Int Arch Occup Environ Health (2005) 78: 475–485 DOI 10.1007/s00420-005-0611-0

O R I GI N A L A R T IC L E

Jens-Uwe Voss Æ Markus Roller Æ Elke Brinkmann Inge Mangelsdorf

Nephrotoxicity of organic solvents: biomarkers for early detection

Received: 7 June 2004 / Accepted: 24 January 2005 / Published online: 13 May 2005 Ó Springer-Verlag 2005

Abstract Objectives: Evidence for a relationship between chronic kidney diseases or progression of already existing diseases (glomerulonephritides) and occupational solvent exposure has been found in case reports, in casecontrol studies and also in cross-sectional studies. An analysis of the available literature was performed with respect to markers measured in cross-sectional studies that might be useful for an early detection of solventinduced effects on the kidney. Methods: The relevant cross-sectional studies were evaluated and the following markers were analyzed with respect to their suitability as biomarker for renal damage: total protein, albumin, transferrin, IgG, b2-microglobulin, retinol-binding protein, N-acetyl-b-D-glucosaminidase, alanine aminopeptidase, b-galactosidase, b-glucuronidase, leucin aminopeptidase, alkaline phosphatase, lysozyme, Tamm-Horsfall protein and laminin fragments in urine as well as E-selectin, laminin and anti-laminin antibodies and anti-glomerular basement membrane antibodies in serum. Results: An increased albumin excretion was observed more frequently in groups of workers exposed to various solvents (like toluene, styrene, aliphatic/aromatic hydrocarbon mixtures, tetrachloroethene, mixtures of chlorinated hydrocarbons) than in controls. No clear pattern emerged for the other markers. Conclusions: The determination of albumin excretion in the urine appears to be a useful parameter for monitoring solvent-exposed workers.

J.-U. Voss Æ M. Roller Æ I. Mangelsdorf (&) Fraunhofer Institute of Toxicology and Experimental Medicine (ITEM), Nikolai-Fuchs-Str. 1, 30625 Hannover, Federal Republic of Germany E-mail: [email protected] Fax: +49-511-5350335 E. Brinkmann German Federal Institute for Occupational Safety and Health, Bundesanstalt fu¨r Arbeitsschutz und Arbeitsmedizin (BAUA), Germany

Keywords Biomarkers Æ Albumin Æ Nephrotoxicity Æ Occupational exposure Æ Organic solvents Abbreviations AAP: Alanine aminopeptidase Æ 1,3-DCP: Dichloropropene Æ b-Gal: b-Galactosidase Æ b-Glu: b-Glucuronidase Æ b2M: b2-Microglobulin Æ NAG: N-acetyl-b-D-glucosaminidase Æ OR: Odds ratio Æ RBP: Retinol binding protein Æ TWA: Time weighted average

Introduction In the USA approximately 4 million workers are exposed to chemicals, which have shown nephrotoxic effects in animal experiments. These chemicals include solvents belonging to a structurally heterogeneous group of chemicals with a wide spread use in a variety of products and at different work places. Renal damage after acute exposure to solvents (turpentine) has been described in case reports as early as more than hundred years ago (Rheinhard 1887; Glaeser 1892). More recent case reports also describe a relationship between solvent exposure and renal damage, primarily Goodpasture’s syndrome or other glomerulonephritides (Beirne 1972; Whitworth et al. 1974). Solvents suspected of causing nephropathy include tetrahydrofurane (Albrecht et al. 1987), toluene (Bosch et al. 1988), Stoddard solvent (Daniell et al. 1988), mineral turpentine (white spirit) (D’Apice et al. 1978), haloalkenes and ketones (Ehrenreich et al. 1974; Ehrenreich 1977), and haloalkanes (Keogh et al. 1984; Nathan and Toseland 1979) including tetrachloromethane (Carlier et al. 1980). The case reports on glomerulonephritides have prompted numerous case-control studies, which have been reviewed (Churchill et al. 1983; Hotz 1994; Phillips et al. 1988; Ravnskov 2000; Voss et al. 2003). No specific solvents could be identified as risk factors, however, a

