2010 THE AUTHORS; BJU INTERNATIONAL Urological Oncology

2010 BJU INTERNATIONAL

HAEMATURIA IN BLADDER CANCER SCREENING AMONG MEN PESCH ET AL.

BJUI
BJU INTERNATIONAL

The role of haematuria in bladder cancer screening among men with former occupational exposure to aromatic amines
Beate Pesch1, Michael Nasterlack2, Friedhelm Eberle2, Nadine Bonberg1, Dirk Taeger1, Gabriele Leng3, Gerhard Feil4, Georg Johnen1, Katta Ickstadt5, Mattias Kluckert6, Harald Wellhuer6, Arnllf Stenzl4, Thomas Brning1 and the UroScreen Group*
Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr University Bochum (IPA), Bochum, 2BASF SE, Occupational Medicine and Health Protection, Ludwigshafen, 3Currenta GmbH & Co.OHG (formerly BAYER AG), Security Health Protection, Leverkusen, 4Department of Urology, EberhardKarls-University, Tbingen, 5Faculty of Statistics, TU Dortmund, Dortmund and 6Berufsgenossenschaft RCI, Fachreferat Arbeitsmedizin, Bereich Prvention, Heidelberg, Germany
Accepted for publication 31 August 2010
1

Study Type Diagnostic (validating cohort) Level of Evidence 1b OBJECTIVE To assess the positive predictive value (PPV) of microhaematuria (H) and gross haematuria (GH) in bladder cancer screening and the inuence of haematuria on tumour tests in a prospective study. PATIENTS AND METHODS From September 2003 to January 2010, 1323 men took part in an annual voluntary bladder cancer screening programme for chemical workers with former exposure to aromatic amines. In 5315 urine samples haematuria was determined with a dipstick, followed by a microscopic blood cell count in the sediment. Haematuria was categorized into traces, H and GH. Urinary leukocytes and other factors were investigated as potential predictors of haematuria using a generalized estimating equation model for repeated urinalysis. The risk of haematuria for positive tumour tests was analysed correspondingly. The bladder cancer risk was estimated for the highest degree of haematuria occurring during the study with Poisson regression.

Whats known on the subject? and What does the study add? Microscopic haematuria (H) is frequently detected in elderly adults. The American Urological Association recommends the follow-up of subjects with H on bladder cancer. Whereas gross haematuria is considered an important sign of the presence of bladder cancer, the disease-predictive value of H is less clear. No association of H with the development of bladder tumours in a prospective screening cohort of chemical workers was observed. The positive predictive value of H for bladder cancer was as low as 1.2%. Haematuria interfered with NMP22 but not with cytology and UroVysionTM test results.

RESULTS As of July 2010, 15 bladder tumours were detected in 14 participants. GH was found in four out of nine highgrade tumours and associated with a rate ratio of 3.82, 95% condence interval (CI) 0.5029.15 for the development of bladder lesions. The PPV of GH was 11.4%, but only 1.2% for H. H occurred in 18.8% of urine samples and was not associated with bladder cancer [rate ratio (RR) 0.72, 95% CI 0.114.78]. Abundant urinary leukocytes were associated with H [odds ratio (OR) 8.34, 95% CI 2.2630.69] and even stronger with GH (OR 22.25, 95% CI 6.4277.06). Haematuria and leukocytes inuenced NMP22 positivity (H: OR 1.63, 95% CI 1.06 2.51, abundant leukocytes: OR 8.90, 95% CI 1.5850.16), but not test results for urine cytology and UroVysionTM.

CONCLUSION While the PPV of H for bladder cancer was low, there was a strong inuence of haematuria and leukocytes on the proteinbased tumour test NMP22. Erythrocytes and leukocytes should be determined at least semi-quantitatively for the interpretation of positive NMP22 test results. In addition, a panel of tumour tests that includes methods not affected by the presence of erythrocytes or leukocytes such as cytology and UroVysionTM would improve bladder cancer screening.

