation of men with LUTS must similarly be based on evidence of increased risk. In response, we developed a replicable protocol with which to review the literature, to answer the following perplexing question: Is the risk of early prostate cancer greater in men with LUTS than in asymptomatic men?

Method

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

Search strategy for identification of studies

At the end of 1998, studies were identified reporting rates of prostate cancer in men with BPH or LUTS, by using a sensitive search strategy combining terms for prostate cancer with terms for urinary symptoms. Because of resource constraints, the search was restricted to literature in English. Medline (1966 onwards) and Embase (1988 onwards) databases were searched using the terms given in Appendix 1 (Silver Platter format). To ensure that the latest unpublished findings were not missed, abstracts from the AUA Conferences for 1997 and 1998 were searched manually. Research presented in earlier conferences was assumed to have already appeared as articles in peer-reviewed journals included in the search.

Abstracts of retrieved articles were screened to identify those pertinent to the research question. When an abstract was not available electronically (e.g. letters to the Editor), the full article was obtained unless the title clearly indicated that the article was irrelevant. Ten exclusion criteria were used to exclude studies that would not potentially provide epidemiological data about rates of prostate cancer (Table 1). If the potential usefulness of an article was unclear from the abstract, the full article was reviewed. In this way, a pool of potentially relevant articles was identified. The reference lists of these articles were reviewed and further potentially relevant articles retrieved.

Table 1.  Primary reason for exclusion of 1553 articles identified during Medline search following review of abstract, and the primary reason for excluding 168 potentially relevant articles
Reason for exclusion Number of excluded articles (n = 1553)
Epidemiological study, rates of prostate cancer not reported66
Immunochemical or biochemical studies714
Histological studies201
Studies reporting the management of disease398
Studies of radiology, not reporting rates of prostate cancer46
Studies of cancers other than prostate cancer31
No empirical data reported65
Individual case reports8
Conference abstracts prior to 19972
Animal studies22
Relevant articles
Rate of prostate cancer not reported:
for men with either BPH or LUTS91
for men without BPH or LUTS (i.e. no control group)41
No primary data30
Article not available within 3 months of interlibrary loan request6

All retrieved articles then were further screened to identify those reporting rates of prostate cancer in men with either BPH or LUTS. Further additional exclusion criteria were applied (Table 1). As LUTS [22,23] and the risk of developing or dying from prostate cancer [24] are strongly associated with increasing age, studies that did not address potential confounding by age through age-standardized analyses or other procedures were then excluded.

Data abstraction and synthesis

For each included study, a critical appraisal form with 28 items (Appendix 2) was completed independently (by J.Y. and D.M.) to assess study type and abstract data. Inter-rater reliability was determined by calculating the kappa statistic [25]. Any disagreements in study categorization and data abstraction then were resolved subsequently by consensus. Statistical pooling of results was intended but not undertaken because of the marked variation in study design and quality.

Results

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

Of 1664 articles identified by the Medline search, 1532 were excluded in the first screening round after a review of the published abstract (Table 1). Of 179 potentially relevant citations (Table 2), 168 subsequently were excluded after a review of the full articles (162) or because the full article was not available within 3 months of the search (six; Table 1). Of the remaining 11 studies, six were then excluded from the review specifically because confounding by age had not been addressed (Table 3) [26–31]. Of the five remaining studies eligible to be included in the review (Table 4), one reported the results of two studies [32]. Inter-rater reliability between J.Y. and D.M. for the included articles was  > 0.8 (‘very good’) for 16 items on the critical appraisal form; 0.6–0.8 for six items (‘good’) and < 0.6 for six items (‘moderate’ or ‘fair’). Complete reference lists of excluded studies are available from the authors on request.

Table 2.  Sources of 179 potentially relevant articles
Source Number of hits Number of potentially relevant articles (= 179)
Medline1664111
Embase1026 27 (excluding duplicates)
Reference reviewN/A 40 (excluding duplicates)
AUA 1998 Conference Abstracts1500  0
AUA 1997 Conference Abstracts1809  1
Table 3.  Details for six studies excluded because of confounding by age
Ref Study type Population Number of men Rate of prostate cancer in men with Comments
LUTS or BPH in controls
  1. RCT, randomized control trial; CS, cross-sectional; PC, prostate cancer.

