Changes in incidence, survival and mortality of prostate cancer in Europe and the United States in the PSA era: additional diagnoses and avoided deaths

(1)

Article

Reference

Changes in incidence, survival and mortality of prostate cancer in Europe and the United States in the PSA era: additional diagnoses

and avoided deaths

NEPPL-HUBER, C, et al . & EUNICE Survival Working Group USEL, Massimo (Collab.)

Abstract

We describe changes in prostate cancer incidence, survival and mortality and the resulting impact in additional diagnoses and avoided deaths in European areas and the United States.

NEPPL-HUBER, C, et al . & EUNICE Survival Working Group, USEL, Massimo (Collab.).

Changes in incidence, survival and mortality of prostate cancer in Europe and the United States in the PSA era: additional diagnoses and avoided deaths. Annals of Oncology , 2012, vol. 23, no. 5, p. 1325-34

PMID : 21965474

DOI : 10.1093/annonc/mdr414

Available at:

http://archive-ouverte.unige.ch/unige:44621

Disclaimer: layout of this document may differ from the published version.

(2)

Annals of Oncology23: 1325–1334, 2012 doi:10.1093/annonc/mdr414 Published online 28 September 2011

Changes in incidence, survival and mortality of prostate cancer in Europe and the United States in the PSA era:

additional diagnoses and avoided deaths

C. Neppl-Huber

¹

, M. Zappa

²

, J. W. Coebergh

³

, E. Rapiti

⁴

, J. Rachtan

⁵

, B. Holleczek

⁶

, S. Rosso

⁷

, T. Aareleid

⁸

, H. Brenner

¹

& A. Gondos

¹

* the EUNICE Survival Working Group

1Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Heidelberg, Germany;²Tuscany Cancer Registry, Clinical and Descriptive Epidemiology Unit, CSPO, Florence, Italy;³Department of Public Health, Erasmus MC Rotterdam, Rotterdam, The Netherlands;⁴Geneva Cancer Registry, Geneva, Switzerland;⁵Cracow Cancer Registry, Cracow, Poland;⁶Saarland Cancer Registry, Saarbru¨cken, Germany;⁷Piedmont Cancer Registry, Turin, Italy;⁸Department of Epidemiology and Biostatistics, National Institute for Health Development, Tallinn, Estonia

Received 11 May 2011; revised 21 July 2011; accepted 4 August 2011

Background:We describe changes in prostate cancer incidence, survival and mortality and the resulting impact in additional diagnoses and avoided deaths in European areas and the United States.

Methods:Using data from 12 European cancer registries and the Surveillance, Epidemiology and End Results program, we describe changes in prostate cancer epidemiology between the beginning of the PSA era (USA:

1985–1989, Europe: 1990-1994) and 2002–2006 among patients aged 40–64, 65–74, and 75+. Additionally, we examine changes in yearly numbers of diagnoses and deaths and variation in male life expectancy.

Results:Incidence and survival, particularly among patients aged<75, increased dramatically, yet both remain (with few exceptions in incidence) lower in Europe than in the United States. Mortality reductions, ongoing since the mid/late 1990s, were more consistent in the United States, had a distressingly small absolute impact among patients aged 40–64 and the largest absolute impact among those aged 75+. Overall ratios of additional diagnoses/avoided deaths varied between 3.6 and 27.6, suggesting large differences in the actual impact of prostate cancer incidence and mortality changes. Ten years of remaining life expectancy was reached between 68 and 76 years.

Conclusion:Policies reflecting variation in population life expectancy, testing preferences, decision aids and guidelines for surveillance-based management are urgently needed.

Key words:prostate cancer, prostate-specific antigen, registries, survival

introduction

Starting from the late 1980s, opportunistic screening for prostate cancer among asymptomatic men by means of prostate-specific antigen (PSA) testing [1–3] caused rapid rises in prostate cancer incidence and survival [4–7] in many Western countries, and by 2008, prostate cancer was estimated to have become the most common malignancy in men in Europe, as well as in North and South America [8]. Evidence from European Randomized Study of Screening for Prostate Cancer (ERSPC) [9] suggested the possibility of reducing prostate cancer mortality through PSA testing, and the better than overall performance of the

intervention in a subpopulation of that trial [10] suggests the possibility of substantial geographic variation in the performance of the intervention.

Treatment triggered by PSA testing can cause nonnegligible harm through overdiagnosis [11,12], overtreatment [13] and side-effects that can reduce quality of life [14,15]. PSA testing trial results convincingly indicate no mortality benefit within the first 10 years after screening [9,12], and the potentially low effectiveness of PSA testing in the population-based setting [16,17] additionally increase the difficulty of weighing the harms and benefits associated with PSA testing.

The more than a decade long utilization of PSA testing in the population enables the comparative assessment of the

population-level effects of the intervention. A recent analysis addressed overall prostate cancer mortality trends in Europe [18]. We provide a detailed international epidemiological comparison of age-specific incidence, survival and mortality between the beginning of the PSA era, defined as the calendar period of 1985–1989 in the United States and 1990–1994 in Europe and the early 21st century. Additionally, we attempt to quantify and compare the absolute impact of incidence and mortality changes in terms of additional diagnoses and avoided deaths and examine population life expectancies.

*Correspondence to:Dr A. Gondos, Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, INF 581, Heidelberg 69120, Germany.

Tel:+49-6221-421348; Fax:+49-6221-421302; E-mail:a.gondos@dkfz.de

The members of the EUNICE Survival Working Group are listed in the acknowledgements section.