KEYWORDS bladder cancer, haematuria, screening, NMP22, UroVysion, cytology

546

BJU INTERNATIONAL

2 0 11 T H E A U T H O R S

2 0 11 B J U I N T E R N A T I O N A L | 1 0 8 , 5 4 6 5 5 2 | doi:10.1111/j.1464-410X.2010.09971.x

HAEMATURIA IN BLADDER CANCER SCREENING AMONG MEN

FIG. 1. Work ow of the UroScreen study. Voided urine samples (annual screening rounds) Dipstick (fresh urine samples): Erythrocytes, hemoglobin, leukocytes Urine processing (centrifugation) Supernatant Microscopy: Erythrocytes, leukocytes If positive test result Recommendation of cystoscopy Cell sediment

measurements in this analysis. Participants with at least one positive test result for cytology, UroVysion or NMP22 received a written recommendation for urethrocystoscopy on a voluntary basis. A positive result for survivin was not considered in diagnostic follow-up. Haematuria was assessed with different methods in the two sub-cohorts. In the present study, data are presented from 1323 males from BASF where detailed urine-status information was available according to a common protocol until January 2010. The database comprised 5583 urine samples. As of January 2010, 15 tumours in 14 participants were detected. Reference pathology was performed for histological classication of the bladder lesions according to the WHO classication of 2004 [8]. URINE PROCESSING, URINE STATUS INFORMATION AND ASSESSMENT OF HAEMATURIA Voided urine samples were collected at the medical service centre of the plant. Urine status, including erythrocytes, haemoglobin (Hb) and leukocytes, was determined in fresh urine with Combur 10 test strips (Roche Diagnostics, Mannheim, Germany). Creatinine was determined using the enzymatic test CREA plus (Roche Diagnostics). After centrifugation at 500g, the cell sediment was analysed microscopically. Erythrocytes and leukocytes were documented semiquantitatively in at least 15 high-power elds. The cells were stabilized for cytology and the UroVysionTM test. The supernatant was stabilized using the Matritech NMP22 Urine Collection Kit (Matritech, Freiburg, Germany). NMP22 samples were stored at 20 C until shipping at 48 C to the laboratory in Tbingen. Haematuria was assessed as none, if neither erythrocytes nor Hb could be detected with any method (microscope or dipstick). Traces were assigned if up to ve erythrocytes were found with any of the two methods. H was dened for blood detection within the range between traces (<5 erythrocytes) and GH (>250 erythrocytes) with the dipstick or abundant erythrocytes in the sediment analysis. A similar semi-quantitative rating was performed for leukocytes with the categories none, traces, non-abundant and abundant.

Quantitative NMP22

UroVysion

Cytology

Survivin

INTRODUCTION Asymptomatic microscopic haematuria (microhaematuria, H) is frequently detected in elderly adults [1]. The American Urological Association recommended the follow-up of subjects with H, in particular on the development of bladder cancer [2,3]. Whereas gross haematuria (GH) is considered an important sign of the presence of bladder cancer, the disease-predictive value of H is less clear. Only a few prospective studies assessed H, and so far the positive predictive value (PPV) for subsequent bladder cancer is likely to be low as a result of other various sources of bleeding [4]. Screening for haematuria is feasible in at-risk populations together with tumour marker tests [5]. Haematuria can also be used to dene at-risk groups for bladder-cancer screening [6]. Dipstick testing for ltered haemoglobin or myoglobin in fresh urine and microscopic analysis of red blood cells in the urine sediment are commonly applied detection methods [7]. Whereas several methodological issues such as sample degradation have been evaluated, predictors of H and the interference of H with tumour tests are less well explored. In order to assess the role of H in bladder cancer screening and its association with tumour tests we analysed a subset of urine samples from the prospective study UroScreen. UroScreen is part of a surveillance programme of the German Social Accident Insurance (Deutsche Gesetzliche

Unfallversicherung, DGUV) among subjects with a previous occupational exposure to aromatic amines.

SUBJECTS AND METHODS STUDY POPULATION Eligible participants of the prospective surveillance cohort UroScreen are subjects who were formerly exposed to aromatic amines in two chemical plants. From 2003 to 2010, 1772 subjects were annually invited by DGUV to take part in a voluntary screening programme for bladder cancer offered by the occupational medical service in charge for each company. Figure 1 presents the work ow of UroScreen. The quantitative determination of the urinary nuclear matrix protein 22 (NMP22) and the UroVysionTM test were offered in addition to urine cytology. Further, the investigation of the performance of survivin as a nonapproved marker was part of the project but was not included in this analysis. A questionnaire was applied to document smoking habits and other relevant diseases. Approval was given by the ethics committee of the Eberhard-Karls-University of Tbingen (No. 1/2003V). All participants gave written informed consent. Participants with a urinary tract infection or insufcient cell yield in the urine sample were asked to repeat the investigation. Because of the low compliance for a second investigation we did not include repeated

2 0 11 T H E A U T H O R S

BJU INTERNATIONAL

2 0 11 B J U I N T E R N A T I O N A L

547

P E S C H ET AL.