[26]Cohort29 133 male smokers participating in RCT of nutritional supplements in Finland317 with PC 29 133 without PC  8.8% of men with PC had BPH at baseline 3.9% of men without PC had BPH at baseline
[27]CSVolunteers to ACS-NPCDP2425, 1218 with urinary symptoms 1207 with no symptoms2.4%1.9%
[28]CohortVolunteers to ACS-NPCDP1383 screened annually for 5 years2.4%1.7%
[29]CSOne medical centre in USA362 with abnormal prostate cancer screening test16%5%
[30]CSEuropean prostate cancer screening trial8603 volunteers for screening  Overall, AUA7 score not associated with PC. For PSA > 10, high AUA7 score was significantly associated with lower rate of PC
[31]CS13 urology clinics1703 men referred to urology clinic  AUA7 symptom score discriminated poorly between BPH and PC
Table 4.  Details of six studies reported in the five included articles
Ref Population Number of men [study factor] Rate of PC in men RR or OR
with LUTS or BPH with no LUTs or BPH
  1. PC, prostate cancer; PY, person-years.

Cohort studies in men with BPH
[32] Inpatients at one hospital in NY 338 with BPH [BPH] 338 age-matched controlsAge-adjusted mortality rate from PC 10.1 per 1000 PY2.7 per 1000 PY3.74
Cohort studies in men after TURP
[33] One hospital in US 838 after TURP [BPH] 802 age-matched controls2.9% (3.0/1000 PY)3.2% (3.1/1000 PY)0.88 (0.50–1.55)
[34] One hospital in Sweden 198 after TURP [BPH] 203 age-matched controls6.6% 3.4%
Case control studies
[32] One hospital in NY 209 cases [BPH] 209 age-matched controls admitted with benign condition  5.1 (Cases had 5 × odds of a previous admission for prostatic disease)
[35] Two clinics in USA 43 432 health plan members undergoing a periodic health examination 238 cases with prostate cancer 238 age-matched controls from cohort (incomplete data for some pairs)[Urinary symptoms (including haematuria)] [Past diagnosis of BPH]  Cases had 5 × odds of previous BPH symptoms OR 1.2 (0.8–1.9) for past history of BPH present for geqslant R: gt-or-equal, slanted 2 years before diagnosis of PC
Cross-sectional analysis
[36] Six medical centres in USA 1167 with an abnormal screening test who, underwent biopsy from a study base of 6630 volunteers for PC screening[Urinary symptoms (including haematuria and haematospermia)]  OR 0.70 (0.51–0.96), adjusted for age, PSA, DRE and TRUS

Of the six included reports, three were cohort studies, two were retrospective case-control series and one a cross-sectional analysis. The cohort studies comprised hospital patients with and without a primary diagnosis of BPH or after prostatectomy

Primary diagnosis of BPH

Armenian et al.[32] identified 345 men who had been admitted to the Roswell Park Memorial Institute (RPMI) in Buffalo, New York, between 1945 and 1965, and discharged with a primary diagnosis of BPH. From information also obtained from the medical record, seven cases were excluded because of malignancy or prostatitis. The authors did not state the criteria used by clinicians to make the primary diagnosis. The proportion of the resulting 338 cases with BPH also undergoing prostatectomy during this admission was not specified. Presenting complaints and the presence of LUTS or other more sinister symptoms (e.g. haematuria) were not reported for cases. For each man with BPH, a male control of the same age (within one year) was selected at random from patients with a non-neoplastic condition discharged from the RPMI in the same year. However, all the patients with BPH and 96% of controls had a DRE to evaluate prostate size. The 36 controls with an enlarged prostate were replaced. The presence of BPH or LUTS among controls was neither established nor reported.

