Population-based analysis of prostate-specific antigen (PSA) screening in younger men (<55 years) in Australia

Authors


Abstract

Objective

  • To analyse the trends in opportunistic PSA screening in Australia, focusing on younger men (<55 years of age), to examine the effects of this screening on transrectal ultrasonography (TRUS)-guided biopsy rates and to determine the nature of prostate cancers (PCas) being detected.

Subjects and Methods

  • All men who received an opportunistic screening PSA test and TRUS-guided biopsy between 2001 and 2008 in Australia were analysed using data from the Australian Cancer registry (Australian Institute of Health and Welfare) and Medicare databases. The Victorian cancer registry was used to obtain Gleason scores.
  • Age-standardized and age-specific rates were calculated, along with the incidence of PCa, and correlated with Gleason scores.

Results

  • A total 5 174 031 PSA tests detected 128 167 PCas in the period 2001–2008.
  • During this period, PSA testing increased by 146% (a mean of 4629 tests per 100 000 men annually), with 80 and 59% increases in the rates of TRUS-guided biopsy and incidence of PCa, respectively.
  • The highest increases in PSA screening occurred in men <55 years old and up to 1101 men had to be screened to detect one incident case of PCa (0.01%).
  • Screening resulted in two thirds of men aged <55 years receiving a negative TRUS biopsy.
  • There was no correlation with Gleason >7 tumours in patients aged <55 years.

Conclusion

  • Despite the ongoing controversy about the merits of PCa screening, there was an increase in PSA testing, especially in men <55 years old, leading to a modestly higher incidence of PCa in Australia.
  • Overall, PSA screening was associated with high rates of negative TRUS-biopsy and the detection of low/intermediate grade PCa among younger patients.
Abbreviations
PCa

prostate cancer

USANZ

Urological Society of Australia and New Zealand

AIHW

Australian Institute of Health and Welfare

ERSPC

European Randomized Study of Screening for Prostate Cancer

SEER

Surveillance, Epidemiology, and End Results

Introduction

Australia and New Zealand together have the highest incidence of prostate cancer (PCa) worldwide with ∼17 835 men being diagnosed with PCa annually in Australia [1]. PCa is the third most prevalent cancer in Australia as a result of low PCa-specific mortality rates, with 5-year survival rates > 90% [2]. Since the introduction of PSA screening for PCa in the USA, little survival benefit has been seen with the resulting early diagnosis and treatment [3]; the relative merits of PSA screening and the benefits and harms from diagnosis and treatment remain highly controversial. More recently, the US Preventive Services Task Force recommended against the use of PSA screening in all age groups because of its limited efficacy in reducing PCa-specific mortality, as well as the harms associated with over-diagnosis, prostate biopsy-related complications and over-treatment [4-6].

As the findings of these trials are based on a slightly older population of men (aged ≥55 years), the impact of PSA screening in younger men has limited evidence. In addition, it is not known if PSA screening has resulted in similar increases in the rates of diagnosis of localized PCa in nations with a high incidence of PCa, such as Australia. We aimed to investigate the trends in the use of PSA testing in Australia, focusing on men aged < 55 years, and to identify its effects on the rates of PCa, and then to analyse the implications of PSA testing, especially in younger men.

Subjects and Methods

Study Population

Our target population was all men in Australia undergoing a screening PSA test. According to the Australian Bureau of Statistics, Australia's average population during the study period from 2001–2008 was 20 343 843, of which 10 104 126 were men. Of those men, 2 914 160 were aged >50 years [7, 8].

The study period was selected, starting from the introduction of the dedicated screening Medicare Benefits Schedule (MBS) code 66655 in 2001 and continuing up to the latest available PCa incidence data in 2008.