TABLE 1 Characteristics of male participants of UroScreen with former occupational exposure to aromatic amines Haematuria* H 408 (30.84%) 63 (3291) Inammation1 Non-abundant 224 (16.93%) 67 (4290)

Characteristics N Age in 2010 (years) Median (range) Age at diagnosis Smoking status at baseline Never Former Current Former bladder cancer

All 1323 60 (2891)

Cases 14 (1.06%) 66 (4481) 64 (3876) (28.57%) 7 (50.00%) 3 (21.43%) 3 (21.43%)

GH 85 (6.42%) 62 (3881)

Abundant 11 (0.83%) 72 (5281 )

447 (33.89%) 491 (37.23%) 381 (28.89%) 21 (1.59%)

143 (35.05%) 148 (36.27%) 117 (28.68%) 9 (2.21%)

26 (30.59%) 29 (34.12%) 30 (35.29%) 3 (3.53%)

64 (28.70%) 72 (32.29%) 87 (39.01%) 6 (26.79%)

5 (45.45%) 2 (18.18%) 4 (36.36%) 0 (0%)

*H, microhaematuria, or inammation non-abundant if 5 250 cells (erythrocytes respectively leukocytes) could be detected with a dipstick or microscopically in the sediment; GH, gross hematuria, or abundant leukocytes if 250 cells could be detected.

TUMOUR MARKER TESTS NMP22 was quantitatively determined in the supernatant with the NMP22 ELISA (Matritech Inc, Newton, MA), which employed two monoclonal antibodies (MAb302-18 and MAb302-22). The analysis was performed according to the manufacturers protocol. The concentration of NMP22 was calculated proportionally to the colour extinction from a standard curve, and 10 units/mL was used as cut-off for positive results. Chromosomal instability in urothelial cells was assessed using the UroVysionTM Bladder Cancer Kit (Abbott Laboratories, Abbott Park, IL, USA). This uorescence in situ hybridization assay was applied according to the manufacturers protocol. DNA was denatured for 3 min at 76 C followed by a hybridization step with a mixture of uorochrome-marked DNA probes (CEP 3, CEP 7, CEP 17 and LSI 9p21) at 39 C for 20 h using a HYBriteTM or ThermoBriteTM (Abbott Molecular, Des Plaines, IL, USA). For each sample, at least 25 morphologically suspicious cells were evaluated. The test was considered positive if at least four nuclei had three signals of two or three chromosomes (3, 7 and 17) or at least 12 nuclei showed one signal for the 9p21 locus. Urinary cytology was performed by successive staining of cytospinned slides using Papanicolaous solutions 1a Harris haematoxylin solution, 2a Orange G solution (OG 6) and 3b polychromatic solution EA 50 (Merck, Darmstadt, Germany) [9]. Samples

were evaluated at 400 magnication. Results were classied as positive, suspicious or negative [10].

for abundant urinary leukocytes. The proportion of never smokers was 33.9% in the cohort, 28.6% in cases, 35.1% in participants with H and 30.6% in men with any GH. Table 2 shows the prevalence of haematuria in cases by histology, together with other information on the last urine sample before diagnosis. GH was detected in four out of nine cases with high-grade bladder cancer while eighth out of all 14 tumours were free of blood or contained only traces. H was present in three of the tumours. However, the interval between screening and diagnosis varied considerably and exceeded 12 months in four cases. Leukocytes could be found at least in traces in most samples except two with low creatinine concentrations. Potential predictors of haematuria are presented in Table 3. Leukocytes were strongly associated with both H (abundant: OR 8.34, 95% CI 2.2630.69) and GH (abundant: OR 22.25, 95% CI 6.4277.06). H was less frequently observed in urines with creatinine <0.5 g/L (OR 0.43, 95% CI 0.340.54). Current smokers had marginally more haematuria (H: OR 1.24, 95% CI 0.971.58; GH: OR 1.72, 95% CI 0.953.12). The risk of haematuria for bladder cancer is depicted in Table 4. Ever H was associated with a RR of 0.72 (95% CI 0.114.78) and ever GH with a RR of 3.82 (95% CI 0.5029.15). The cancer-predictive value of H was low (1.24%) and slightly higher for GH (11.43%) (Table 5).

