Can Simple Tests Performed in the Primary Care Setting Provide Accurate and Efficient Diagnosis of Benign Prostatic Hyperplasia? Rationale and Design of the Diagnosis Improvement in Primary Care Trial

J. Carballido; R. Fourcade; A. Pagliarulo; C. Cricelli; F. Brenes; A. Pedromingo-Marino; R. Castro


Int J Clin Pract. 2009;63(8):1192-1197. 

In This Article

Rationale for the Diagnosis Improvement in Primary Care Trial (D-IMPACT)

Accurate diagnosis of BPH is essential before associated LUTS can be effectively treated and quality of life improved. Although the progressive nature of BPH provides clinicians with an opportunity to initiate early and effective preventative therapy or allocate more suitable treatment, this approach to the therapeutic management of BPH is also dependent on the availability of accurate and efficient diagnostic strategies. Expected increases in prevalence accompanying the gradual aging of the population further highlight the need for effective BPH diagnosis.

The increasing prevalence of BPH, coupled with greater use of medical therapy, mean that GPs are likely to have an increasingly important role in managing the condition. While diagnosis of BPH in the primary care setting is often based on assessment of symptoms, a recent study has indicated that a standardised assessment (e.g. use of a validated tool such as the IPSS questionnaire) is performed in less than half of men who present to their general practitioner (GP) with LUTS.[17] This is despite the fact that worsening of symptoms is the most common disease progression event in BPH,[18] and that European guidelines recommend use of the IPSS to measure symptoms and bother in order to monitor disease progression and allocate treatment.[9] In addition, there appears to be insufficient focus on risk factors for disease progression such as PSA level and prostate volume; only two-thirds of patients in another study reported having their PSA level measured.[19] These same studies also reported marked national differences in the assessment of patients with BPH and LUTS. For example, PSA measurement was performed in less than 50% of patients in Germany and over 80% of patients in Spain.[19] Furthermore, nearly 40% of Italian patients did not receive a DRE, despite the fact that current EAU and Italian guidelines consider this examination a minimal requirement. It is against this background of suboptimal and conflicting diagnostic practices that the implementation of simple, non-invasive tests for accurate diagnosis of BPH in the primary care setting becomes particularly attractive.

The validity of BPH diagnostic tools that are easily employed in primary care was assessed in a pilot study performed in 20 specialist urology clinics across Spain.[20] This study recruited 363 patients with suspected BPH who were referred to the participating urology clinics because of the presence of LUTS. Patients were excluded if they had previously undergone BPH diagnostic tests; other exclusion criteria included a history of prostate cancer, neurological disorders, pelvic surgery or trauma, previous invasive treatment for urinary flow obstruction, poorly controlled diabetes mellitus and/or diabetic neuropathy, the presence of sexually transmitted diseases, or previous treatment with alpha-blockers or 5α-reductase inhibitors (5ARIs). Patients were initially evaluated by employing the following simple tests in sequence: medical history and physical examination, IPSS questionnaire, DRE and serum PSA measurement. After each of these tests, the participating urologists provided a preliminary diagnosis regarding the presence or absence of BPH. The study demonstrated a high correlation between BPH diagnoses provided following completion of the four simple tests (medical history, IPSS, DRE and serum PSA), and those based on a full battery of tests including ultrasonography and uroflowmetry (Table 1). These findings show that, in specialist urology clinics, valid and accurate diagnosis of BPH can be achieved using a simple diagnostic algorithm involving the sequential collection of information via medical history, IPSS questionnaire, DRE and PSA measurement. The suitability of similar, simple diagnostic strategies applied in the primary care setting will be investigated in the Diagnosis Improvement in Primary Care Trial (D-IMPACT) in three European countries.

Study Design

D-IMPACT is an ongoing prospective, multicentre, epidemiological study that aims to identify the optimal subset of simple tests applied by GPs in the primary care setting to diagnose BPH in men who spontaneously report obstructive (voiding) and/or irritative (storage) LUTS. The validity of the GP diagnosis will be assessed following comparison with a gold-standard diagnosis provided by specialist urologists. The prevalence of BPH in men attending primary care clinics and spontaneously reporting LUTS will also be evaluated.

