SEARCH

SEARCH BY CITATION

Keywords:

  • Prostate cancer;
  • familial;
  • prognosis;
  • detection bias

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Objectives

To compare traditional prognostic characteristics of familial vs sporadic prostate cancers and to investigate potential detection biases arising from differences in the use of screening and investigative procedures.

Patients and methods

Familial and sporadic cancers were identified in a population-based sample of incident prostate cancers (total 318) in Auckland, New Zealand. To examine the potential for detection biases in these comparisons, the sociodemographic and clinical characteristics were determined according to family history status for a sample of 959 patients newly referred to Auckland urology clinics by general practitioners for the investigation of prostate-related conditions.

Results

Compared with sporadic prostate cancers, familial cancers were more likely to be diagnosed in patients at a younger age (P=0.05), after asymptomatic serum prostate-specific antigen (PSA) screening (P=0.02), and to include a lower proportion with extraprostatic disease (P=0.009) and serum PSA levels before diagnosis of >20 ng/mL (P=0.04). This was consistent with the observed trend for patients referred to urology clinics with a positive family history to be of higher socio-economic and educational status and to more frequently undergo screening and biopsy investigation.

Conclusion

Familial prostate cancers appeared to be diagnosed at an earlier stage of disease progression in this study population, possibly as the result of the higher socio-economic status and greater use of screening and investigative procedures amongst patients reporting a positive family history. These features reduce the validity of cross-sectional comparisons of prognostic variables for familial vs sporadic prostate cancer and emphasize the need for further longitudinal prognostic studies.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Previous studies have reported two- to three-fold increases in prostate cancer risk amongst men who have a positive family history involving one first-degree relative diagnosed with this disease, and the risk increases with younger age of onset and the number of relatives affected [ 1]. Clinicians are encouraged to routinely inquire about familial factors in assessing patients’ prostate cancer risk, but the prognostic significance of familial prostate cancer has not been clearly defined. A longitudinal investigation of biochemical relapse after radical prostatectomy has suggested that familial prostate cancers may follow a more aggressive clinical course than sporadic cancers [ 2], but cross-sectional studies based mainly on patients undergoing radical prostatectomy have reported few differences in traditionally used prognostic clinicopathological characteristics of familial compared with sporadic prostate cancers [ 2[3][4][5][6][7]–8]. Evidence suggests that the incidence of prostate cancer, particularly early-stage tumours, is influenced by the level of screening and clinical investigation [ 9, 10], but previous studies of prognostic variables in familial and sporadic prostate cancers have not considered the potential role of detection biases arising from more intensive investigation of men reporting positive family histories.

The aims of the current study were two-fold: first, to undertake a population-based cross-sectional comparison of the clinicopathological prognostic variables of familial and sporadic prostate cancers, in advanced as well as clinically localized cases; and second, to systematically describe the sociodemographic and clinical characteristics of a population-based sample of men undergoing urological investigation for prostate cancer, according to their reported family history status. This second aim was to investigate the potential role of detection biases in these analyses.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Recruitment of prostate cancer cases

The Auckland Prostate Study aimed to identify all men aged 40–80 years with newly diagnosed prostate cancer in the greater metropolitan area of Auckland, New Zealand over the 13-month study period from March 1996. As almost all men in Auckland in this age group with a new diagnosis of prostate cancer attend urologists, urology clinic attendance was used as the basis to ascertain prostate cancer cases. All public hospital urology clinic attendees and patients attending five of the seven private clinic urologists in Auckland were eligible for participation (an estimated 91% of all patients in the study population). The study identified prostate cancer cases from two recruitment groups of men attending the urology clinics. Most cases were identified by systematic follow-up of the clinic records of patients who had been recruited prospectively during February 1996 to February 1997, at the time of their first GP-referred urology clinic attendance for investigation of prostate-related conditions (prostatic symptoms, raised serum PSA level, or suspicious DRE). Remaining prostate cancers diagnosed at the participating urology clinics over the study period were identified retrospectively from laboratory histology records, biopsy clinic and theatre lists. These cases included: re-attendance at clinics for follow-up appointments, referrals from other than GPs (from wards and other specialist clinics), and diagnoses after TURP (where original assessment and assignment to a surgical waiting list pre-dated the study recruitment period). Men with a previous diagnosis of prostate cancer were excluded and all prostate cancer cases were confirmed histologically.