As all study subjects were residents of New York State, death certificates and cancer registrations at the New York State Health Department and RPMI records for cases and controls were searched until 1972 for mortality outcomes. At the final follow-up, survival data were available for 87.6% of patients with BPH and 88.5% of controls. The authors acknowledged that latent prostate cancer was more likely to be diagnosed among men with BPH because there was a greater likelihood of prostatic surgery in this group. Therefore, all cases of latent prostate cancer (three in the BPH group and one in the control) were excluded from their analysis. The proportions of men in either the case or control group who had undergone a prostatectomy after the index admission was not reported. Furthermore, the proportion of control patients who subsequently developed BPH was also not reported. Although autopsy reports confirmed the primary cause of death for one in five control patients, rates of coexisting prostate cancer were not reported.

This study found that the age-adjusted mortality rate for prostate cancer was 3.7 times greater among cases than controls (10.1 vs 2.7 per 1000 person years). Further analysis found that the mortality rate from prostate cancer was 5.1 per 1000 person years among 77 patients with BPH who had undergone a prostatectomy during the index hospital admission, 11.3 per 1000 person years among the 229 patients with BPH who did not undergo prostatectomy during the index admission and 2.7 per 1000 person years among the controls. The authors provided no confidence intervals for their estimates.

This study concluded that men admitted with BPH were significantly more likely to die from prostate cancer than men who did not have an enlarged prostate when hospitalized for other reasons. Most at risk were men with BPH who did not undergo prostatectomy during their index admission.

After prostatectomy

From the medical records of patients admitted to the Lahey Clinic Foundation in Boston (Massachusetts, USA), Greenwald et al.[33] identified 838 aged < 80 years who had undergone a subtotal prostatectomy for histologically confirmed BPH between 1942 and 1955. The presence of LUTS among cases was not reported. For each of these patients, a control patient in the same 5-year age group who had undergone surgery other than prostatectomy in the same month and who did not have known prostate disease or malignancy was selected from the hospital admissions index. Neither the criteria to detect BPH nor the presence of LUTS were reported for the control group. A matching control was not available for 36 (4.9%) cases. Patients undergoing prostatectomy (88%) and controls (91%) were followed up until death or the end of the study in 1970. The presence of prostate cancer was determined by medical record review, pathology reports, correspondence from physicians, death certificates and autopsy reports.

During the follow-up, 24 of 828 (2.9%) prostatectomy patients developed prostate cancer, compared with 26 of 802 controls (3.1%; relative risk, RR, 0.88, 95% CI 0.50–1.55). However, among the control group, 78 (9.7%) developed BPH during the follow-up period. The rate of prostate cancer was 3.8% among controls who developed BPH, 3.2% among controls who did not develop BPH and 2.9% among prostatectomy patients. These rates were not significantly different from each other or from that expected using age-specific population rates from local cancer registry data, suggesting that men undergoing a prostatectomy for BPH were no more likely to develop prostate cancer than men undergoing other surgery.

In another study, Hammarsten et al.[34] sought to test the hypothesis that men with latent prostate cancer who undergo subtotal prostatectomy may be at increased risk of developing clinical prostate cancer because of tumour promotion resulting from this surgery. Their study-base comprised 198 men aged <80 years who had undergone TURP at Varberg Hospital, Sweden between 1980 and 1984. The presence of LUTS before TURP was not reported for these cases. Men with clinical prostate cancer at the time of surgery were excluded. However, unlike the study by Greenwald et al.[33], men with evidence of early (stage T1) prostate cancer were included. These men were age-matched with 203 controls who had undergone an inguinal hernia repair at the same hospital during the same period. Neither the proportion of LUTS nor BPH in the control group was reported, but 26 control patients underwent TURP during the follow-up period.