ªThe Author 2011. Published by Oxford University Press on behalf of the European Society for Medical Oncology.

at Institut universitaire de hautes etudesinternationales - Bibliotheque on November 28, 2014http://annonc.oxfordjournals.org/Downloaded from

(3)

materials and methods

data sources

Cancer incidence data came from 12 long-standing population-based cancer registries participating in the European Network for Indicators of Cancer (EUNICE) Survival Cooperation, the NORDCAN database [19], the Surveillance, Epidemiology and End Results (SEER) Cancer Query Systems (CANQUES) for the United States and the International Agency for Research on Cancer (IARC) (for years up to 1997) [20]. National-level age-specific prostate cancer mortality data were obtained from the World Health Organization (WHO) mortality database [21], along with additional regional data from the cancer registries.

survival analysis

For European regions, the database of EUNICE Survival Cooperation was used; data preparation, inclusion criteria for survival analysis and population life tables used to calculate relative survival estimates have been described in detail elsewhere [22]. Briefly, this analysis includes men aged‡40 diagnosed with malignant tumors of the prostate [International Classification of Diseases (ICD)-10 code C61] between 1985 and 2004, excluding diagnoses confirmed by autopsy or death certificate only. For the United States, the public use database of the SEER Program was used. White male prostate cancer patients from the SEER 9 registries were selected using the same criteria as for European patients. Life tables for USA whites, obtained from the US National Center for Health Statistics, for the years 1990, 2000 and 2004 were used to cover the periods of 1990–1995, 1996–2002 and 2003–2004, respectively.

Five-year relative survival estimates were calculated for the age groups of 40–64, 65–74 and 75+, for the periods 1990–1994, 1995–1999, and 2000–

2004. Relative survival estimates were derived as a ratio of the absolute survival of the cancer patients divided by the expected survival of an age- matched group of the underlying male general population [23]. Expected survival estimates were derived using the Ederer II method [24]. Life expectancy was calculated on the basis of the population-specific life tables.

We used a saturated Poisson regression model equivalent to period analysis to derive survival estimates [25]. To obtain a test for trend, the above model was extended: the logarithm of the excess number of deaths was modeled as a function of calendar period and year of follow-up, with the logarithm of the person-years (p.y.) at risk as offset [26]. For registries with incidence data up to 2003, but follow-up data through 2004, hybrid

analysis was used to take advantage of mortality follow-up data that were more up to date than incidence data [27]. Survival calculations were carried out with the SAS statistical software package using adapted versions of previously described SAS macros [28].

age adjustment of incidence and mortality data and calculation of changes in the numbers of surplus diagnoses and reductions in deaths of prostate cancer in the PSA era

Truncated (ages 40+) age-specific incidence and mortality data adjusted to the World Standard Population (within the age group) were available from the IARC/WHO databases [20,21]. Cancer registry data were age adjusted to the World Standard [29]. Net differences between incidence/mortality rates at the beginning of the PSA era (1985–1989 in the United States and 1990–1994 in Europe) and 2002–2006 were calculated, and these were used to calculate age-specific change in the yearly number of diagnoses and number of men dying of prostate cancer, based on a hypothetical underlying population of 1 million men with the age distribution of the 2000 USA male population (equivalent to 300 000, 60 000 and 44 000 men in the age groups 40–64, 65–74 and 75, respectively) [30]. Overall ratios of net standardized change in incidence and mortality were calculated to quantify the absolute effect of incidence and mortality changes in terms of additional diagnoses and avoided deaths in the PSA era.

results

Table 1provides case numbers and median age in each participating cancer registry in 1990–1994, 1995–1999 and 2000–2004. Overall, 401 162 prostate cancers were included in the analysis. Large increases in the numbers of cases in all registries were accompanied by, except in Estonia and Lithuania, a decrease in the median age at diagnosis of between 1 and 5 years, with particularly strong changes in Geneva (from 73 to 68), Finland (from 74 to 70), Norway (from 75 to 72) and in the United States (from 71 to 68).

Table 2presents age-specific results of the survival analysis.

With the exception of the oldest age group in Slovenia and Geneva, the 5-year relative survival of patients has increased

Table 1. Participating cancer registries, number of included patients aged 40+for the periods of 1990–1994, 1995–1999 and 2000–2004, median age of cases, and % change in overall truncated (ages 40+) age-adjusted incidence

Registry Country Base

population (millions)

1990–1994 Median age

% change in age-adjusted overall incidence^a

Cracow Poland 0.8 256 71 460 71 619 70 +125.3

Estonia Estonia 1.4 1212 69 1543 71 1635 71 +108.8

Lithuania Lithuania 3.6 2241 72 3446 72 6828 72 +280.3

Slovenia Slovenia 1.9 1293 72 2113 70 2364 70 +130.1

Turin Italy 1.0 1184 73 2131 72 3265 71 +145.4

Tuscany Italy 1.2 1931 74 2764 74 2799 73 +74.7

Eindhoven The Netherlands 1.0 1444 72 3720 71 5353 70 +64.4

Scotland UK 5.1 7784 74 10 129 73 11 735 72 +53.4

Finland Finland 5.2 8563 74 14 057 72 20 226 70 +114.1

Norway Norway 4.5 10 958 75 13 404 74 15 653 72 +63.7

Geneva Switzerland 0.4 691 73 1049 70 1427 68 +74.5

Saarland Germany 1.0 1639 71 2188 69 3342 69 +68.7

USA SEER 9 whites USA 19.0 78 178 71 71 827 70 79 678 68 28.8

SEER, Surveillance, Epidemiology and End Results.

aChange between 1990–1994 and 2002–2006

(4)

statistically significantly and meaningfully in each age group and registry between 1990–1994 and 2000–2004. In 2000–2004, in Europe, estimates ranged between 64% (Lithuania) and 95%

(Geneva) for the age group 40–64, between 67% (Cracow) and 95% (Turin) for the age group 65–74 and between 34%

(Cracow) and 84% (Saarland) in the age group 75+. The United States, where survival was already very high in 1990–1994, had the highest survival in all age groups in both 1990–1994 and 2000–2004.