476

positive relationship between solvent exposure and nonsystemic glomerulonephritides has been found in 20 out of 25 studies (also including the positive study of Huber et al. (2000), which was not yet included in the reviews). A recently published meta-analysis (Ravnskov 2000) found a significantly albeit weakly increased odds ratio (OR) of 1.6 (95% CI: 1.2–2.0) for all studies combined. The odds ratios were higher when studies with more than 1% drop out due to death were excluded (OR: 3.7; 2.9–4.6). Furthermore, the odds ratio correlated with the degree of renal failure and was highest in studies that included patients with end-stage renal disease (OR: 5.9; 3.8–9.3). Unlike case-control studies, cohort mortality studies failed to show a relationship between solvent exposure and kidney disease (Hotz 1994; Phillips et al. 1988; Delzell et al. 1988). Also a meta-analysis of 55 mortality studies on workers exposed to organic solvents did not show an increased risk (Chen and Seaton 1996). It is likely that the study power was not sufficient for analyzing rare diseases such as glomerulonephritis (Churchill et al. 1983). Furthermore, the diagnostic categories applied in many of these studies were rather broad (‘‘nephritis and nephrosis’’ or ‘‘genitourinary diseases’’), including other diseases in addition to kidney diseases, which might thus obscure a possible relationship. Thus, the results of the cohort mortality studies do not contradict the results of the case-control studies. Unfortunately, cohort morbidity studies on glomerulonephritis, which would be more suitable, are not available. Cross-sectional studies in solvent-exposed workers have led to a complicated result (Hotz 1994; Voss et al. 2003; Lauwerys and Bernard 1989; Hotz and Boillat 1989; Mutti 1996). Renal alterations, indicating effects at the tubular or the glomerular site, were seen in some studies, whereas such alterations were not found in others. A comparison of the results from the great number of studies is impossible because the study groups differed with respect to working site, type of solvents used, concentration, frequency and duration of exposure as well as parameters used to monitor renal alterations. However, in contrast to the exposed groups, signs of renal alterations were seen in control groups only in few cases. Therefore, the results of the studies are compatible with the hypothesis of a nephrotoxic effect of organic solvents. Due to the large reserve capacity of the kidney, kidney diseases are diagnosed at a rather late stage, which can be irreversible and end up in end-stage renal failure. It is therefore important to detect kidney defects as early as possible to prevent progression of the disease. The aim of a research project funded by the German Federal Institute of Occupational Health and Safety was to find out, which biomarker might be useful for an early diagnosis of solvent-induced kidney disease at the workplace (Voss et al. 2003). The most important results will be presented in the following.

c Fig. 1 a Comparison of albumin concentration in urine (mg/g creatinine) of solvent exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part I, continued in Fig. 1b and c. The data points represent mean or median values, bars indicate the range of individual values. Note logarithmical scale. Asterisks indicate statistically significant differences according to the authors of the studies. The sequence of studies within the blocks of substances is arranged according to exposure concentration (in ascending order). Abbreviations of study authors: Ho90 Hotz et al. (1990), Fr83 Franchini et al. (1983), St98 Stengel et al. (1998), Ng90 Ng et al. (1990), Vi87a Viau et al. (1987a), VH98 Verplanke and Herber (1998), two groups; Vy89 Vyskocil et al. (1989), Mu81 Mutti et al. (1981), Vi87b Viau et al. (1987b), Ge97 Ge´rin et al. (1997). cf. Tables 1 and 4.1. b Comparison of albumin concentration in urine (mg/g creatinine) of solvent-exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part II, continued from Fig. 1a. The sequence of studies with tetrachloroethene exposure is arranged according to TWA exposure concentration (in ascending order). Abbreviations of study authors: Vy91 Vyskocil et al. (1991), Ro93 Rocskay et al. (1993), Gr92 Gruener (1992), Ve99 Verplanke et al. (1999), Fr83 Franchini et al. (1983), Mu92 Mutti et al. (1992), SR91 Solet and Robins (1991), La83 Lauwerys et al. (1983), Vy90 Vyskocil et al. (1990), Bo93 Boogard et al. (1993); Boogard and Caubo (1994). cf. Tables 1 and 4.1. c Comparison of albumin concentration in urine (mg/l) of solvent exposed workers (full square symbols) and of nonexposed controls (open square symbols)—part III, continued from Fig. 1a and 1. Abbreviations of study authors: As81a Askergren et al. (1981a), Br91 Brouwer et al. (1991), Kr85 Krusell et al. (1985) and Ho90 Hotz et al. (1990). cf. Tables 1 and 4.1