All patients were followed up for 10 years using hospital and general practice records, autopsy reports, death certificates and cancer registry data. Of the 198 patients in the TURP group, six developed clinical prostate cancer (3.0%) compared with five of 203 (2.5%) in the control group. During the follow-up, three men in the TURP cohort and three controls died from prostate cancer. When men with early (stage T1) prostate cancer were included, 13 of 198 (6.6%) in the TURP cohort developed prostate cancer, compared with seven of 203 (3.4%) in the control group. However, as 177 men in this latter group did not undergo TURP, early prostate cancer would not have been diagnosed. The authors concluded that TURP for clinically benign prostate disease does not increase the risk of developing clinical prostate cancer in the ensuing 10 years.

Retrospective case-control studies

The review identified two case-control studies in which men with prostate cancer were matched with men without prostate cancer and their histories of BPH or LUTS compared. Using a hospital-based case-control design, Armenian et al.[32] identified 290 men admitted to the RPMI between 1957 and 1965 with a final diagnosis of prostate cancer. These cases were age-matched (within one year) with a control patient with a benign final diagnosis. Information about previous admissions for prostatic disease was obtained from cases and controls through questionnaires and interviews with patients. This information was coded for reason for admission and ‘index of suspicion of cancer’ by coders unaware of the individual's diagnosis. Using a matched-pair analysis, the RR of a preceding admission for ‘possible BPH’ was reported as 5.1 for cases with prostate cancer compared with controls. ‘Possible BPH’ was not defined. Confidence intervals around the estimates of RR were not reported.

Hiatt et al.[35] conducted a case-control study nested within a cohort of 43 432 men in the Kaiser Permanente Health Plan in California and who had undergone a health examination between 1979 and 1985. Men with incident prostate cancer (n = 238) between 1979 and 1985 identified from cancer registry data were matched with a control subject from the cohort for age (within one year), race and date of health examination (within one year). Patient questionnaires completed during health examinations and medical record audit conducted unaware of the man's case or control status provided information about ‘BPH symptoms’, history of genitourinary infections and other conditions including family history of cancer. ‘BPH symptoms’ were defined as frequency, hesitancy, dribbling, incontinence and weak stream, but also haematuria. Initial analysis found that men with prostate cancer were significantly more likely to ever have had BPH symptoms. However, when the analysis was restricted to a past diagnosis of BPH at least 2 years before the diagnosis of prostate cancer, the association was not significant (RR 1.2, 95% CI 0.8–1.9).

Cross-sectional analysis

Catalona et al.[36] reported a cross-sectional analysis of data from a multicentre trial of prostate cancer screening. Advertisements were placed in the lay media, encouraging volunteers aged geqslant R: gt-or-equal, slanted50 years to attend one of six medical centres for a measurement of PSA and a DRE by a urologist or medical oncologist. Of the 6630 men self-selecting to participate, 1710 had either an abnormal DRE or a serum PSA level of > 4 ng/mL and were recommended for quadrant TRUS-guided needle biopsy of the prostate. Of 1167 who underwent biopsy, 264 were found to have prostate cancer. Logistic regression analysis found that absence of symptoms (frequency, urgency, weak urinary stream, dysuria, haematuria or haematospermia) was predictive of the presence of prostate cancer among men with an abnormal finding. The adjusted odds ratio (OR) for the presence of prostate cancer in symptomatic men was 0.70 (95% CI 0.51–0.96), i.e. significantly lower than the risk for asymptomatic men. However, men with a normal DRE and PSA level did not undergo biopsy, so the presence of early prostate cancer in this group is unknown.

Discussion

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

As the management of men with LUTS consumes substantial health resources [7,9,10], it is crucial that an evidence-based approach be adopted in response to controversies and clinical uncertainties. Unfortunately, the present effort to systematically review the international literature found an impoverished evidence base from which to inform clinical decision-making. Of the entire international database searched, only 11 articles promised data of any relevance to the clinical question posed at the outset of the review. Six of these failed to account for age as a confounder, rendering their findings invalid.