Prostate cancer incidence (Table 3) rose strongly with age, with the oldest age group having a 4–11 times higher age-

specific incidence than the youngest in 2002–2006. With the exception of the United States, incidence was always highest in the oldest age group. Among 40- to 64-year-olds, incidence was lower in Central and Southern Europe as well as in Scotland and Eindhoven (rates between 40 and 93 per 100 000 p.y.) than in Northern Europe, Saarland and Geneva (rates between 106 and 146 per 100 000 p.y.). Incidence in the United States was 30% higher in this age group than the highest incidence in Europe. In the age group 65–74, incidence in Cracow was markedly lower (264 per 100 000 p.y.) than in Tuscany, Slovenia, Scotland and Estonia (470–550 per 100 000 p.y.);

rates were between 640 and 700 per 100 000 p.y. in Turin, Eindhoven and Saarland, and highest in Geneva, Lithuania, Norway and Finland (760–870 per 100 000 p.y.). Again, incidence in the United States, at920 per 100 000 p.y., was highest. In the oldest age group, patterns were more uniform:

except the much lower incidence in Cracow than elsewhere, rates varied between670–815 per 100 000 p.y. in all registries except Lithuania, Norway and Finland, where rates exceeded the 1000 per 100 000 p.y. mark and were higher than the rate seen in the United States.

Rises in the incidence between the beginning of the PSA era and 2002–2006 had, except for Estonia, a similar pattern in Europe: the strongest rise, in percentage, was seen in the youngest age group while changes were smallest in the oldest age group. In Europe, incidence increased in all registries in the two youngest age groups; in contrast, essentially unchanged incidence was seen in the oldest age group in Eindhoven, Scotland, Norway and Saarland, and incidence decreased by 10% in Geneva. In the United States, incidence nearly tripled in the age group 40–64, increased by 44% in the age group 65–74, but decreased by 27% in age group 75+.

For mortality, age-specific rates were between 36 and 80 times higher in the oldest age group than in the youngest in 2002–2006. However, the absolute variation between registries was small: rates were between 4 and 13 per 100 000 p.y. in the youngest age group and between 54 and 139 per 100 000 p.y.

and 319 and 619 per 100 000 p.y. in the age groups 65–74 and 75+, respectively. Mortality rose in all age groups in Cracow, Estonia, Lithuania and Slovenia, remained largely stable in Scotland, decreased in Tuscany (Italy) in the two youngest age groups by20%, and decreased by between 5% and 32% in all three age groups in the remaining registries. In the United States, rates declined by 23%–38%.

Table 3also shows the age-specific change in the numbers of men diagnosed and dying of prostate cancer in 1 year (based on 1 million men with the USA age structure as underlying population) by age group and overall. In the age group 40–64, the yearly number of additional diagnosis was between 82 and 354 while changes in the yearly number of men dying of prostate cancer did not exceed 8. In the age group 65–74, numbers of additional diagnoses ranged between 85 and 349 while the change in the number of deaths ranged between+21 and226. In the oldest age group, corresponding ranges for additional diagnoses were2133 and+301, and+106 and289 for deaths. Reductions in the number of death exceeded the number of additional diagnosis in Saarland and Eindhoven, they were approximately equal in Norway, and declines in additional diagnosis were accompanied by even stronger

Table 2. Period estimates of 5-year relative survival of patients with prostate cancer, by participating registry, age group, and calendar period

Registry Age group

Survival

1990–1994 1995–1999 2000–2004

PE SE PE SE PE SE Diff P-value

Cracow 40–64 40.1 7.7 44.9 6.8 75.0 4.7 34.9 <0.001 65–74 29.7 5.3 38.4 6.2 67.3 4.4 37.6 <0.001 75+ 27.7 7.7 33.7 8.3 34.5 5.3 6.8 0.025 Estonia 40–64 45.9 4.6 60.2 3.1 66.6 3.2 20.7 <0.001 65–74 48.3 3.9 63.3 2.9 68.2 2.4 19.9 <0.001 75+ 36.9 4.5 54.1 4.5 55.5 3.7 18.6 0.001 Lithuania 40–64 22.5 4.5 44.4 2.5 63.7 2.1 41.2 <0.001 65–74 34.7 4.2 49.8 2.1 69.0 1.6 34.3 <0.001

75+ 43.4 5.2 47.9 2.7 62.2 2.4 18.8 <0.001

Slovenia 40–64 41.7 4.0 54.2 3.3 77.6 2.1 35.9 <0.001 65–74 49.0 3.3 52.2 2.5 67.7 1.9 18.7 <0.001 75+ 45.0 3.7 49.0 3.5 46.8 2.9 1.8 0.845 Turin 40–64 54.8 4.3 78.5 3.0 90.6 1.8 35.8 <0.001 65–74 62.5 3.5 78.4 2.3 94.9 1.4 32.4 <0.001