STATISTICAL ANALYSIS The distribution of data is presented by median and inter-quartile range (IQR). Odds ratios (ORs) and 95% CIs were estimated for the predictors of haematuria and of positive tumour tests with generalized estimating equation (GEE) models. The bladder cancer risk of haematuria was estimated for the highest category observed until diagnosis. The bladder cancer risk of haematuria was estimated as the rate ratio (RR) using Poisson regression, adjusted for potential confounders. All calculations were performed with SAS/STAT and SAS/IML software, version 9.2 (SAS Institute Inc., Cary, NC, USA). RESULTS Between September 2003 and January 2010, 1323 men took part at least once in the extended surveillance programme UroScreen. So far, 15 bladder tumours have been detected in 14 participants. Table 1 depicts demographic and other characteristics of the screening participants at baseline, of cases with bladder tumours, and of subjects with at least one occurrence of erythrocytes or leukocytes during this period. The median age of cases with bladder cancer was 64 years. The median age of the cohort was 60 years in 2010, for participants with H 63 years and 62 years for GH. The median age was 72 years

548

BJU INTERNATIONAL

2 0 11 T H E A U T H O R S

2 0 11 B J U I N T E R N A T I O N A L

HAEMATURIA IN BLADDER CANCER SCREENING AMONG MEN

Table 6 shows the association of haematuria with tumour tests. Both H and GH inuence NMP22 positivity (OR 1.63, 95% CI 1.062.51 and OR 3.94, 95% CI 1.938.03, respectively). Urinary leukocytes had a strong effect on the NMP22 test results (abundant: OR 8.90, 95%

CI 1.5850.16). Erythrocytes and leukocytes did not show an inuence on the cytology and UroVysionTM test. Incident bladder cancer was associated with test positivity (NMP22: OR 4.25, 95% CI 1.6910.68, UroVysion: OR 11.14, 95% CI 3.1839.04, cytology: OR 32.23,

TABLE 2 Histopathological ndings, haematuria and inammation in the last screening round before diagnosis of 15 bladder lesions in 14 cases in the prospective study UroScreen Histopathological nding Papilloma High grade Papilloma Low grade Low grade Low grade Low grade High grade High grade High grade High grade High grade High grade High grade High grade Months before diagnosis 10 24 2 10 2 2 3 7 26 1 14 0 2 18 2 Creatinine (g/L) 0.35 0.87 0.77 0.23 1.59 0.44 0.74 1.28 1.01 0.24 2.92 1.53 0.32 2.93 0.72

95% CI 10.02103.65). Smoking and former bladder cancer were associated with positivity of cytology and the UroVysionTM test, but not with NMP22 where current smokers were less frequently associated with positive results (OR 0.51, 95% CI 0.280.90). The UroVysionTM test was more frequently positive in urines with creatinine <0.5 g/L (OR 2.73, 95% CI 1.634.57).

DISCUSSION
Erythrocytes* H GH None H Traces None H GH None None None GH GH Traces None Leukocytes* Traces Traces Traces Traces Traces None Traces Traces Traces None Traces Non-abundant Non-abundant Traces Traces

Case 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14

The role of haematuria in bladder cancer screening was evaluated in a prospective cohort of 1323 chemical workers between 2003 and 2010. Results from dipstick and subsequent sediment analysis were combined to classify urine samples as free of blood or with traces of H or GH. A similar assessment was performed for leukocytes which were a strong predictor of haematuria. Our results do not indicate an association of H with bladder tumours. The PPV of H for bladder cancer was as low as 1.2%. Haematuria and leukocytes interfered with NMP22 but not with cytology and UroVysionTM test results. Bladder cancer screening in the general population is not yet an accepted standard as a result of the low incidence of the disease, raising concerns about cost effectiveness [11]. However, the availability of tumour tests and

*H, microhaematuria, or inammation non-abundant if 5 250 cells (erythrocytes respectively leukocytes) could be detected with a dipstick or microscopically in the sediment; GH, gross hematuria, or abundant leukocytes if 250 cells could be detected.