Methodological critique

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

Only two articles included in this review assessed urinary symptoms [35,36]; the remaining three, reporting four studies, focused on the presence of BPH [32–34] which was inconsistently defined, and seldom with explicit criteria sufficient to permit replication. As there is no clear relationship between LUTS and prostate size, it would not have been unreasonable to have rejected these studies outright in this review. While persevering with their inclusion, profound methodological weaknesses compromise their value. Specifically, in selecting study subjects from a single hospital, the cohort study of Armenian et al.[32] was probably unrepresentative of men with BPH in the community. Men with BPH who did not require hospitalization, e.g. those with milder symptoms, would not have been included in that study.

Second, there was potential for misclassification of both BPH and prostate cancer in that study [37]. Criteria to establish a diagnosis of BPH were not provided by the authors. Although all cases had a DRE, histological confirmation of BPH was not reported. For each study subject, the presence of prostate cancer was determined from death certificates and cancer registry reports. However, death certificates may not be accurate for ascertaining prostate cancer [37,38]. Autopsy or other pathology reports were available for only about a third of either BPH cases (33.7%) or controls (30.2%).

Third, no outcome data were available for 12.4% of BPH cases and 11.5% of controls. If rates of prostate cancer were systematically different among men lost to follow-up compared with those for whom data were available, the study results would be subject to selection bias.

The case-control analysis reported by Armenian et al.[32] was also subject to bias. As the study base was poorly defined, there was potential for selection bias. As before, cases were men with prostate cancer admitted to a single institution. Controls selected from among hospital inpatients may not have represented the same study base as the cases. Recall bias cannot be excluded as it was unclear whether men with prostate cancer were unaware of their diagnosis when completing questionnaires about previous admissions and urinary symptoms. Thus, the positive association between BPH and prostate cancer shown by Armenian et al. in both cohort and case-control studies may be subject to bias.

Two studies found no increase in rates of prostate cancer among men who had undergone subtotal prostatectomy for BPH [33,34]. Both studies involved patients admitted to single institutions rather than drawn from a defined population. Neither reported the presence of urinary symptoms. Methodological weaknesses of the study of Greenwald et al.[33] included loss to follow-up (12% of cases, 9% of controls) and failure to provide data for the histological confirmation of prostate cancer.

The strengths of the study by Hammarsten et al.[34] include a complete follow-up of all study subjects and the use of cancer registry data in addition to medical records, death certificates and autopsy reports to determine the development of prostate cancer. The authors concluded that neither BPH nor TURP was a risk factor for the development of prostate cancer. However, 13% of men in the control group subsequently underwent TURP at an unspecified time during the 10 years of follow-up. These men were analysed as controls, potentially biasing the study findings towards the null hypothesis.

Studies assessing the relationship between specific LUTS and prostate cancer provide some conflicting results. Nested within a large cohort, cases and controls in the case-control study by Hiatt et al.[35] were selected from the same study base. Another strength was the ascertainment of urinary symptoms from health-plan records unaware of the subject's case or control status. This study failed to find a significant association between prostate cancer and either BPH or specific urinary symptoms (including haematuria) present for > 2 years before the diagnosis of cancer. Conversely, analysis of these data to include symptoms present within 2 years of cancer diagnosis found RRs of 4–5 for the development of malignancy. However, the impact of including haematuria in this latter analysis was unclear.

The only attempt at a community-based study recruited volunteers to a multicentre trial of prostate cancer screening [36]. The authors reported that urinary symptoms were associated with a reduced risk of cancer among those men with an abnormal screening test. However, this study also has several weaknesses. As men with negative screening tests were not biopsied, the association between symptoms and latent prostate cancer in this group could not be assessed. Overall, 32% of men with a positive PSA test or DRE had no biopsy despite a recommendation to do so. If the association between symptoms and prostate cancer was systematically different for these men, bias may have been introduced. Despite the inclusion of symptoms inconsistent with a diagnosis of LUTS [10], i.e. haematuria, haematospermia and bone pain, no association with prostate cancer was found.