75+ 65.8 3.8 69.0 3.4 82.9 2.6 17.1 <0.001

Tuscany 40–64 59.7 3.9 82.0 2.6 88.0 1.8 28.3 <0.001 65–74 60.7 2.6 76.9 1.9 89.2 1.4 28.5 <0.001

75+ 51.1 2.8 56.9 2.5 69.5 2.2 18.4 <0.001

Eindhoven 40–64 64.3 3.9 82.0 2.1 86.1 1.3 21.8 <0.001 65–74 70.8 3.3 79.2 1.9 84.3 1.3 13.5 <0.001 75+ 59.8 4.8 77.3 3.4 75.1 2.5 15.3 0.019 Scotland 40–64 54.8 1.9 69.1 1.5 76.6 1.2 21.8 <0.001 65-74 54.1 1.3 64.8 1.1 76.7 1.0 22.6 <0.001

75+ 45.2 1.5 56.0 1.4 64.4 1.3 19.2 <0.001

Finland 40–64 58.2 1.7 73.1 1.3 88.7 0.7 30.5 <0.001 65–74 66.2 1.3 76.4 0.9 86.8 0.7 20.6 <0.001

75+ 60.9 1.5 68.4 1.3 77.7 1.2 16.8 <0.001

Norway 40–64 60.2 1.8 74.6 1.3 85.9 0.8 25.7 <0.001 65–74 63.9 1.1 72.7 0.9 83.6 0.8 19.7 <0.001

75+ 55.9 1.2 62.6 1.1 68.7 1.1 12.8 <0.001

Geneva 40–64 64.0 5.7 78.6 4.2 94.8 1.7 30.8 <0.001 65–74 67.2 4.1 76.7 3.1 91.7 2.0 24.5 <0.001 75+ 61.7 4.1 58.9 4.6 58.1 4.0 23.6 0.600 Saarland 40–64 76.9 3.0 82.7 2.2 92.8 1.4 15.9 <0.001 65–74 80.4 2.9 83.0 2.1 91.4 1.6 11.0 <0.001

75+ 65.1 4.4 76.7 3.8 84.5 3.3 19.4 <0.001

USA SEER whites

40–64 89.0 0.4 96.7 0.2 98.8 0.2 9.8 <0.001 65–74 91.5 0.4 97.1 0.2 99.5 0.2 8.0 <0.001

75+ 86.3 0.6 88.7 0.5 92.1 0.5 5.8 <0.001

PE, period estimate; SE, standard error; SEER, Surveillance, Epidemiology and End Results. Diff=change in 5-year relative survival between 1990–

1994 and 2000–2004, in % units.

Volume 23|No. 5|May 2012 doi:10.1093/annonc/mdr414|¹³²⁷

(5)

declines in the numbers of deaths in Geneva and the United States. Overall, the number of additional diagnosis was3 to 8 times higher than reductions in the number of deaths Geneva, the United States, Eindhoven and Saarland;10 to 12 times higher in Turin and Norway;17 to 20 times higher in Scotland and Finland; and28 times higher in the Tuscany.

In the absence of mortality declines, ratios remained negative for Eastern European populations.

Figure 1presents yearly values of incidence and mortality rates with the underlying survival trends during the study period. Incidence appeared to rise in the youngest age group in most registries up to 2006, with higher rates in 2006 than the average rate for 2002–2006 in all registries except Finland, Saarland and the United States. The striking rise in incidence in Lithuania in the year 2006 is plausibly due to the starting of a national PSA-based early detection program in that year [31].

Table 3. Changes in truncated (ages 40+) age-adjusted age-specific prostate cancer incidence and mortality rates, resulting rise or decline in the annual numbers of diagnoses and deaths, and the ratio of these, by age group and participating registry