Method All relevant available cross-sectional studies on kidney disease from solvent exposure at the workplace were assessed. Relevant studies were identified from a bibliographic search in the MEDLINE-database and TOXLINE-database. For a comprehensive overview, keywords for solvent exposure (solvent and hydrocarbon) were combined with various keywords regarding kidney disease and function (renal disease, kidney disease, glomerulonephritis, nephrotoxicity, nephropathy, protein excretion and biomarker). Additionally, older publications were identified from several reviews, e.g. Bernard and Lauwerys (1991), Hotz (1994) and Mutti (1996). The analysis included, if available, numbers and age of workers exposed and of controls, duration of exposure (mean, range), type of work, compounds involved, and exposure concentration (TWA, range). The following biomarkers were assessed: total protein, albumin, transferrin, N-acetyl-b-D-glucosaminidase (NAG), IgG, b2microglobulin (b2M), retinol binding protein (RBP), alanine aminopeptidase, b-galactosidase, b-glucuronidase, leucin aminopeptidase, alkaline phosphatase, lysozyme, Tamm-Horsfall protein, and laminine fragments. Also, studies providing data for E-selectin, laminin and anti-laminin antibodies and anti-glomerular basement membrane antibodies in serum were evaluated. The database for most of these biomarkers was rather small, preventing detailed evaluations and further conclusions. In the following, mainly the results of albumin,

477

b2-M and NAG are presented. These biomarkers are measured frequently and represent important functions of the kidney, i.e. enhanced albumin excretion indicates deficiencies in glomerular filtration of macromolecules,

b2M indicates defects in tubular reabsorption, and NAG general tubular toxicity. Details of the studies and the numbers underlying Figs. 1, 2 and 3 are available in Voss et al. (2003).

478 Table 1 Explorative statistical analysis of the frequency of ‘‘high’’ values for variables of urinary excretion in cross-sectional studies with solvent exposure Variable

Albumin

b2M d

NAG

Cut-off limit

37 mg/g creatininec, or 19 mg/l urinec 100 mg/g creatinine, or 100 mg/l urinee 200 lg/g creatininec, or 300 lg/l urinec 5 U/g creatininec 20 U/g creatininee

Frequency of groups containing individuals with ‘‘high’’ valuesa

p-Valueb (Fisher’s exact test, one-tailed)

Exposed

Non-exposed

14/14 (100%)

7/12 (58%)

0.01

9/14 (64%)

2/12 (17%)

0.02

7/9 (78%)

5/8 (63%)

0.4

7/9 (78%) 3/9 (33%)

4/7 (57%) 0/7 (0%)

0.4 0.2

b

‘‘High’’ means individual values greater than a defined cut-off limit, whereby the cut-off limits were chosen under consideration of the normal ranges given by Boege (1998) a Number of studies (groups within a study) in which the upper limit of the range of individual values is greater than the cut-off limit, related to the number of studies for which the range of individual values is given in the paper

Comparison exposed versus non-exposed Upper limit of the normal range of Boege (1998) The studies by Brogren et al. (1986) and Rasmussen et al. (1993) were not included in this analysis, because these authors used a different method for determining NAG activity e Higher cut-off limits set arbitrarily for a less sensitive comparison of frequency of ‘‘high’’ values