Data collection and analysis

Self-administered questionnaires were used to collect personal, sociodemographic, anthropometric and lifestyle data. The questionnaire was used to determine whether first-degree relatives had been diagnosed with cancer (including a ‘don’t know’ option), and for each cancer, the site and age at which it had been diagnosed. Questionnaires were completed immediately before urology clinic attendance by the prospectively identified cases and within 3 weeks of diagnosis by those identified retrospectively. Clinical data were collected from referral letters, clinic medical records and pathology reports. Combined Gleason scores and pathological staging data were identified from prostatic needle biopsy, radical prostatectomy and TURP histology reports, and seminal vesical and lymphoid tissue histology reports, where available. Positive bone scintigraphy reports were also used to determine extraprostatic disease status. Positive predictive values (PPVs) for prostate cancer amongst the initial GP referrals were defined as the proportion of these men diagnosed with prostate cancer following initial investigations, according to their reported family history status. Sociodemographic, clinical and pathological characteristics according to family history status were compared statistically using the chi-square or t-test, with significance defined at P<0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The study recruited 318 prostate cancer cases diagnosed at Auckland urology clinics during the 13-month period, including 191 cases recruited prospectively (77% response rate) and 127 retrospectively (72% response rate). The prospectively recruited cases were identified from amongst 959 patients attending urology clinics after GP referrals for the investigation of prostate-related conditions. In all, 208 prostate cancers cases were diagnosed in this group after initial investigations and surgery carried out during the extended period from March 1996 to June 1997.

The clinicopathological features of the 318 prostate cancer cases diagnosed in Auckland over the study period are presented in Table 1. Familial prostate cancer was identified in 26 cases reporting at least one affected first-degree relative (including six cases with two or more affected relatives). The mean age at diagnosis was lower for familial than for sporadic cancers (P=0.05). The diagnosis of familial prostate cancer was more likely than sporadic cancer to have followed asymptomatic serum PSA screening before the urology clinic attendance (P=0.02), and to be treated with radical prostatectomy (P=0.06). Familial and sporadic cancers had a similar prevalence of coexistent high-grade prostatic intraepithelial neoplasia and similar combined Gleason scores. However, there was a smaller proportion of familial cancers with prediagnostic serum PSA levels of >20 ng/mL (P=0.04) and with pathological or radiological evidence of extraprostatic disease (P=0.009).

Table 1.   (A) Clinical and pathological characteristics of 318 prostate cancers diagnosed in Auckland from March 1996 to March 1997, and (B) sociodemographic and clinical characteristics of 959 men referred to Auckland urology clinics for the investigation of prostate-related conditions, from February 1996 to February 1997, by family history of prostate cancer Thumbnail image of

Table 1 also presents a comparison of the sociodemographic and clinical characteristics of GP-referred patients attending urology clinics for investigation of prostate-related conditions, by family history status. There was little difference in mean age between the groups (data not presented). Patients with unknown family histories of cancer were characterized by lower socio-economic status and educational attainment than those reporting a negative family history, whilst those reporting a positive family history had higher socio-economic status and education levels. Although not reaching statistical significance at the usually accepted levels, men reporting a positive family history were more likely to be referred after asymptomatic serum PSA screening by GPs (P=0.27) and to undergo prostatic biopsy (P=0.16). The PPV for a positive family history of prostate cancer was 37.5% overall, compared with 23% for men with a negative family history, and this increased with prediagnostic serum PSA levels.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this population-based sample of prostate cancers, familial cancers were more likely to be diagnosed at a younger age after asymptomatic screening and there was a lower proportion of familial cancers in men with with extraprostatic disease and prediagnostic serum PSA levels of >20 ng/mL. The investigation of new patient referrals by GPs to urology clinics (the major source of prostate cancer cases) suggests that men reporting a positive family history of prostate cancer were more likely to be of higher socio-economic and educational status and to undergo asymptomatic screening and diagnostic biopsy.