Thus, the present literature review found no compelling evidence of an association between LUTS and prostate cancer. Two studies (from the same institution) found a significant association between BPH and prostate cancer [32], whereas two failed to find an increased rate of subsequent malignancy among men after prostatectomy for BPH [33,34]. Of the two studies assessing urinary symptoms, one found an association between recent onset symptoms and cancer, but not for symptoms present for > 2 years [35]. The remaining study found that absence of urinary symptoms predicted prostate cancer. All studies had methodological weaknesses which are likely to have biased their results.

Implications for clinical practice

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

We acknowledge that there are two types of response to the absence of evidence, i.e. ‘grey zones’[39]. Specifically, an interventionist approach assumes patients are best served by proceeding with an investigation, procedure or treatment, even if the empirical evidence of its value is weak, while a non-interventionist approach accepts that the potential for harm always outweighs the likely benefit until there is a solid and uncontroversial body of accumulated evidence [39]. The value of prostate cancer screening is as yet speculative [10]. Hence, as the evidence to date of an association between LUTS and early prostate cancer also is weak, PSA testing of men with LUTS is similarly a ‘grey zone’.

The present authors find it inexplicable as to why such an important clinical problem has fostered so little rigorous research. It also is inexplicable that most studies were hospital-based, recruiting cases and controls or cohorts from a biased study base. Men with LUTS and their GPs require evidence from community-based research. Only one study involved a community study base. Other studies are potentially ‘fatally flawed’ in their generalizability because of selection biases caused by access, health insurance, sampling and matching. If another criterion had been imposed, i.e. that of a community study base, the present systematic review would have been limited to one study. If we had further imposed a reasonable requirement for representative community sampling, even this study would have been rejected.

The paucity of rigorously acquired evidence with which to determine the relationship between LUTS and early prostate cancer is cause for concern. In its absence, advice to men about whether or not to be tested for early prostate cancer as part of managing their LUTS will probably be left to individual clinical discretion. While there is considerable promotion of the need for ‘informed consent’ when asymptomatic, far less concern has been aired about PSA testing in the context of LUTS. However, from the present review of the evidence we believe it is reasonable to advise men with LUTS that there are no data to suggest that they have an increased risk of prostate cancer. In the absence of compelling evidence to the contrary, PSA testing in the context of LUTS management is likely to do greater harm than good.

Conclusion

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

The Australian guidelines stated ‘Men with uncomplicated LUTS should be advised that there are no data to suggest that they have an increased risk of prostate cancer’[10]. Given conflicting results from poor-quality studies, we see no reason to revise this guideline. Developers of guidelines in other countries should also find the results of the present systematic review useful for decision-making. Furthermore, this review also highlights the potential deficiencies of an investigator-driven research agenda. Compared with the weight of total citations about prostate disease, studies addressing this specific clinical question are few. With some foresight, the Australian guidelines recommended epidemiological research ‘to determine the association between urinary symptoms and early prostate cancer’[10]. We endorse a multidisciplinary research collaboration between epidemiologists and urologists to design and conduct a rigorous and definitive study to answer this important clinical conundrum. In the absence of such research, men will continue to be ill-advised about the relationship between LUTS and early prostate cancer.

Acknowledgements

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

We thank Mr Geof Hirst for content expertise and Mr Michael Otim and Mrs Mary Bourke for diligently retrieving and photocopying articles for this review.