Incidence^a Mortality^b

Registry Age group 2002–2006 % change since 1990–1994^c

Change in number of diagnoses^d

2002–2006 % change since 1990–1994^c

Change in number of deaths^d

Overall ratio^e

Cracow 40–64 40.0 +222.1 +82 7.7 +5.3 1

65–74 263.7 +114.3 +85 87.8 +6.4 3

75+ 389.3 +73.5 +73 321.2 +58.7 53 24.2

Estonia 40–64 70.2 +117.3 +113 13.0 +17.0 6

65–74 554.7 +120.4 +182 139.4 +33.3 21

75+ 794.1 +83.9 +160 492.2 +94.3 106 23.4

Lithuania 40–64 96.6 +417.0 +233 12.3 +21.3 6

65–74 787.8 +278.6 +349 133.8 +17.7 12

75+ 1034.7 +192.5 +301 444.8 +70.5 81 28.8

Slovenia 40–64 63.7 +293.0 +142 9.5 +29.0 6

65–74 470.6 +143.4 +167 117.4 +28.4 16

75+ 666.3 +44.6 +91 555.3 +45.0 76 24.1

Turin 40–64 89.6 +268.8 +195 4.0 239.3 28

65–74 640.7 +176.3 +246 54.4 237.5 220

75+ 688.0 +35.9 +80 322.2 214.2 224 10.2

Tuscany 40–64 64.9 +163.6 +120 4.4 218.1 23

65–74 468.5 +83.8 +128 55.1 223.7 210

75+ 642.0 +15.2 +38 319.3 +2.0 3 27.6

Eindhoven 40–64 92.6 +133.3 +158 7.1 26.6 21

65–74 645.5 +76.4 +168 95.4 213.1 29

75+ 732.5 +1.7 +6 457.4 219.4 249 5.6

Scotland 40–64 68.9 +150.4 +124 8.2 +3.5 1

65–74 482.8 +63.3 +112 94.4 211.1 27

75+ 725.0 21.7 26 405.9 23.7 27 17.4

Finland 40–64 143.0 +318.7 +325 8.1 210.7 23

65–74 870.2 +111.3 +276 96.5 216.5 211

75+ 1216.9 +26.4 +112 452.5 29.9 222 19.6

Norway 40–64 118.1 +158.4 +216 8.6 216.4 25

65–74 792.2 +69.9 +196 118.5 217.8 215

75+ 1039.1 +2.9 +13 619.3 25.4 216 11.8

Geneva 40–64 146.0 +250.4 +312 6.0 227.9 27

65–74 759.4 +61.4 +174 92.8 231.9 226

75+ 815.5 29.9 239 491.4 229.1 289 3.6

Saarland 40–64 105.8 +133.4 +181 7.0 210.4 22

65–74 695.7 +80.0 +186 80.0 223.7 215

75+ 759.2 +3.7 +12 395.3 215.1 231 7.8

USA SEER whites 40–64 186.1 +175.5 +354 6.1 231.1 28

65–74 918.5 +44.0 +169 66.8 238.1 225

75+ 831.3 226.6 2133 312.4 222.8 241 5.3

SEER, Surveillance, Epidemiology and End Results; Changes in the number of diagnoses/deaths are based on 1 million men in the 2000 USA population.

aIncidence and mortality data are rates per 100 000 and age standardized using the World Health Organization World Standard.

bFor Cracow, Tuscany, Eindhoven, Geneva and Saarland, national mortality data were used.

cFor the United States: % change since 1985–1989 for both incidence and mortality.

dChanges in the number of diagnosis/deaths are based on 1 million men in the 2000 US population.

eRatio of surplus diagnosis over mortality decline; negative ratios indicate no mortality decline or declines in both incidence and mortality.

(6)

Figure 1. Trends in age-specific incidence, mortality and 5-year relative survival of prostate cancer between the beginning of the prostate-specific antigen era and 2006, by registry. Incidence and mortality rates are adjusted to the World Standard within each age group.

Volume 23|No. 5|May 2012 doi:10.1093/annonc/mdr414|¹³²⁹

(7)

Figure 1. (Continued).

(8)

Volume 23|No. 5|May 2012 doi:10.1093/annonc/mdr414|¹³³¹

(9)

In the age group 65–74, a peak and subsequent decline in incidence has already manifested in Saarland and the United States, and may be under way in Finland and Geneva; incidence may be stabilizing in several other populations. In the oldest age group, stable or even slightly declining incidence is seen since the mid/late 1990s in Turin, Eindhoven, Scotland, Finland, Norway, Geneva, and Saarland, while the strongest decline is found in the United States. These incidence declines were uniformly accompanied by declines in mortality as well.

InTable 4, life expectancies of men in the populations included in our analysis are shown. Both overall and conditional life expectancy varied considerably among the populations, with the exact age at which cumulative 10-year life expectancy became<50% varying between 68 and 76 years.

discussion

The PSA era brought about, particularly among patients<75 years of age, strong rises in prostate cancer incidence and survival in both Europe and the United States. PSA utilization spread much quicker in the United States (the rise in incidence was similar to that in Lithuania in 2006, the starting year of the national PSA- based prostate cancer early detection program [31] in that country), and both incidence (with very few exceptions) and survival (uniformly) remain higher in the United States than in Europe. The impressive rises in relative survival are clearly overwhelmingly explained by lead-time and overdiagnosis effects, which strongly compromise the epidemiological interpretability of the changes. Relative declines in mortality were larger and more consistent in the United States compared with European areas. In contrast to other populations, mortality rises were seen in Eastern European populations, and it was argued that this may be an indication of increased risk [18]. Our results (Table 3,Figure 1) indicate that in all of these four countries, mortality rises affect the oldest age group far stronger than younger ones. It is difficult to explain the geographic disparity and the disproportionate increase by age. Possible explanations include differential degrees of death misattribution by country, age and time period [32].

Relative (and sometimes even absolute) incidence rises were largest in the age groups of 40–64 and 65–74, the absolute

mortality effect was by far weakest in the age group 40–64 and strongest among patients aged 75+. Everywhere, the number of additional diagnosis in the age group 40–64 alone exceeded, often many times over, the total reduction in the numbers of deaths over the entire age range, as achieved by mortality reductions up to 2002–2006. The large variation in overall additional diagnoses versus reductions in prostate cancer mortality indicates considerable geographic differences in the overall effect of disease measure changes by 2006. For many European populations, ratios are considerably less favorable than for the Unites States. The use of age-adjusted incidence and mortality rates mean that the effect of population aging is completely removed from the trends in our analyses.