The question of statistical significance of a possible difference in urinary excretion of biomarkers in solventexposed workers and non-exposed controls was investigated using two different approaches. In a first step the statistical significance as given by the authors of the individual studies was assessed. The second analysis was based on the observation that the differences between mean or median values in individual studies may appear relatively small but maximum values of individuals are partially considerably higher in exposed workers than in the corresponding group of non-exposed controls. Therefore, the frequency of ‘‘high’’ individual values was examined. ‘‘High’’ was defined under consideration of the ‘‘normal’’ ranges given by Boege (1998). It would have been favourable for such an analysis to have the original data with individual values and the number of urine samples of exposed and non-exposed persons, respectively. Of course, these data usually are not given in the publications. Instead, a simplified procedure was used by counting over all studies the number of groups (exposed or non-exposed) in which ‘‘high’’ individual values occur and by relating them to the total number of groups in the studies. Fisher’s one-tailed exact test was performed to detect statistically significant differences at the 5% level.

ics. Workplaces analyzed included a rubber tire factory, shoe production, paint manufacturing, painting and spraying, oil refineries, printing, dry cleaning, reinforced plastic industry, boat manufacturing, soil fumigation, and chemical companies (Voss et al. 2003). Past exposure was usually assessed by years of employment or years of exposure. In several studies, duration of exposure was not reported. Many studies that did report duration of exposure covered durations ranging from several months to more than 10 years. Current exposure was assessed by regular measurement of the exposure concentrations at the workplaces, and in some studies additionally by biomonitoring. If no measurements were available, special exposure scores had been developed, which considered the different types of work (Hotz et al. 1990, 1991, 1993, 1997; Yaqoob et al. 1993a, b). Urine samples were taken at different times and frequencies, i.e. spot urine not further specified (Solet and Robins 1991; Viau et al. 1987), overnight urine (Franchini et al. 1983; Verplanke and Herber 1998; Verplanke et al. 1999), second morning urine (Mutti et al. 1992), end-of-shift urine (Viau et al. 1987; Verplanke and Herber 1998; Vyskocil et al. 1990; Lauwerys et al. 1983), pre-shift and post-shift urine (Verplanke and Herber 1998). More extended analyses included urine upon arrival at work and at the end of the work shift on the first and last day of a typical work week (Viau et al. 1987a, b; Vyskocil et al. 1989) or several samples during one work week (Verplanke and Herber 1998; Verplanke et al. 1999). In some studies, the parameters were also followed over time (Brouwer et al. 1991; Ge´rin et al. 1997; Osterloh and Feldmann 1993; Rocskay et al. 1993; Stengel et al. 1998). Since the same spectrum of parameters was measured in both these longitudinal observations and in the cross-sectional studies, the results are described and evaluated together.

Overview of studies Detailed overviews of the cross-sectional studies are given by Hotz (1994) and Voss et al. (2003). In these studies, workers were exposed to a variety of different solvents and more often to solvent mixtures. Solvents included hydrocarbon mixtures such as white spirit and mineral oils, toluene or toluene/xylene mixtures, styrene, methyl ethyl ketone, butoxyethanol, ethylene glycol, tetrachloroethene, trichloroethene, 1,3dichloropropene, several other organochlorine aliphat-

c

d

479

Fig. 2 a Comparison of b2-microglobulin concentration in urine (lg/g creatinine) of solvent exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part I, continued in Fig. 2b. The data points represent mean or median values, bars indicate the range of individual values. Note logarithmical scale. The sequence of studies with tetrachloroethene exposure is arranged according to TWA exposure concentration (in ascending order). The data point for the study of Ge´rin et al. (1997) consists of two squares—almost completely molten together—for two sampling periods: end-of-shift values after de-icing and in period without de-icing, respectively. Abbreviations of study authors: St98 Stengel et al. (1998), Vi87a, b Viau et al. (1987a, 1987b), Vy89 Vyskocil et al. (1989), Ro93 Rocskay et al. (1993), Vy91 Vyskocil et al. (1991), Ge97 Ge´rin et al. (1997), Mu92 Mutti et al. (1992), La83 Lauwerys et al. (1983), Vy90 Vyskocil et al. (1990), Na89 Nagaya et al. (1989) and Bo94 Boogard and Caubo (1994). cf. Tables 2 and 4.1. b Comparison of b2-microglobulin concentration in urine (lg/l urine) of solvent exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part II, continued from Fig. 2a. Abbreviations of study authors: As81a Askergren et al. (1981a), Br91 Brouwer et al. (1991) and Kr85 Krusell et al. (1985). cf. Tables 2 and 4.1