References

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices
  • 1
    Ward J & Sladden M. Urinary symptoms among older men, their investigators and treatment: is there an epidemic of undetected morbidity in the waiting room? Fam Pract 1994; 11: 2519
  • 2
    McNaughton Collins M, Barry M, Bin L et al. Diagnosis and treatment of benign prostatic hyperplasia: practice patterns of primary care physicians. J Gen Int Med 1997; 12: 2249
  • 3
    Hansen M & Zdanowski A. The urological management of men with lower urinary tract symptoms, recorded using an interactive computer program. Br J Urol 1997; 80: 20510
  • 4
    Hansen M & Zdanowski A. The agreement among urological experts on the diagnostic management of patients with common urological problems. Br J Urol 1997; 80: 78792
  • 5
    Gewalli F & Hansen M. variations in the diagnostic process among urologists in one region recorded using an interactive computer program. Scan J Urol Nephrol 1998; 32: 3641
  • 6
    Barry M, Fowler F, Bin L, Oesterling J. A nationwide survey of practising urologists: current management of benign prostatic hyperplasia and clinically localised prostate cancer. J Urol 1997; 158: 48892
  • 7
    McConnell J, Barry M, Bruskewitz R et al. Benign prostatic hyperplasia: diagnosis and treatment. Clinical practice guideline No 8: Agency for Health Care Policy and Research, Public Health Service. Rockville, Maryland, US Department of Health and Human Services. AHCPR Publication No 94–0582, 1994
  • 8
    Cockett A, Aso Y, Chatelain C et al. eds. Proceedings of the 2nd International Consultation on Benign Prostatic Hyperplasia. Channel Islands: Scientific Communication International Ltd, WHO, 1994
  • 9
    NHS Centre for Reviews and Dissemination. Benign prostatic hyperplasia. Effective Health Care 1995; 2 (December 1–16)
  • 10
    National Health and Medical Research Council. Clinical practice guidelines: The management of uncomplicated lower urinary tract symptoms in men. Canberra: AGPS, 1996
  • 11
    Abrams P. New words for old: lower urinary tract symptoms for ‘prostatism’. Br Med J 1994; 308: 92930
  • 12
    Woolf S. Should we screen for prostate cancer? Br Med J 1997; 314: 98990
  • 13
    Australian Health Technology Advisory Committee. Prostate cancer screening. Canberra: AHTAC, 1996
  • 14
    Ward JE & Girgis S. General practitioners' estimates of men's risks of prostate cancer and expectations of screening. Aust NZ J Pub Hlth 1999; 23: 21920
  • 15
    Puech M, Sladden M, Hirst G, Ward J. Management of uncomplicated lower urinary tract symptoms in men before release of national guidelines: whose practice needs to change? Aust Fam Phys 2000; in press
  • 16
    Ward J, Hughes AM, Hirst G, Winchester L. Men's estimates of prostate cancer risk and self-reported rates of screening. Med J Aust 1997; 167: 2503
  • 17
    Pinnock C, Weller D, Marshall V. Self-reported prevalence of prostate-specific antigen testing in South Australia: a community study. Med J Aust 1998; 169: 258
  • 18
    Perkins J, Sanson-Fisher RW, Clarke SM, Youman P. An exploration of screening practices for prostate cancer and the associated community expenditure. Br J Urol 1998; 82: 5249
  • 19
    Ward J, Gupta L, Taylor N. Do general practitioners use prostate-specific antigen as a screening test for early prostate cancer? Med J Aust 1998; 169: 2931
  • 20
    Slevin T, Donnelly N, English D, Clarkson J, Ward JE. Prostate cancer testing: behaviour, motivation and attitudes among Western Australian men. Med J Aust 1999; 171: 1858
  • 21
    Selley S, Donovan J, Faulkner A, Coast J, Gillat D. Diagnosis, management and screening of early localised prostate cancer: a systematic review. Health Technol Assessment 1997; 1: 2
  • 22
    Garraway W, Russell E, Lee R et al. Impact of previously unrecognised benign prostatic hyperplasia on the daily activities of middle-aged and elderly men. Br J Gen Pract 1993; 43: 31821
  • 23
    Pinnock C & Marshall V. Troublesome urinary symptoms in the community: a prevalence study. Med J Aust 1997; 167: 725
  • 24
    Australian Institute of Health and Welfare (AIHW) and Australasian Association of Cancer Registries (AACR). Cancer in Australia 1991–94 (with Projections To 1999) Canberra: AIHW (Cancer Series No 7), 1998
  • 25
    Altman D. Practical Statistics for Medical Research. London: Chapman & Hall, 1991
  • 26
    Hartman TJ, Albanes D, Rautalahti M et al. Physical activity and prostate cancer in the alpha-tocopherol, beta-carotene (ATBC) cancer prevention study (Finland). Cancer Causes Control 1998; 9: 118
  • 27
    Mettlin C, Lee F, Drago J, Murphy GP et al. The American Cancer Society National Prostate Cancer Detection Project. Findings on the detection of early prostate cancer in 2425 men. Cancer 1991; 67: 294958
  • 28
    Mettlin C, Murphy GP, Babaian RJ et al. The results of five-year early prostate cancer detection intervention. Cancer 1996; 77: 1509
  • 29
    Babaian RJ, von Eschenbach AC, Miyashita H, Ramirez EI, Evans RB. Early detection program for prostate cancer: results and identification of high-risk patient population. Urology 1991; 37: 1937
  • 30
    Rietbergen JB, Kranse R, Boeken Kruger AE, Kirkels WJ, Schröder FH. Additional value of the AUA7 symptom score in prostate cancer (PC) detection. AUA Conference Abstract No 467 1997 :120
  • 31
    Stoevelaar HJ, van der Beek C, Nijs HG, Casparie AF, McDonnell J, Janknegt RA. The symptom questionnaire for benign prostatic hyperplasia: an ambiguous indicator for and ambiguous disease. Br J Urol 1996; 77: 1815
  • 32
    Armenian HK, Lilienfeld AM, Diamond EL, Bross ID. Relation between benign prostatic hyperplasia and cancer of the prostate. Lancet 1974; 2: 1157
  • 33
    Greenwald P, Kirmss V, Polan AK, Dick VS. Cancer of the prostate among men with benign prostatic hyperplasia. J Nat Cancer Inst 1974; 53: 33540
  • 34
    Hammarsten J, Andersson S, Holmén A, Högstedt B, Peeker R. Does transurethral resection of clinically benign prostate gland increase the risk of developing clinical prostate cancer? A 10-year follow-up study. Cancer 1994; 74: 234751
  • 35
    Hiatt RA, Armstrong MA, Klatsky AL, Sidney S. Alcohol consumption, smoking, and other risk factors and prostate cancer in a large health plan cohort in California (United States). Cancer Causes Control 1994; 5: 6672
  • 36
    Catalona WJ, Richie JP, Ahmann FR et al. Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: results of a multicentre clinical trial of 6630 men. J Urol 1994; 151: 128390
  • 37
    Williams RD & Blackard CE. Benign prostatic hyperplasia and cancer of the prostate. Lancet 1974; 23: 1265
  • 38
    Byar DP, Huse R, Bailar JC. An exponential model relating censored survival data and concomitant information for prostatic cancer patients. J Natl Cancer Inst 1974; 52: 3216
  • 39
    Davidoff F. Evangelists and snails redux: the case of cholesterol screening. Ann Int Med 1996; 124: 5134