TABLE 3 Potential predictors of microhaematuria in 5067 urine samples from 1312 male participants (samples with gross haematuria not included) and gross haematuria in 5181 urine samples from 1323 male participants of the UroScreen cohort Microhaematuria N (samples) 1609 3152 298 8 834 3947 286 2898 2169 1789 2084 1194 4948 119 4998 69 5035 32 Gross haematuria N (samples) OR 1628 1 3220 1.53 322 4.65 11 22.25 843 0.61 4042 1 296 1.47 2967 1 2214 0.81 1819 1 2129 1.36 1233 1.72 505 1 125 2.65 5106 1 75 4.19 514 1 37 6.62

Variable at sampling Leukocytes

Creatinine

Age Smoking status

Former urogenital tumour except bladder Former bladder cancer Incident bladder cancer

Category None Traces Non-abundant Abundant <0.5 g/L 0.52.5 g/L >2.5 g/L <60 years 60 years Never Former Current None Yes None Yes None Yes

OR 1 1.10 2.28 8.34 0.43 1 1.06 1 1.18 1 0.95 1.24 1 1.33 1 1.95 1 1.24

95% CI 0.971.26 1.762.94 2.2630.69 0.340.54 0.851.32 0.961.45 0.771.18 0.971.59 0.832.12 0.973.93 0.443.45

95% CI 0.982.38 2.448.84 6.4277.06 0.321.15 0.772.78 0.491.33 0.812.30 0.953.12 0.917.76 1.919.20 2.4418.01

2 0 11 T H E A U T H O R S

BJU INTERNATIONAL

2 0 11 B J U I N T E R N A T I O N A L

549

P E S C H ET AL.

treatments improving survival fulll general requirements for cancer screening. Haematuria and exposure to aromatic amines have been considered for establishing at-risk populations. So far, few screening cohort studies have been conducted (e.g. [4,5,12,13]). Because chemical workers with exposure to aromatic amines were at excess risk for bladder cancer [14,15] DGUV offers an annual surveillance programme with investigation of the urine status and cytology. UroScreen became part of this programme and included the determination of NMP22 and UroVysionTM. A major advantage of UroScreen is the prospective design with repeated investigations because performance measures such as PPV have only be estimated in a longitudinal design [16,17]. Several biases of screening studies have been described [17]. Selection bias is of minor concern because the participation rate was about 90%. All participants of UroScreen were recruited from an on-going surveillance programme. Therefore, prevalent cases were less likely. However, the compliance to participate regularly in the annual screening rounds was low. A challenge in screening for bladder cancer is the low incidence and therefore the need for dening at-risk populations [18]. Although we enrolled workers with former exposure to aromatic amines, we observed only 15 bladder tumours in 14 participants. The average age of the cohort was several years younger than the age at diagnosis (as presented in the German cancer database at Robert Koch Institute (http://www.rki.de). Because the surveillance programme was established for all workers with occupational exposure, exclusion criteria such as young age, former or newly detected bladder cancer or never smoking were not applied. While historical exposure resulted in very high risks for bladder cancer in chemical workers [15], technological progress and the ban of certain aromatic amines in the dyestuff production may have reduced the risks as observed in this and another study in Germany [19]. Other screening cohorts also face a small number of cases, for example in British men [12]. In a US cohort many participants had already undergone a recent urinalysis where prevalent cases are less likely to be detected [13]. This holds also for UroScreen. Initial efforts at bladder cancer screening have focused on the detection of haematuria

TABLE 4 Risk of haematuria and other factors on newly detected bladder lesions in male participants of the UroScreen cohort Potential risk factor Haematuria until diagnosis Never Traces Microhaematuria Ever gross haematuria Leukocytes until diagnosis Never or traces Non-abundant or ever abundant Age at entry <60 years 60 years Smoking status Never Ever Former bladder cancer None Yes Person-Years 436.05 2444.69 1557.63 310.11 3809.16 939.33 2930.64 1817.85 1411.15 3337.33 4676.77 71.72 Cases 2 2 6 4 11 3 7 7 4 10 11 3 Rate Ratio 1.0 0.15 0.72 3.82 1.0 0.62 1.0 1.66 1.0 0.84 1.0 31.66 95% CI

0.021.42 0.114.78 0.5029.15

0.132.86

0.475.85

0.213.43

6.18182.12

TABLE 5 Evaluation of the cancer-predictive value of haematuria for bladder lesions in the UroScreen cohort All lesions N = 15 3 8 1039 239 1.24% 99.24% 4 11 1278 31 11.43% 99.15% Papilloma N=2 1 1 Low grade N=4 2 2 High grade N=9 0 5