Guidelines for PSA Screening and TRUS-Guided Biopsies and the Current Practice in Australia

No clear guidelines for PSA screening were available during the period of this study. TRUS-guided biopsies were proposed for abnormalities detected by PSA tests (commonly set at 4 ng/mL) or by DRE [9]; the Urological Society of Australia and New Zealand (USANZ) policy in 2009 recommended starting a single PSA test and DRE at 40 years and a TRUS-guided biopsy for men with a PSA level above the age-specific median level, also taking into account family history, ethnicity, DRE and PSA derivatives such as PSA velocity and free:total PSA ratio [10].

Although individual practices may vary, it is usual for the decision to proceed to a prostate biopsy in Australia to be made on the basis of a PSA reading above the upper limit of the age-adjusted normal range for the patient concerned. This level is quoted on the patients' results by the pathology laboratory undertaking the analysis, based on data for the particular assay being used to determine the PSA level.

Databases

(1) MBS database: rates of PSA testing in the community

The MBS is the Australian national database which records all health-related information and codes all tests and procedures conducted throughout Australia for billing purposes. As there was no method of differentiating between screening and follow-up PSA tests, a new item billing code ‘66655’, defined as a single PSA test performed during a 12-month period, was introduced in 2001 by the MBS as an initial ‘screening’ PSA test and the code was used to monitor its use. The MBS codes for patients who have prostatic disease including follow-up PSA testing for PCa is coded under different billing codes. Thus the billing code 66655 is solely used for ‘men who do not have previous prostate disease’ and negates the possibility of multiple counting of one individual in a year; however, this code does not determine if members of the target population are symptomatic. It is therefore only possible to determine the rates of opportunistic screening; the code does not identify only asymptomatic men who have undergone a screening test.

The annual number of PSA tests conducted throughout Australia between 2001 and 2008 were obtained in 10-year age bands using the MBS item code 66655 described above. Age-specific rates were calculated per 100 000 men and age-standardized rates were calculated using Segi's world population [11]. Similarly, age-specific and age-standardized rates of PSA testing were calculated per 100 000 men for each Australian state.

(2) MBS database: TRUS-guided prostate biopsies performed

To determine any relationship between serum PSA testing and the performance of a TRUS-guided prostate biopsy, the MBS billing code ‘37219’ defined as ‘PROSTATE, needle biopsy of, using prostatic ultrasound techniques and obtaining 1 or more prostatic specimens’ was used. Age-specific and age-standardized rates were calculated as previously described.

(3) The Australian Institute of Health and Welfare Cancer Registry database: incidence of PCa

The Australian Institute of Health and Welfare (AIHW) database is the central cancer registry for all cancer incidence and deaths, and collates data by all state-based cancer registries. AIHW cancer incidence and screening data are validated and reputable [8]. The number of PCas were obtained for each 10-year age band using the AIHW datasets, and age-specific and age-standardized rates were calculated for the period 2001–2008.

Analysis of PCa Incidence with the Gleason Scores

Three Australian state-based cancer registries collect detailed records of PCa grading, but none have validated recording of staging of PCa. Of these, the Victorian cancer registry has a well-established and long-standing database of the Gleason scores of diagnosed prostate tumours [12].With 2 484 490, the second largest population of Australian men (24%) residing in Victoria, the Victorian cancer registry was the most useful database to be representative of Australia; therefore, to investigate the grading of the diagnosed PCas, we examined the age -specific and age-standardized rates of PCa and Gleason scores in Victoria.

Statistical Analyses

The overall percentage increases in PSA, TRUS-guided biopsy and PCa incidence rates were calculated over the 7 years. The trends and the annual percentage change in PSA rates and PCa incidence were calculated using Joinpoint regression analysis software (version 3.5.4) [13]. The mean 7-year ratios of PSA, TRUS-guided biopsy and PCa incidence were determined. Spearman's correlation was used to calculate the correlation coefficient between the rates of PSA, PCa incidence and Gleason scores in Victoria with spss statistical software (version 17.0).