Usually only healthy workers were included in the studies. Workers with a history of kidney disease or nonwork related risk factors for developing kidney disease

such as diabetes mellitus, hypertension, chronic nephritis, renal stones, use of analgesics or possibly nephrotoxic drugs (e.g. diclofenac and ibuprofen) were excluded from several studies (Hotz 1994; Hotz et al. 1990; Yaqoob et al. 1993a; Viau et al. 1987a; Verplanke and Herber 1998; Verplanke et al. 1999; Mutti et al. 1992; Stengel et al. 1998; Meyer et al. 1984; Nagaya et al. 1989; Zschiesche and Triebig 1990).

Results Albumin and other markers of glomerular filtration Excretion of albumin in the urine was assayed most frequently in the cross-sectional studies. The results of the studies that provided the corresponding data are presented in Fig. 1a–c. Figure 1a and b show the results in mg/g creatinine, and in Fig. 1c the albumin concentration is given in mg/l urine. Several studies found statistically significant differences between the mean or median albumin excretion in

480 Table 2 Solvent-exposed groups with individual values of albumin concentration in urine exceeding normal range (37 mg/ g creatinine or 19 mg/l, respectively, according to Boege (1998))

Industry/work

Main compounds

Reference

Oil refinery/laboratory technicians, truck drivers, bulk pant and refinery operators Shoe production/gluers Paint manufacturing and spraying Photogravure printing Floor layers Paint manufacturing Reinforced plastic industry Plastic boat manufacturing Dry cleaning Dry cleaning Dry cleaning Organochlorine plant/shift workers

Aliphatic and aromatic hydrocarbons

Viau et al. (1987a)

Petroleum naphtha, toluene Toluene Toluene Toluene and methanol Xylene/toluene Styrene Styrene Tetrachloroethene Tetrachloroethene Tetrachloroethene Allyl chloride, epichlorhydrin, 1,3-dichloropropene, hexachlorocyclopentadiene

Vyskocil et al. (1991) NG et al. (1990) Askergren et al. (1981) Hotz et al. (1990) Askergren et al. (1981) Vyskocil et al. (1989) Askergren et al. (1981) Verplanke et al. (1999) Solet and Robins (1991) Vyskocil et al. (1990) Boogaard et al. (1993)

Soil fumigation in flower bulb culture

urine of workers exposed to solvents and of controls (Hotz et al. 1990; Viau et al. 1987a; Verplanke and Herber 1998; Mutti et al. 1992; Ge´rin et al. 1997; Askergren et al. 1981; Boogaard et al. 1993; Boogaard and Caubo 1994; Hashimoto et al. 1991; Normand et al. 1989, 1990). The range of values or the standard deviation was frequently higher in workers than in controls and, in addition, higher than the range considered as normal. This indicated an increases of the albumin excretion for individual workers (Hotz et al. 1990; Solet and Robins 1991; Verplanke and Herber 1998; Verplanke et al. 1999; Brouwer et al. 1991; Hashimoto et al. 1991; Ng et al. 1990; Vyskocil et al. 1991; Piscator 1989). Therefore, and in addition to the statistical evaluation of the original papers, we examined the frequency of ‘‘high’’ individual albumin values in comparison to the ‘‘normal’’ ranges given by Boege (1998). The results of the test are given in Table 1; the p-values for albumin excretion indicate ‘‘significance’’. Table 2 presents an overview of the workplaces with increased albumin excretion and the compounds to which the workers were exposed. It is remarkable that the effects are not restricted to certain compounds. Effects have been found with both non-chlorinated and chlorinated compounds. In addition to an increased albumin excretion, statistically significant increases in total protein in the urine were also found in some studies (Franchini et al. 1983; Mutti et al. 1981) as well as high ranges of total protein concentrations (Verplanke et al. 1999). As albumin is the main protein in urine, the results of these studies may also indicate increased levels of albumin in the urine. Transferrin is a marker of glomerular filtration as well. For this parameter increases were found in the studies by Gruener (1992) only in the group of workers exposed for more than 10 years. In the study by Mutti et al. (1992), both albumin and transferrin were increased. In the study by Vyskocil et al. (1991), all three parameters were measured in parallel. Marginally in-