Authors

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

J.M. Young, MBBS, MPH, Postgraduate Fellow.

D. Muscatello, BSc, MPH, Public Health Officer.

J.E. Ward, PhD, FAFPHM, Director.

Appendices

  1. Top of page
  2. Introduction
  3. Method
  4. Results
  5. Discussion
  6. Methodological critique
  7. Implications for clinical practice
  8. Conclusion
  9. Acknowledgements
  10. References
  11. Authors
  12. Appendices

Appendix 1

Search terms for the literature search.

#1 explode PROSTATIC NEOPLASMS/all subheadings

#2 (prostat* near cancer) or (prostat* near carcinoma)

#3 prostate specific antigen or PSA

#4 digital rectal examination or DRE

#5 protat* near screening

#6 #1 or #2 or #3 or #4 or #5

#7 lower urinary tract symptoms or LUTS

#8 benign prostatic hyperplasia or BPH

#9 prostatism

#10 nocturia

#11 urin* near urgency

#12 urin* near frequency

#13 urin* near incontinence

#14 urin* near hesitancy

#15 postmicturition dribble

#16 bladder outlet obstruction

#17 AUA*

#18 IPSS*

#19 #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or  #15 or #16 or #17 or #18

#20 English in la

#21 #6 and #19 and #20

Appendix 2