Results

PSA Testing and PCa Diagnosis

A total of 5 174 031 screening tests that were carried out on an average annual population of 10 104 126 men detected 128 167 PCas between 2001 and 2008. There was a 146% increase in the rates of PSA testing during this period, with a mean of 4629 tests performed per 100 000 men. By comparison, there were only 80 and 59% overall increases in the rates of TRUS-guided biopsy and PCa incidence, respectively, during the period of analysis (Fig. 1A). On Joinpoint regression analysis, the annual percentage increases in PSA testing and TRUS-guided biopsy were 11.8% (P < 0.05) and 9.6% (P < 0.05) respectively, while the annual percentage increases in PCa incidence were 10.2% in 2001–2004 and 4.87% in 2004–2007 (P < 0.05 [Table 1]).

Figure 1.

(A) Rates of PSA testing, TRUS-guided biopsy and PCa incidence in the period 2001–2008 in Australia. Rates were calculated per 100 000 men. The incidence of PCa was calculated using Segi's world population. (Note: there is a break in the Y-axis from 200 to 2000 to illustrate the differences between the rates). (B) Crude rates of PSA testing in different states in Australia.

Rates were calculated per 100 000 men. NSW, New South Wales; VIC, Victoria; QLD, Queensland; SA, South Australia; WA, Western Australia; TAS, Tasmania; ACT, Australian Capital Territory; NT, Northern Territory.

Table 1. Rates of PSA testing leading to TRUS-guided biopsy, and PCa incidence rates
 PSA testing rateTRUS-guided biospy rateIncidence of PCaNo. PSA tests per biopsyNo. TRUS-guided biospies per PCa detectedNo. PSA tests per PCa detectedPercentage PSA tests per PCa, %
  1. Rates were calculated per 100 000 men. The ASR incidence of PCa was calculated using Segi's world standardized rate.
Age band, years       
35–442 951.19.03.0329.33.11011.00.01
45–5411 648.5150.570.276.82.2164.90.61
55–6419 576.6697.4402.728.11.748.62.06
65–7421 689.41009.4841.221.51.225.63.91
75–8415 129.4480.3963.931.70.515.66.41
≥858 548.2136.21038.563.30.18.312.05
Age-standardized rate4 628.7132.0102.634.91.344.52.25
Annual percentage change11.89.610.2–4.6    
Overall increase (%)1468059    

Table 1 shows the number of PSA tests conducted per PCa detected by TRUS-guided biopsy and the number of TRUS-guided biopsies used per diagnosis of PCa. For every 34.9 men who had a PSA test performed, only one person underwent a TRUS-guided biopsy (2.9%), indicating that 97.1% of the PSA testing did not result in further investigation for PCa with TRUS-guided biopsy. Similarly only one PCa was detected per 44.5 men who underwent PSA testing (2.2%), showing that 97.8% of PSA tests in Australia did not result in PCa diagnosis. Furthermore, one PCa was detected in every 1.3 men who underwent a TRUS-guided biopsy (Table 1).

Populations Being Tested with Serum PSA

A total of 66% of the overall PSA testing was conducted in men aged < 65 years. The highest rates of PSA use were in the age group 65–74 years (21 689 per 100 000), followed by the age group 55–64 years (19 577 per 100 000), with PCa detection rates of 3.9 and 2.1%, respectively. The highest rates of TRUS-guided biopsy were also seen in these two age groups (1009 and 697 per 100 000, respectively); however, the highest incidence of PCa was seen in the ≥85 years (1038 per 100 000) and 75–84 years (964 per 100 000) age groups (Table 1). The number of PSA tests used to detect one PCa decreased with increasing age. Over the period of analysis, there was a 17.8% decrease in the rates of PCa in men aged > 85 years (Table 1).

Victoria had the fifth highest mean rate of PSA testing (6294 PSA tests per 100 000) and the third highest increase in use of PSA testing (183%), contributing to 24% of the entire rate of PSA testing in Australia and 23% of the TRUS-guided biopsy rate (Fig. 1B). Analysis of the Victorian Cancer registry database showed that the highest rates of PSA testing were seen in the age groups 65–74 years (21 202 per 100 000) and 55–64 years (19 511 per 100 000), similar to the Australian data. There was a significant positive correlation between PSA rates and PCa incidence in all age groups, from 35–84 years.