Microhaematuria

Gross haematuria

Result True positive False negative True negative False positive Positive predictive value Negative predictive value True positive False negative True negative False positive Positive predictive value Negative predictive value

0.42% 99.90% 0 2

0.83% 99.81% 0 4

0% 99.52% 4 5

0% 99.84%

0% 99.69%

11.43% 99.61%

[1]. Especially high-grade tumors can become muscle invading, which may lead to haematuria. GH is therefore considered an important sign of bladder cancer. A classication into H and GH is important because four out of nine UroScreen cases with high-grade bladder cancer showed abundant erythrocytes until diagnosis. While visible blood is a relatively rare symptom, asymptomatic occult urinary bleeding is a common nding. H is an intermittent symptom in urine samples, where the probability of detection increases with the number of samples per subject. Various

factors inuence the detection of blood in urine and thus the measures of prevalence or risk. Urinary leukocytes turned out as major predictor of both H and, even stronger, GH. While dipsticks can also detect Hb, the analysis of the sediment allows only counting intact cells. A detailed technical assessment of the diagnostic tests has been given by Rodgers et al. [7]. Dipsticks are point-of-care tests where readouts are possible that provide an automated calling of semi-quantitative results on the content of erythrocytes and leukocytes. Microscopic evaluation needs expertise and can be prone to observer bias.

550

BJU INTERNATIONAL

2 0 11 T H E A U T H O R S

2 0 11 B J U I N T E R N A T I O N A L

HAEMATURIA IN BLADDER CANCER SCREENING AMONG MEN

TABLE 6 Potential inuences of haematuria and other factors on positivity of tumour tests in urine samples from male participants of the UroScreen cohort NMP22 N = 5167 Category None Traces Microhaematuria Gross haematuria None Traces Non-abundant Abundant <0.5 g/L 0.52.5 g/L >2.5 g/L None Yes None Yes Never Former Current <60 years 60 years Cytology N = 4911 n (npos) OR (95% CI) 3818 (14) 1 1004 (7) 4522 (20) 300 (1) 1.20 (0.473.04) 1 1.19 (0.245.82) UroVysion N = 4822 n (npos) OR (95% CI) 2359 (36) 1 1459 (20) 1.09 (0.621.92) 894 (12) 0.96 (0.501.84) 110 (2) 0.81 (0.232.87) 4522 (65) 1 300 (5) 760 (25) 3790 (44) 271 (1) 4752 (64) 70 (6) 4785 (64) 37 (6) 1679 (15) 1990 (35) 1153 (20) 2741 (39) 2081 (31) 1.17 (0.453.03) 2.73 (1.634.57) 1 0.28 (0.032.40) 1 4.33 (1.3513.90) 1 11.14 (3.1839.04) 1 2.00 (1.053.81) 2.14 (1.064.33) 1 1.05 (0.611.80)

Variable at sampling Haematuria*

Leukocytes

Creatinine

Prevalent bladder Cancer Incident bladder Cancer Smoking Status Age

n (npos) 2529 (48) 1555 (43) 969 (40) 114 (14) 1623 (29) 3212 (78) 321 (35) 11 (3) 840 (12) 4032 (117) 295 (16) 5094 (141) 73 (4) 5130 (140) 37 (5) 1814 (53) 2124 (71) 1229 (21) 2959 (60) 2208 (85)

OR (95% CI) 1 1.15 (0.741.78) 1.63 (1.062.51) 3.94 (1.938.03) 1 1.06 (0.691.65) 4.32 (2.367.90) 8.90 (1.5850.16) 0.55 (0.291.06) 1 1.86 (1.063.27) 1 1.09 (0.393.01) 1 4.25 (1.6910.68) 1 1.13 (0.751.69) 0.51 (0.280.90) 1 1.46 (0.982.17)

4752 (17) 70 (4) 4785 (16) 37 (5) 1679 (2) 1990 (14) 1153 (5) 2741 (10) 2081 (11)

1 9.35 (2.0243.42) 1 32.23 (10.02103.7) 1 6.90 (1.6029.85) 4.64 (0.8425.83) 1 1.38 (0.563.44)

*Risk of haematuria for positivity of cytology: none or traces as a reference group vs the combined group of microhaematuria and gross haematuria; Risk of leukocytes for positivity of cytology and UroVysionTM test: none or traces as the reference group vs the combined group of non-abundant and abundant leukocytes.