1,3-Dichloropropene

Boogaard and Caubo (1994) Brouwer et al. (1991)

creased ranges were found for total protein and albumin but not for transferrin. b2-Microglobulin (b2M) As is shown in Fig. 2a and b, in contrast to albumin, no significant increase in the mean b2M excretion in solvent-exposed workers was observed in most of the studies. Only one study reported a higher prevalence of abnormally high values in workers exposed to tetrachoroethene (Mutti et al. 1992). Also, an analysis of the frequency of groups containing individuals with ‘‘high’’ values failed to show a relationship (Table 1). N-acetyl-b-D-glucosaminidase (NAG) and other markers of tubular toxicity The results of the cross-sectional studies analyzing NAG are presented in Fig. 3a and b. A statistically significant increase in NAG in the urine has been reported in some cross-sectional studies (Vyskocil et al. 1990, 1989; Meyer et al. 1984; Zschiesche and Triebig 1990; Normand et al. 1989, 1990; Brogren et al. 1986). A higher prevalence of elevated values was found in some other studies (Yaqoob et al. 1993a; Pai et al. 1996, 1998). In a study on printers exposed to toluene, a very slight but statistically significant increase in NAG with a cumulative toluene exposure index (ppm years) was observed (Stengel et al. 1998). No relationship between exposure index and hypertension was observed, but the relation for NAG did not persist after hypertensive patients had been excluded from the study. In the studies by Vyskocil et al. (1989, 1991), NAG was increased in workers exposed to styrene and petroleum hydrocarbons, respectively. The increase was at the end of the last workday of the workweek but not in the arithmetic mean of the four samples per worker, which were taken on the first and

481

Fig. 3 a Comparison of N-acetyl-b-D-glucosaminidase (NAG) activity in urine of solvent-exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part I, continued in Fig. 3b. The data points represent mean or median values, bars indicate the range of individual values. Note logarithmical scale. Asterisks indicate statistically significant differences according to the authors of the studies (full asterisks indicate significant elevation; in the study of Zschiesche and Triebig (1990), significantly lower values were noted for the exposed workers). Abbreviations of study authors: St98 Stengel et al. (1998), VH98 Verplanke and Herber (1998), two groups; Vy89 Vyskocil et al. (1989), Vi87 Viau et al. (1987b), Ro93 Rocskay et al. (1993), Vy91 Vyskocil et al. (1991), ZT90 Zschiesche und Triebig (1990) and Ya93 Yaqoob et al. (1993b). b Comparison of N-acetylb-D-glucosaminidase (NAG) activity in urine of solvent exposed workers (full square symbols) and of non-exposed controls (open square symbols)—part II, continued from Fig. 3a. Abbreviations of study authors: Ya93 Yaqoob et al. (1993b), Gr92 Gruener (1992), Ve99 Verplanke et al. (1999), Br86 Brogren et al. (1986), Ra93 Rasmussen et al. (1993), Se93 Selde´n et al. (1993), Bo93 Boogard et al. (1993), Bo94 Boogard and Caubo (1994) and Mu92 Mutti et al. (1992)

last day of the workweek upon arrival to work and at the end of the workday (Vyskocil et al. 1989). The overall analysis of the frequency of groups containing

individuals with ‘‘high’’ values did not show a statistically significant relationship (Table 1). Retinol-binding protein (RBP), which is considered to be a more sensitive marker of tubular changes, was not increased. In contrast in another study on workers exposed to various solvents in the paint production, both NAG and RBP were increased (Normand et al. 1989, 1990).