Impact of PSA Testing in Younger Men

The largest increases in PSA testing during the period of the study were seen in men aged <55 years, where there were increases of 284 and 194% in the age groups 35–44 and 45–54 years, respectively (Fig. 2). Furthermore the largest increases in TRUS-guided biopsy rates were also seen in the age groups 35–44 years (174%) and 45–54 years (104% [data not shown]). Among those aged <45 years, 1101 men were needed to undergo PSA tests to detect one PCa (0.01%), 329 PSA tests were needed for one indication for TRUS-guided biopsy (0.3%) and 3.1 men received TRUS-guided biopsy (32.2%) per PCa detected. In the age group 45–54 years, 164 men were needed to undergo PSA testing (0.61%), 76.8 PSA tests were needed for one indication for TRUS-guided biopsy (1.3%) and 2.2 men needed to undergo TRUS-guided biopsy to detect one PCa (45% [Table 1]).

Figure 2.

Percentage increase in rates of PSA testing and PCa incidence from 2001 to 2008 in Australia.

Rates were calculated per 100 000 men. The incidence of PCa was calculated using Segi's world population.

Analysis of the Victorian Cancer database showed that the highest increase in PSA testing (220%) was seen in men aged <55 years, in keeping with the Australia-wide data. A third of PCas in Australia were detected in men aged <65 years and, among the PCas detected in men aged <65 years in Victoria, 76% were Gleason score ≤7 (Table 2). In patients aged <55 years, increased PSA testing was significantly correlated with the diagnosis of low grade (45%) and, in the age group 45–54 years, with low and intermediate grade tumours. A positive correlation with higher grade tumours (Gleason >7) was only seen in men aged >55 years.

Table 2. Correlation between PSA rates and Gleason scores for each age group in Victoria in the period 2001–2007
Age groupGleason scoreAverage number per yearPercentage tumours found in each age group (%)Correlation coefficientCIP
  1. Rates for each Gleason score were calculated per 100 000 men. Tumours were then expressed as a percentage of the total number of tumours in each age group in column.
35–44 yearsOverall91000.70.02–0.940.037
≤ 64430.80.250.960.010
72260.3−0.48–0.840.387
>717−0.1−0.77–0.630.705
Unknown2240.1−0.67–0.740.839
45–54 yearsOverall2191001.00.99–1.00.000
≤ 699450.80.12–0.950.021
779360.90.38–0.970.002
>71570.5−0.26–0.900.139
Unknown26120.2−0.61–0.780.662
55–64 yearsOverall10031001.00.99–1.00.000
≤ 6398400.90.55–0.980.000
7361360.90.39–0.970.002
>796100.7−0.03–0.940.047
Unknown148140.6−0.19–0.920.102
65–74 yearsOverall13251001.00.87–1.00.000
≤ 6441330.80.25–0.960.010
7463351.00.75–1.000.000
>7199150.90.39–0.970.002
Unknown222170.6−0.19–0.920.102
75–84 yearsOverall9051000.90.65–0.990.000
≤ 618921−0.5−0.88–0.340.207
7265300.80.12–0.950.021
>7227250.4−0.44–0.860.321
Unknown224240.80.25–0.960.010
≥85 yearsOverall2831000.50.88–0.340.207
≤ 62810−0.4−0.86–0.440.321
75620−0.1−0.77–0.620.705
>77025−0.6−0.92–0.150.086
Unknown12945−0.3−0.83–0.500.423

Discussion

To date, none of the previously published major clinical trials have assessed the clinical value of PCa screening in younger patients (i.e. patients aged <55 years) [5, 14]; therefore, most current guidelines such as those of the American Urology Association, the European Association of Urology, American Cancer Society, National Comprehensive Cancer Network and USANZ, do not recommend screening for patients <40 years [10, 15]. The Australian general practitioner guidelines during this period also did not advocate PSA testing in men aged <50 years [16]. Despite these recommendations against screening, our results showed a large increase in PSA testing in Australia, with a significant proportion of screening occurring in men aged <55 years. In comparison with the European Randomized Study of Screening for Prostate Cancer (ERSPC) study, where 17.7 screening tests were required to detect one PCa (5.65%) [4, 17], the rates seen in the present study were much lower, with 44.5 men being screened to detect one PCa in all age groups (2.25%) and 165–1011 men undergoing PSA testing to detect one PCa in men aged <55 years (0.01–0.61%).