Automated investigations are possible but not commonly applied in practice. Further, microscopy was evaluated as generating more false-negative results, and dipstick testing might generate more false-positive results. Here we combined results from both methods. Only 40% of our samples were free of erythrocytes or Hb, and 2% demonstrated GH. Misclassication of haematuria is possible by low cell content in urines with <0.50 g/L creatinine. This cut-off was chosen according to the 10th percentile of creatinine in elderly German men [20]. Low creatinine might serve as a potential proxy for retention time and uid consumption. Current guidelines for US urologists recommend cystoscopic evaluation of elderly adults with H. Former and incident bladder cancer were predictors of GH but not of H. Further, H was not associated with a bladder cancer risk. This was also reected in the low PPV of 1.2% for H that increased up to 11.4% for GH, in particular for high-grade tumours. A low PPV has also been reported for NMP22 BladderChek [13]. These low PPVs

should be critically taken into account in clinical routine although PPVs can be improved in subjects at excess risk for bladder cancer [21]. However, dipsticks are important tools to determine urinary blood, leukocytes and creatinine for tumour tests because these parameters inuenced the results. The interference of erythrocytes and leukocytes with NMP22 positivity conrmed observations from an experimental approach by adding blood to urine samples and from subjects with a urinary tract infection [22]. Leukocytes were the strongest predictor of NMP22 positivity. These results need consensus on diagnostic algorithms before cystoscopy should be recommended. We recommended a second examination after treatment of an infection but the compliance was low. Further, an experience of the invasiveness of cystoscopy reduced the compliance for subsequent cystoscopies. However, neither erythrocytes nor leukocytes showed an inuence on the detection of aberrant urothelial cells with cytology or the UroVysionTM test. An evaluation of the NMP22

or UroVysion results will be subject to another analysis. In conclusion, the UroScreen study revealed for H a low PPV and lack of an obvious association with bladder cancer. However, there was a strong inuence of haematuria and leukocytes on the protein-based tumour test NMP22. Erythrocytes and leukocytes should be determined at least semiquantitatively for the interpretation of positive NMP22 test results. In addition, a panel of tumour tests that includes methods not affected by the presence of erythrocytes or leukocytes such as cytology and UroVysionTM would improve bladder cancer screening.

ACKNOWLEDGEMENTS The study received grants from German Social Accident Insurance (DGUV), Sankt Augustin, Germany. Abbott GmbH & Co. KG, Wiesbaden, Germany, supplied the UroVysionTM kits at no cost.

2 0 11 T H E A U T H O R S

BJU INTERNATIONAL

2 0 11 B J U I N T E R N A T I O N A L

551

P E S C H ET AL.

CONFLICT OF INTEREST None declared.

REFERENCES 1 Messing EM, Vaillancourt A. Hematuria screening for bladder cancer. J Occup Med 1990; 32: 83845 Grossfeld GD, Litwin MS, Wolf JS et al. Evaluation of asymptomatic microscopic hematuria in adults: the American Urological Association best practice policypart I: denition, detection, prevalence, and etiology. Urology 2001; 57: 599603 Grossfeld GD, Litwin MS, Wolf JS Jr et al. Evaluation of asymptomatic microscopic hematuria in adults: the American Urological Association best practice policypart II: patient evaluation, cytology, voided markers, imaging, cystoscopy, nephrology evaluation, and follow-up. Urology 2001; 57: 604 10 Messing EM, Madeb R, Young T et al. Long-term outcome of hematuria home screening for bladder cancer in men. Cancer 2006; 107: 21739 Steiner H, Bergmeister M, Verdorfer I et al. Early results of bladder-cancer screening in a high-risk population of heavy smokers. BJU Int 2008; 102: 291 6 Grossman HB, Messing E, Soloway M et al. Detection of bladder cancer using a point-of-care proteomic assay. JAMA 2005; 293: 8106 Rodgers M, Nixon J, Hempel S et al. Diagnostic tests and algorithms used in the investigation of haematuria: systematic reviews and economic evaluation. Health Technol Assess 2006; 10: iii259 Montironi R, Lopez-Beltran A. The 2004 WHO classication of bladder tumors: a summary and commentary. Int J Surg Pathol 2005; 13: 14353 Papanikolaou GN, Marshall VF. Urine sediment smears as a diagnostic procedure in cancers of the urinary tract. Science 1945; 101: 51921