Unfavorable ratio of surplus incidence over reduction in mortality through PSA testing was found before in the European trial of PSA testing [9], as well as in population-based analysis [13]. On the population level, beside PSA utilization, other factors may have influenced incidence and mortality. Some

Table 4. Life expectancy (birthdays between which 50% cumulative overall mortality is reached) among men at birth; at 65, 70, and 75 years of age; and oldest exact age with not<10 years of life expectancy in the populations of the participating registries

Life table Male life expectancy

Registry Period At birth At 65 At 70 At 75 Exact age

Cracow 2000–2004 72–73 12–13 9–10 6–7 70

Estonia 2000–2004 67–68 10–11 8–9 6–7 68

Lithuania 2000–2003 68–69 11–12 9–10 6–7 70 Slovenia 2000–2004 73–74 13–14 9–10 6–7 71

Turin 2000–2004 79–80 16–17 12–13 8–9 74

Tuscany 2000–2003 79–80 16–17 12–13 9–10 75 Eindhoven 2000–2004 77–78 14–15 10–11 7–8 71 Scotland 1998–2003 75–76 13–14 10–11 7–8 71

Finland 2001–2004 77–78 15–16 11–12 8 73

Norway 2000–2004 78–79 15–16 11–12 8–9 73

Geneva 2000–2004 80–81 17–18 13–14 9–10 76 Saarland 2003–2007 77–78 16–17 11–12 10–11 73

USA 2004 78–79 16–17 12–13 9–10 75

(10)

natural rise in incidence may have occurred in the pre-PSA era already [33], making underlying incidence rates higher even in the absence of PSA testing. However, early detection, e.g. through transurethral resection of the prostate (TURP) for benign prostatic disease, was shown to explain most of the increase in incidence seen in the United States between 1973 and 1986 [34].

Comparable TURP trends, though with a smaller explanatory power, were reported from Europe also [35].

Data from clinical trials suggest that PSA testing is unlikely to produce a mortality benefit within the first 10 years after testing [9,10]. Consequently, it is probably only after 2002 that a part of mortality decline may be due to PSA testing in the United States (where testing peaked in 1992), while in Europe, where utilization was lower in the 1990s, any larger scale effect will probably manifest in even later years. With mortality declines in the two older age groups often ongoing since the early or mid 1990s (Figure 1), other factors, rather than PSA testing, are probably responsible for most of the mortality decline.

Clinical progress with neoadjuvant and adjuvant hormonal therapy, which improves disease-specific survival and

sometimes also overall survival [36], may have contributed to lower mortality rates. Furthermore, the long-term use of aspirin and other nonsteroidal anti-inflammatory drugs, reducing the risk of prostate cancer [37–39], as well as advancement of diagnosis to younger age groups may have contributed to incidence declines seen in the oldest age group since the mid/late 1990s. On the other hand, for mortality, statins, increasingly utilized since the late 1990s [40], that have been suggested to reduce the risk of advanced prostate cancer [41, 42] may also have played a role. The share of PSA testing in incidence rises is probably very substantial; its effect on mortality reduction up to 2006 is probably partial, suggesting that a worse total balance than implied by the calculated ratios.

Since 2008, the US Preventive Task Force has recommended against PSA testing in men>75 years of age, as with<10 years of life expectancy, any benefit is unlikely [43]. Data indicate that a substantial proportion of men are being tested even if the available evidence clearly indicates that they cannot obtain a benefit [44–47]. For many European populations, lower life expectancy suggests that PSA testing should often stop already 70 years of age. Shared and informed decision making is therefore essential to determine patient preferences and the best course of action before and after testing [48]. In practice, this recommendation often remains unfulfilled [49] to which lack of understanding and overestimation of screening benefits in the general population [50] certainly contributes. As guidelines for shared decision making exist [51,52], implementing this approach should have a very high priority.

Several limitations of this study require careful consideration.

A considerable proportion of men dying within a given year from prostate cancer (and so contributing to mortality) have been diagnosed during preceding years, and this, as well as misclassification [53] turn mortality into an unspecific measure. Furthermore, due to the long lead time associated with a PSA-related diagnosis and the fact that prostate cancer is often a slowly progressing disease, a long time might be necessary until progress against cancer translates into

a mortality decline on the population level, and the latter may even occur in a different age group than the diagnosis. It cannot

be excluded that PSA testing associated mortality declines will occur in the future—the more time this should take, the more likely it is that such decline will affect older ages. More detailed data, particularly uniformly assessed stage, could have enabled deeper insight into reasons for survival differences.

In summary, highly elevated incidence, large geographical variation in the dynamics of prostate cancer, considerable international differences in life expectancy, the need for tools that enable informed and shared decision making, as well as evidence that testing continues to be widespread in older populations that are likely to be only harmed suggest that the complex issue of prostate cancer screening may be best addressed at the policy level.

acknowledgements

Members of the EUNICE Survival Working Group: Tiiu Aareleid (National Institute for Health Development, Estonia), Freddie Bray (Cancer Registry of Norway, Norway), Hermann Brenner (German Cancer Research Center, Germany), David H. Brewster (Scottish Cancer Registry, UK), Jan Willem Coebergh

(Eindhoven Cancer Registry, The Netherlands), Emanuele Crocetti (Tuscany Cancer Registry, Italy), Adam Gondos (German Cancer Research Center, Germany), Timo Hakulinen (Finnish Cancer Registry, Finland), Bernd Holleczek (Saarland Cancer Registry, Germany), Maryska Janssen-Heijnen (Eindhoven Cancer Registry, The Netherlands), Margit Ma¨gi (Cancer Registry of Estonia, Estonia), Jadwiga Rachtan (Cracow Cancer Registry, Poland), Roberto Zanetti (Piedmont Cancer Registry, Torino, Italy), Giedre Smailyte (Lithuanian Cancer Registry, Lithuania), Massimo Usel (Geneva Cancer Registry, Switzerland), Maja Primic Zˇakelj (Cancer Registry of Slovenia).

funding

This study was partly supported by a grant from the European Commission (Directorate of SANCO Luxemburg) for the European Network for Indicators of Cancer (EUNICE) by the intramural funding program (K207) of German Cancer Research Center and by a grant from the German Cancer Aid (108 257).

disclosure

The authors declare no conflict of interest.

references

1. Baade PD, Youlden DR, Krnjacki LJ. International epidemiology of prostate cancer:

geographical distribution and secular trends. Mol Nutr Food Res 2009; 53: 171–184.