Our findings are similar to the findings of the Surveillance, Epidemiology, and End Results (SEER) programme, which has shown that an additional million men are being over-diagnosed and treated for PCa as a result of PSA screening and an increasing incidence of PCa in younger men [3]. One possible reason for the increased rates of screening may be attributed to the increased screening in other family members because of the increasing awareness of the risk of familial PCa; however, this may cause a cascading effect as increasing numbers of younger men with low grade PCa are being detected, as seen in our study. One study showed that a single raised PSA level at or before 50 years may be used to predict the development of advanced PCa up to 30 years later and that 19% of men who have an initial PSA below the median level at ages 44–45 years will also develop advanced PCa [18]. Evidence from the Malmo Preventative Project shows that measurement of PSA concentration in early mid-life can identify a small group of men who are at increased risk of PCa metastasis several decades later. Men with PSA levels in the highest 10th of the distribution of concentrations at age 45–49 years contribute to nearly half of PCa deaths over the next 25–30 years. At least half of men can be identified as being at low risk, and probably need no more than three lifetime PSA tests (mid to late 40s, early 50s, and 60s) [17]. Our results showed that, as a result of the large proportion of men <55 years being screened and PCa detection rates of only 0.01–0.61%, there are increasing rates of negative TRUS-guided biopsy being performed and a large number of low grade tumours being detected in younger men. The stability of the Gleason score in younger patients in the present study shows the natural history of the biopsy-detected PCas. This further adds to the evidence that the progression of aggressiveness of PCa is likely to start at age ∼55 years. While some of these early diagnoses and treatment of PCa may still improve mortality rates [4], the overall risks and benefits should be considered. If PSA testing is to be performed in men aged <55 years, a strategy for follow-up similar to the recommendations of the Malmo Preventative Project should be undertaken so that those at low-risk of PCa mortality should not be subjected to yearly PSA testing.

The PSA test, when used appropriately, provides clinicians with valuable information to aid in the diagnosis and treatment of PCa. Currently, there is not a comparable test or diagnostic tool available for this purpose. Even in men aged > 55 years, there is no clear evidence as to whether early detection and treatment by PSA screening improves PCa-specific survival. The SEER Registry data shows a 75% reduction in the proportion of men with PCa who present with metastatic disease, as well as a 42% reduction in age-adjusted PCa mortality during the PSA era [19]; however, this could be a distortion because of lead-time bias with over-testing of PSA, detecting more indolent disease which may never progress to dangerous metastatic disease. Thus, as observational data can be flawed, more emphasis should be placed on randomized controlled studies of screening. The ERSPC trial showed a significant relative reduction in PCa mortality in the screening group of 29% at 11-year follow-up as a result of PSA screening [4, 17]. By contrast, the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial (which has its own important limitations, such as incomplete long-term outcome data and contamination of the control group) and a few smaller scale studies failed to show a benefit of PSA screening [15, 20]. Despite the reduction in PCa mortality in the ERSPC trial, these benefits were diminished by the loss of quality-adjusted life years resulting from post-diagnosis long-term effects [4, 17].

In addition to the absence of clear survival benefits with treatment, patients should be made aware of the disadvantages of screen-detected PCa, such as the treatment complications associated with prostatectomy and radiation therapy [6]. Active surveillance is an accepted alternative for localized and low-risk PCa and may avoid many of these complications, with no increase in reported anxiety or depression and good satisfaction rates in smaller scale studies [21, 22], but, with the advances in focal therapy which are more likely to be curative, studies suggest that younger men with low grade PCa are more likely to undergo definitive treatment rather than active surveillance [23]. Furthermore, the detection of low grade tumours in younger men increases the prevalence of PCa as a direct result of screening, with both active surveillance and radical treatment having been shown to place a significant burden on the economy [24]. With only a modest sensitivity of 20% and specificity of 94% at the ‘normal’ level of 4.1 ng/mL of a PSA test [15], the adverse psychological effects associated with a false-positive PSA test [25] should also not be underestimated.