The study investigators - GPs and urologists from France, Italy and Spain - will assess men aged ≥ 50 years who attend the primary care clinic and spontaneously report obstructive (voiding) and/or irritative (storage) LUTS. The key inclusion and exclusion criteria for patients participating in D-IMPACT are listed in Table 2. Patients with endocrine, neurological, inflammatory or infectious disease for which BPH tests have already been performed will not be eligible. Previous treatment for BPH, or any prior treatment with alpha-blockers, 5ARIs or phytotherapy, will also preclude patient enrolment.

The study design and procedures to be performed at each study visit are summarised in Figure 1. When a patient attends a participating primary care clinic and spontaneously reports LUTS, the GP will assess the eligibility of the patient for inclusion in the study. Eligible patients who have provided informed consent will then undergo an assessment of medical history and LUTS. Immediately after the evaluation of medical history, GPs will provide their opinion about the possible diagnosis of BPH (both on a subjective probability scale [0-100%] and in a yes/no format). In addition, the severity of LUTS, and the impact of these symptoms on patient quality of life, will be assessed during the first study visit via completion of IPSS and bother score questionnaires, respectively. PSA analysis and urinalysis will also be requested at this visit.

Figure 1.

Diagnosis Improvement in Primary Care Trial (D-IMPACT) study design and procedures.

Patients will return to the primary care clinic for the second study visit once the results of the PSA analysis and urinalysis are available. A DRE will be performed at this visit. GPs will then be asked again to provide their opinion about the possible diagnosis of BPH, both in terms of probability (0-100%) and in a yes/no format. This latest diagnosis will be based on all evaluations performed in the primary care setting, i.e. medical history, initial assessment of symptoms, IPSS and bother score questionnaires, PSA analysis, urinalysis and DRE.

The third and final study visit will be performed following referral of the patient to a participating specialist urology clinic and will take place within 3 months of the second study visit. At the final visit, a urologist will provide the final, gold-standard diagnosis. This diagnosis will be formed following a review of the evaluations performed by the GP, and after the urologist has performed a physical examination (including DRE) and additional BPH diagnostic tests. The additional tests (abdominal ultrasound and uroflowmetry) will allow measurement of postvoid residual volume, prostate size and peak urinary flow rate.

Two primary objectives have been defined for D-IMPACT. The first of these is to identify the optimal subset of simple tests employed by GPs for the diagnosis of BPH and evaluate, in a multivariate manner, its diagnostic performance. This will be achieved by comparison with gold-standard diagnoses provided by urologists using the same tests plus the additional investigations. The second primary objective is to assess the prevalence of BPH in a population of men presenting to GPs and spontaneously reporting LUTS, based on the final diagnosis provided by urologists.

The following secondary study objectives have been defined: (i) to describe the sociodemographic and clinical characteristics of a population of men with suspected BPH attending primary care clinics; (ii) to compare the quality of life between patients with a confirmed and non-confirmed BPH diagnosis by the urologist; (iii) to compare the prevalence of risk factors for BPH progression (age, PSA > 1.5 ng/ml and symptoms assessed with the IPSS questionnaire) between patients with and without suspicion of BPH; and (iv) to describe, in an exploratory manner, differences between countries regarding the diagnostic performance of BPH tests and the prevalence of BPH.

An exploratory assessment will also be performed of the diagnostic performance of a simpler model comprising only those tests that are non-subjective and non-invasive, and quick and easy to perform in a GP's office as part of everyday clinical practice. This simpler model will therefore include age, PSA level, IPSS and urinalysis.

Sample size calculations for the second primary objective of the study, assessment of BPH prevalence in men presenting with LUTS, dictate that 760 patients should be recruited into the study. The per protocol population will be the primary population for analysis and will include all patients meeting the study eligibility criteria and undergoing gold-standard diagnosis by the participating urologists. The diagnostic performance of the individual tests performed by GPs will be expressed as the positive predictive value (PPV) and the negative predictive value (NPV); these values represent the proportion of patients with a positive and negative diagnosis, respectively, who are correctly diagnosed. Diagnostic performance will also be expressed as the test sensitivity and specificity, with sensitivity representing the proportion of patients with BPH (according to the gold-standard diagnosis) who test positive, and specificity representing the proportion of patients without BPH who test negative. Using a multivariate logistic regression model, the diagnostic contribution of each test (with the exception of DRE) will be independently established by means of its adjusted odds ratio; as DRE is a subjective evaluation and its interpretation is likely to be influenced by knowledge of other test results, its diagnostic contribution will not be independently established. Upon completion of the study, it will be possible to identify the most efficient combination of tests to maximise diagnosis of BPH in the primary care setting.