2. Hankey BF, Feuer EJ, Clegg LX et al. Cancer surveillance series: interpreting trends in prostate cancer—part I: evidence of the effects of screening in recent prostate cancer incidence, mortality, and survival rates. J Natl Cancer Inst 1999;

91: 1017–1024.

3. Potosky AL, Miller BA, Albertsen PC, Kramer BS. The role of increasing detection in the rising incidence of prostate cancer. JAMA 1995; 273: 548–552.

4. Jemal A, Clegg LX, Ward E et al. Annual report to the nation on the status of cancer, 1975-2001, with a special feature regarding survival. Cancer 2004;

101: 3–27.

5. Verdecchia A, Guzzinati S, Francisci S et al. Survival trends in European cancer patients diagnosed from 1988 to 1999. Eur J Cancer 2009; 45:

1042–1066.

Volume 23|No. 5|May 2012 doi:10.1093/annonc/mdr414|¹³³³

(11)

6. Brenner H, Arndt V. Long-term survival rates of patients with prostate cancer in the prostate-specific antigen screening era: population-based estimates for the year 2000 by period analysis. J Clin Oncol 2005; 23: 441–447.

7. Gondos A, Bray F, Hakulinen T, Brenner H. Trends in cancer survival in 11 European populations from 1990 to 2009: a model-based analysis. Ann Oncol 2009; 20: 564–573.

8. World cancer report 2008; Chapter 1.1. Introduction: needs and prospects for cancer control. In Boyle P, Levin B (eds), World Cancer Report 2008. Lyon, France: International Agency for Research on Cancer 2008; 12–39.

9. Schroder FH, Hugosson J, Roobol MJ et al. Screening and prostate-cancer mortality in a randomized European study. N Engl J Med 2009; 360: 1320–1328.

10. Hugosson J, Carlsson S, Aus G et al. Mortality results from the Goteborg randomised population-based prostate-cancer screening trial. Lancet Oncol 2010; 11: 725–732.

11. Etzioni R, Penson DF, Legler JM et al. Overdiagnosis due to prostate-specific antigen screening: lessons from U.S. prostate cancer incidence trends. J Natl Cancer Inst 2002; 94: 981–990.

12. Draisma G, Etzioni R, Tsodikov A et al. Lead time and overdiagnosis in prostate- specific antigen screening: importance of methods and context. J Natl Cancer Inst 2009; 101: 374–383.

13. Welch HG, Albertsen PC. Prostate cancer diagnosis and treatment after the introduction of prostate-specific antigen screening: 1986-2005. J Natl Cancer Inst 2009; 101: 1325–1329.

14. Brawley OW, Ankerst DP, Thompson IM. Screening for prostate cancer. CA Cancer J Clin 2009; 59: 264–273.

15. Wilt TJ, MacDonald R, Rutks I et al. Systematic review: comparative effectiveness and harms of treatments for clinically localized prostate cancer. Ann Intern Med 2008; 148: 435–448.

16. Wever EM, Draisma G, Heijnsdijk EA et al. Prostate-specific antigen screening in the United States vs in the European Randomized Study of screening for prostate cancer—Rotterdam. J Natl Cancer Inst 2010; 102: 352–355.

17. Albertsen PC. Efficacy vs effectiveness in prostate-specific antigen screening. J Natl Cancer Inst 2010; 102: 288–289.

18. Bray F, Lortet-Tieulent J, Ferlay J et al. Prostate cancer incidence and mortality trends in 37 European countries: an overview. Eur J Cancer 2010; 46: 3040–3052.

19. Engholm G, Ferlay J, Christensen N et al. NORDCAN - a Nordic tool for cancer information, planning, quality control and research. Acta Oncol 2010; 49(5):

725–736.

20. International Agency for Research on Cancer. (IARC). CI5 I-VIII Detailed Database.

Lyon, France: IARC 2005.

21. International Agency for Research on Cancer (IARC). WHO Mortality Database.

Lyon, France: IARC 2008.

22. Gondos A, Bray F, Brewster DH et al. Recent trends in cancer survival across Europe between 2000 and 2004: a model-based period analysis from 12 cancer registries. Eur J Cancer 2008; 44: 1463–1475.

23. Ederer F, Axtell LM, Cutler SJ. The relative survival rate: a statistical methodology. J Natl Cancer Inst Monogr 1961; 6: 101–121.

24. Ederer F, Heise H. Instructions to IBM 650 Programmers in Processing Survival Computations. In Methodological note no. 10. Bethesda, MD: End Results Section, National Cancer Institute 1959.

25. Brenner H, Gefeller O, Hakulinen T. Period analysis for ’up-to-date’ cancer survival data: theory, empirical evaluation, computational realisation and applications. Eur J Cancer 2004; 40: 326–335.

26. Brenner H, Hakulinen T. Up-to-date and precise estimates of cancer patient survival: model-based period analysis. Am J Epidemiol 2006; 164: 689–696.