10 Rathert P, Roth S eds. Urinzytologie. Praxis Und Atlas. Berlin: Springer, 1991 11 Lotan Y, Svatek RS, Malats N. Screening for bladder cancer: a perspective. World J Urol 2008; 26: 138 12 Britton JP, Dowell AC, Whelan P. Dipstick haematuria and bladder cancer in men over 60: results of a community study. BMJ 1989; 299: 10102 13 Lotan Y, Elias K, Svatek RS et al. Bladder cancer screening in a high risk asymptomatic population using a point of care urine based protein tumor marker. J Urol 2009; 182: 527 14 Vineis P, Pirastu R. Aromatic amines and cancer. Cancer Causes Control 1997; 8: 34655 15 Pira E, Piolatto G, Negri E et al. Bladder cancer mortality of workers exposed to aromatic amines: a 58-year follow-up. J Natl Cancer Inst 2010; 102: 10969 16 Brenner DE, Normolle DP. Biomarkers for cancer risk, early detection, and prognosis: the validation conundrum. Cancer Epidemiol Biomarkers Prev 2007; 16: 191820 17 Baker SG, Kramer BS, McIntosh M, Patterson BH, Shyr Y, Skates S. Evaluating markers for the early detection of cancer: overview of study designs and methods. Clin Trials 2006; 3: 4356 18 Lotan Y, Svatek RS, Sagalowsky AI. Should we screen for bladder cancer in a high-risk population?: a cost per life-year saved analysis. Cancer 2006; 107: 98290 19 Pesch B, Haerting J, Ranft U, Klimpel A, Oelschlgel B, Schill W. Occupational risk factors for urothelial carcinoma: agent-specic results from a case-control study in Germany. MURC Study Group. Multicenter urothelial and renal cancer. Int J Epidemiol 2000; 29: 23847 20 Becker K, Schulz C, Kaus S, Seiwert M, Seifert B. German Environmental Survey 1998 (GerES III). Environmental pollutants in urine of the German population. Int J Hyg Environ Health 2003; 206: 1524 21 Lotan Y, Capitanio U, Shariat SF, Hutterer GC, Karakiewicz PI. Impact of clinical factors, including a point-of-care nuclear matrix protein-22 assay and cytology, on bladder cancer detection. BJU Int 2009; 103: 136874

22 Ats N, Ekici S, Oge OO, Ergen A, Haselik G, Ozen H. False-positive results of the NMP22 test due to hematuria. J Urol 2002; 167: 5558 Correspondence: Beate Pesch, Institut fr Prvention und Arbeitsmedizin der Deutschen Gesetzlichen Unfallversicherung (IPA), Brklede-la-Camp-Platz 1, 44789 Bochum, Germany. e-mail: pesch@ipa-dguv.de Abbreviations: PPV, positive predictive value; H, microhaematuria; GH, gross haematuria; RR, rate ratio; OR, odds ratio; GEE, generalized estimating equation; IQR, interquartile range.

APPENDIX *UroScreen Group: BASF SE, Occupational Medicine and Health Protection, Ludwigshafen: Bernd Scheuermann, Friedhelm Eberle, Thomas Mayer, Michael Nasterlack Berufsgenossenschaft RCI, Fachreferat Arbeitsmedizin, Bereich Prvention, Heidelberg: Harald Wellhuer, Matthias Kluckert; Organisationsdienst fr nachgehende Untersuchungen (ODIN), Heidelberg: Reinhard Detzner Institute for Prevention and Occupational Medicine of German Social Accident Insurance, Institute of Ruhr University Bochum (IPA), Bochum: Beate Pesch, Dirk Taeger, Nadine Bonberg, Heike Bontrup, Georg Johnen, Judith Delbanco, Evelyn Heinze, Thomas Brning Currenta GmbH & Co.OHG, Security Health Protection, Leverkusen: Martin Pelster, Gabriele Leng Department of Urology, Eberhard-KarlsUniversity, Tbingen: Gerhard Feil, KarlDietrich Sievert, Sverine Huber, Margarete Geiger, Erika Senger, Valentina Gerber, Andrea Hohneder, Gundi Beger, Ursula Kuehs, Jrg Hennenlotter, Arnulf Stenzl

552

BJU INTERNATIONAL

2 0 11 T H E A U T H O R S

2 0 11 B J U I N T E R N A T I O N A L

Copyright of BJU International is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.