The clinicopathological characteristics of familial compared with sporadic cancers has been reported in several clinic-based cross-sectional studies of men undergoing radical prostatectomy [ 2, 5[6][7]–8] and a few population-based studies [ 3], including a USA-based study of men referred for investigation after prostate cancer screening [ 4] which was similar to the present, but involved a group of higher-risk patients. These studies have reported either no differences in the clinicopathological characteristics of familial vs sporadic cancers or weak trends in stage, grade or prediagnostic serum PSA levels which were similar to the present findings.

These findings are inconsistent with the observations of a poorer disease-free survival of men with familial cancers after radical prostatectomy [ 2], although this has not been supported by a more recent smaller study [ 8] and must be regarded as preliminary at this stage. However, we propose three possible explanations for this apparent inconsistency between prognostic data from longitudinal and cross-sectional studies. First, most cross-sectional studies have been restricted to clinically localized tumours in men undergoing radical prostatectomies. In the present study the most significant differences between familial and sporadic prostate cancers were for more advanced stage and grade tumours which would usually be excluded from radical prostatectomy series. Second, the influence of a subgroup of biologically more aggressive hereditary prostate cancers may not have been delineated in previous cross-sectional studies. Hereditary prostate cancer is likely to involve inheritance of highly penetrant genes [ 11] and initial gene-mapping studies have identified susceptibility loci on chromosome 1 (HPC1) [ 12] and the X chromosome (HPCX) [ 13]. A cross-sectional study of hereditary prostate cancer in families potentially linked to the HPC1 locus concluded that there was a greater proportion of higher grade and advanced stage tumours compared with potentially unlinked families [ 14]. That study had the additional advantage of including a wider range of cancer stages and treatment modalities. Third, cross-sectional comparisons of familial and sporadic cancer stage and grade are potentially misleading if there is a tendency for men reporting positive family histories to be investigated more intensively, possibly leading to diagnosis at an earlier point in the natural history of the disease. In the present study, men with familial prostate cancers were more likely to have been initially referred after asymptomatic screening, and urology clinic attendees with a positive family history were more likely to have undergone serum PSA screening and prostatic biopsy. This trend persisted when the comparison was restricted to a subgroup of 571 attendees with serum PSA levels of >4 ng/mL or a suspicious DRE, despite the lower mean serum PSA levels on presentation amongst men in this group with positive family histories (data for this subgroup not presented). In addition, there was a trend amongst the urology clinic attendees and a population-based sample of men (from a former case-control study control group, data not presented) for men reporting a positive family history to be of higher socio-economic and educational status, features which are also compatible with greater use of screening and investigative procedures. We are unaware of biologically plausible explanations for the association observed between positive family histories of prostate cancer and higher socio-economic and educational status. As there is no reason to believe that differential selection of recruited participants occurred on the basis of family histories of prostate cancer in the present study, we suggest that this finding is likely to represent differing patterns of family history recall according to socio-economic and educational status. We are not aware of previous reports of socio-economic factors or patterns of prediagnostic clinical activities in relation to familial and sporadic variants of prostate cancer. In the future, a more valid interpretation of prognostic markers from cross-sectional studies may be derived from comparison groups (familial vs sporadic cancer) with similar socio-economic, screening and investigational characteristics.

The present study has several possible limitations. As staging in the study was limited to pathological and radiological (bone scintigraphy) findings, it is probable that the extent of extraprostatic disease was underestimated in men not undergoing radical prostatectomy. Under-staging of cancer is less likely to have involved patients with a positive family history as they were younger and more likely to undergo radical surgical treatment. Bias arising from this source is therefore expected to have lead to an underestimation of the true differences in pathological stage between familial and sporadic cancers.

A strength of the study is the systematic and comprehensive collection of family history data, but in common with most previous studies of familial prostate cancer, we were unable to validate self-reported family histories against other sources. The present study provides indirect evidence for concern about the validity of self-reported family history status. We identified and excluded from the analyses a significant proportion of participants (including several with clinically significant cancers) whose family histories were reported as unknown. Most previous studies have not specified how such participants were classified in relation to analyses. A comparison of self-reported and database-linked histories has reported a 70% sensitivity for reported prostate cancer in relatives [ 15]. Reduced specificity of reported family histories of prostate cancer is also a concern as there is potential for commonly occurring benign prostatic conditions in family members to be confused with prostate cancer. However, a previous study which carried out a limited validation of probands’ recalled family histories against telephone interview or the medical records of 32 affected siblings, reported only one brother with BPH mis-reported as cancer [ 3].