27. Brenner H, Hakulinen T. Model based hybrid analysis of cancer patient survival.

Eur J Cancer 2007; 43: 921–927.

28. Brenner H, Gefeller O, Hakulinen T. A computer program for period analysis of cancer patient survival. Eur J Cancer 2002; 38: 690–695.

29. Cancer Incidence in Five Continents Volumes I to IX. IARC June 2010; http://

ci5.iarc.fr/CI5i-ix/glossary.htm (26 August 2011, date last accessed).

30. U.S. Census Bureau. QT-P1: Age Group and Sex: 2000.

31. Jankevicius F, Adomaitis R, Jievaltas M et al. The Launch of the Lithuanian Early Prostate Cancer Detection Program (EPCDP). Eur Urol Suppl 2007;

6(2): 174.

32. Hoffman RM, Stone SN, Hunt WC et al. Effects of misattribution in assigning cause of death on prostate cancer mortality rates. Ann Epidemiol 2003; 13:

450–454.

33. Post PN, Stockton D, Davies TW, Coebergh JW. Striking increase in incidence of prostate cancer in men aged<60 years without improvement in prognosis. Br J Cancer 1999; 79: 13–17.

34. Merrill RM, Feuer EJ, Warren JL et al. Role of transurethral resection of the prostate in population-based prostate cancer incidence rates. Am J Epidemiol 1999; 150: 848–860.

35. Andren O, Garmo H, Mucci L et al. Incidence and mortality of incidental prostate cancer: a Swedish register-based study. Br J Cancer 2009; 100: 170–173.

36. Kumar S, Shelley M, Harrison C et al. Neo-adjuvant and adjuvant hormone therapy for localised and locally advanced prostate cancer. Cochrane Database Syst Rev 2006; (4): CD006019.

37. Harris RE, Beebe-Donk J, Doss H, Burr Doss D. Aspirin, ibuprofen, and other non-steroidal anti-inflammatory drugs in cancer prevention: a critical review of non-selective COX-2 blockade (review). Oncol Rep 2005; 13:

559–583.

38. Jacobs EJ, Rodriguez C, Mondul AM et al. A large cohort study of aspirin and other nonsteroidal anti-inflammatory drugs and prostate cancer incidence. J Natl Cancer Inst 2005; 97: 975–980.

39. Mahmud SM, Franco EL, Aprikian AG. Use of nonsteroidal anti-inflammatory drugs and prostate cancer risk: a meta-analysis. Int J Cancer 2010; 127(7):

1680–1691.

40. Walley T, Folino-Gallo P, Stephens P, Van Ganse E. Trends in prescribing and utilization of statins and other lipid lowering drugs across Europe 1997-2003. Br J Clin Pharmacol 2005; 60: 543–551.

41. Bonovas S, Filioussi K, Sitaras NM. Statin use and the risk of prostate cancer:

a metaanalysis of 6 randomized clinical trials and 13 observational studies. Int J Cancer 2008; 123: 899–904.

42. Platz EA, Leitzmann MF, Visvanathan K et al. Statin drugs and risk of advanced prostate cancer. J Natl Cancer Inst 2006; 98: 1819–1825.

43. U.S. Preventive Services Task Force. Screening for prostate cancer: U.S.

Preventive Services Task Force recommendation statement. Ann Intern Med 2008; 149: 185–191.

44. D’Ambrosio GG, Campo S, Cancian M et al. Opportunistic prostate-specific antigen screening in Italy: 6 years of monitoring from the Italian general practice database. Eur J Cancer Prev 2010; 19: 413–416.

45. Drazer MW, Huo D, Schonberg MA et al. Population-based patterns and predictors of prostate-specific antigen screening among older men in the United States. J Clin Oncol 2011; 29: 1736–1743.

46. Zappa M, Visioli C, Crocetti E et al. Practice of opportunistic PSA screening in the Florence District. Eur J Cancer Prev 2003; 12: 201–204.

47. Zeliadt SB, Hoffman RM, Etzioni R et al. Influence of publication of US and European prostate cancer screening trials on PSA testing practices. J Natl Cancer Inst 2011; 103: 520–523.

48. Barry MJ, Mulley AJ Jr. Why are a high overdiagnosis probability and a long lead time for prostate cancer screening so important? J Natl Cancer Inst 2009; 101:

362–363.

49. Lin GA, Aaronson DS, Knight SJ et al. Patient decision AIDS for prostate cancer treatment: a systematic review of the literature. CA Cancer J Clin 2009; 59:

379–390.

50. Gigerenzer G, Mata J, Frank R. Public knowledge of benefits of breast and prostate cancer screening in Europe. J Natl Cancer Inst 2009; 101: 1216–1220.

51. Elwyn G, O’Connor A, Stacey D et al. Developing a quality criteria framework for patient decision aids: online international Delphi consensus process. BMJ 2006;

333: 417.

52. Wolf AM, Wender RC, Etzioni RB et al. American Cancer Society guideline for the early detection of prostate cancer: update 2010. CA Cancer J Clin 2010; 60:

70–98.

53. Newschaffer CJ, Otani K, McDonald MK, Penberthy LT. Causes of death in elderly prostate cancer patients and in a comparison nonprostate cancer cohort.

J Natl Cancer Inst 2000; 92: 613–621.

54. Brawley OW. Prostate cancer screening; is this a teachable moment? J Natl Cancer Inst 2009; 101: 1295–1297.