One stark finding of the present study is the increased rate of TRUS-guided biopsy associated with PSA testing, where 50–66% of men aged <55 years who underwent TRUS-guided biopsy had no PCa. Furthermore, our study did not take into account the PCas detected using alternative methods such as TURP, endoscopic biopsy and cystoprostatectomy, suggesting that the rates of negative biopsy could be even higher than 66%. These results are similar to those of the ERSPC trial where 76% men who had TRUS-guided biopsy with a raised PSA did not result in PCa [4]. Although TRUS-guided biopsy plays a pivotal role in the diagnosis of PCa, rates of septic complications have been increasing in recent years in the USA [26] and, in Australia, overall complication rates could be as high as 22% after transperineal and 20% after transrectal biopsies, and overall septic complications up to 1.2% [27]. Thus, the risk profile of the patient, examination findings and symptomatic status should also be given strong consideration in younger men before biopsy. Although a number of risk calculators have been developed to help determine which patients should undergo a biopsy [15], none have been validated in Australian men.

The decreasing number of PSA tests used to detect one PCa in older age groups is probably attributable to the natural history of the disease [28]. Although most guidelines recommend the use of screening only in patients with a life expectancy of >10 years and discourage the use of routine screening in men aged >75 years [29-33], some men diagnosed with PCa after 70 years of age die from high grade disease [34]. This should not, however, be an argument for increased PSA screening in the older age groups; the decision to screen or treat should be based on health status rather than age [35].

Although we were unable to ascertain if the patients who underwent PSA testing were asymptomatic or at high risk for PCa, another Australian study showed that a high number of asymptomatic men aged >70 years, were still being screened [36], and, as a large majority of PSA testing did not warrant further TRUS-guided biopsy and there was no resulting increase in PCa, it is likely that these were true screening tests. A further limitation of the present study is that the coding used for TRUS-guided biopsy may have included double counting of a single patient; however, this is unlikely as it is very rare that one person would undergo two TRUS biopsies in 1 year. Furthermore, the threshold values for PSA testing and PSA values which ‘triggered’ a TRUS-guided biopsy were unavailable. The present study has biases inherent with observational data and lack of granularity of claims and, in addition, we did not look at outcomes regarding survival; however, it should be noted that our findings are similar to the SEER data. A large proportion of the Victorian cancer registry coded as ‘unknown’ especially in the higher age groups, may have also influenced some of the analysis.

In conclusion, despite the lack of clear survival benefits of PSA screening and the lack of evidence regarding screening younger men, there was a significant increase in the use of opportunistic PSA screening compared with the rates of PCa in Australia in the period 2001–2008, especially in men aged <55 years. These PSA testing increases resulted in high rates of negative TRUS-guided biopsy and increased diagnosis of low grade disease in younger age groups. Our results further demonstrate the stability of Gleason scoring in younger men, indicating that the aggressiveness of biopsy-detectable PCa is likely to start to increase at age ≥55 years.

In light of our findings, further robust studies focusing on PSA screening in younger men are warranted. PSA tests should only be performed in asymptomatic men aged < 55 years who understand and are willing to accept these risks, so as to minimize the negative consequences of PSA testing and reduce the exposure to an unnecessary biopsy. This study further highlights the need for better screening tests for PCa.

Acknowledgements

We would like to thank Vicky Thursfield, Cancer Control Information Manager of the Victorian Cancer Registry for data provision and Dr Mark Short, Senior Data Analyst, Cancer and Screening Unit, AIHW, for advice regarding data.

Conflict of Interest

None declared.

Ancillary