In conclusion, we suggest that patients referred to urology clinics who report positive family histories of prostate cancer comprise a higher socio-economic and educational status group in whom prostate cancer was diagnosed at an earlier stage of disease progression, possibly as the result of greater use of investigative clinical practices. These findings raise questions about the validity of cross-sectional comparisons of prognostic variables in familial and sporadic prostate cancers, arising from detection bias, and emphasize the need for further longitudinal studies to clarify the prognostic significance of familial variants of this disease.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was supported by the Cancer Society of New Zealand and the Health Research Council of New Zealand. The assistance of the following urologists with the recruitment of participants is acknowledged: Mr John Boulton, Mr Jon Cadwallader, Mr Roger Chambers, Mr Russell McIlroy, Mr Derek Rothwell and Mr Michael Rice. Special acknowledgement is made of statistical advice provided by Dr Roger Marshall and technical assistance given by Ms Cherie Mulholland, Ms Susan Hawkins and Mrs Fleur O’Keefe.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Walsh PC & Partin AW Family history facilitates the early diagnosis of prostate carcinoma. Cancer 1997;80 1871 4
  • 2
    Kupelian PA, Klein EA, Witte JS et al. Familial prostate cancer: a different disease? J Urol 1997;158 2197 201
  • 3
    Steinberg GD, Carter BS, Beaty TH et al. Family history and risk of prostate cancer. The Prostate 1990;17 337 47
  • 4
    Aprikian AG, Bazinet N, Plante M et al. Family history and the risk of prostatic carcinoma in a high risk group of urological patients. J Urol 1995;154 404 6
  • 5
    Carter BS, Bova GS, Beaty TH et al. Hereditary prostate cancer: epidemiological and clinical features. J Urol 1993;150 799 802
  • 6
    Bastacky SI, Wojno KJ, Walsh PC et al. Pathological features of hereditary prostate cancer. J Urol 1995;153 987 92
  • 7
    Keetch DW, Humphry PA, Smith DS et al. Clinical and pathological features of hereditary prostate cancer. J Urol 1996;155 1841 43
  • 8
    Bauer JJ, Srivastava S, Connelly RR et al. Significance of familial history of prostate cancer to traditional prognostic variables, genetic biomarkers, and recurrence after radical prostatectomy. Urology 1998;51 970 6
  • 9
    Newcomer LM, Stanford JL, Blumenstein BA et al. Temporal trends in rates of prostate cancer: declining incidence of advanced stage disease 1974–94. J Urol 1997;158 1427 30
  • 10
    Threlfall TJ, English DR, Rouse IL Prostate cancer in Western Australia: trends in incidence and mortality from 1985 to 1996. Med J Aust 1995;169 21 4
  • 11
    Carter BS, Beaty TH, Steinberg GD et al. Mendelian inheritance of familial prostate cancer. Proc Natl Acad Sci USA 1992;89 3367 71
  • 12
    Smith JR, Freije D, Carpten JD et al. A genome-wide search reveals a major susceptibility locus for prostate cancer on chromosome 1. Science 1996;274 1371 3
  • 13
    Xu J, Myers D, Freije D et al. Evidence for a prostate cancer susceptibility locus on the X chromosome. Nature Genetics 1998;20 175 9
  • 14
    Gronberg H, Isaacs SD, Smith JR et al. Characteristics of prostate cancer in families potentially linked to the hereditary prostate cancer 1 (HPC1) locus. JAMA 1997;278 1251 5
  • 15
    Kerber RA & Slattery ML Comparison of self-reported and database-linked family history of prostate cancer data in a case-control study. Am J Epidemiol 1997;146 244 8
  • 16
    Johnston RA revision of economic indices for New Zealand. Wellington: New Zealand Council for Educational Research 1983