Discussion Dose–response No parameter was increased consistently after exposure to a certain chemical, therefore no single nephrotoxic compound could be identified. Different exposure concentrations may be one reason for these discrepancies. However, no consistent pattern is emerging. For example there were some studies with current exposure concentrations of styrene from 86 to 225 mg/m3. Albumin was increased in the studies with

482

TWAs of 113 (47–320) mg/m3 (Verplanke and Herber 1998) and of 20—100 mg/m3 (peaks to 925 mg/m3) (Askergren et al. 1981), which were the studies with the lower exposure concentrations. Chronic exposure could not be compared because the parameters years exposed and years employed as measures of chronic exposure were not given in most of the studies. Similarly, with respect to tetrachloroethene no explanation can be given at first why albumin is increased in the study by Mutti et al. (1992) with a median exposure concentration of 100 mg/m3, but not in the study by Vyskocil et al. (1990) with an exposure to 157 mg/m3. Again, chronic exposure could not be compared due to lack of data. Also, when comparing differently exposed groups, the authors of the studies often did not find a dose–response relationship with respect to exposure concentration or duration of exposure (Solet and Robins 1991; Verplanke et al. 1999; Rocskay et al. 1993; Vyskocil e al. 1991). The prevalence of elevated albuminuria, e.g., was not related to any specific job title or to the duration of employment (Viau et al. 1987a; Askergren 1981; Askergren et al. 1981). However, as is stressed by Hotz and Boillat (1989), a straight-forward dose–response relationship between exposure and renal damage may be too simplistic an assumption, since peak exposures, intermittent exceeding of thresholds, and latency periods might be as important as intensity and duration of exposure. Especially, short-term high exposure concentrations, which are not reflected by TWA exposure concentrations, may be important as demonstrated by the case reports. Furthermore, short-term dermal exposure may be substantial at many solvent-exposed workplaces and may contribute considerably to the overall exposure depending on the type of work and the solvent used. Although some studies mention the importance of this route (Hashimoto et al. 1991; Haufroid et al. 1997), dermal exposure was not taken into account or quantified in any cross-sectional study. Selection of study population Several studies have not analyzed persons with already existing kidney disease or risk factors for kidney disease. However, an existing kidney disease may very well result from the exposure. Therefore, this approach may lead to an underestimation of the effect. Significance of results Our simplified statistical analysis (Table 1) does not allow consideration of the sample sizes of the groups, and the type of analysis was chosen after the data had been evaluated. Therefore the character of the analysis is only preliminary and explorative. However, and despite these limitations, a comparison of the frequencies of ‘‘high’’ values according to Table 1—together with the statisti-

cal significances reported in some studies—can be considered as a support of the hypothesis that occupational exposure to organic solvents and effects on the kidney may be associated. The effects seem to be reflected particularly by increased urinary albumin excretion and may especially or solely occur after high exposure and/or in sensitive subpopulations. Mode of action b2-Microglobulin (b2M) is a small protein (11,800 Da) present at the cell membrane as part of histocompatibility antigens (Bernard and Lauwerys 1991). The high sensitivity for detecting proximal tubular insult by determining urinary b2M results from the fact that under normal conditions its tubular reabsorption is with 99.7% nearly complete (Bernard and Lauwerys 1991; Piscator 1989). Thus, a decrease in reabsorption capacity of 0.1% results in a threefold increase in b2M excretion in urine (Piscator 1989). One advantage of b2M is that the relative clearance can be estimated if the protein is also determined in serum. Since excretion of b2M is a very sensitive parameter for indicating renal tubular damage, the results described could suggest that the tubulotoxic potential of solvent exposure is low in the studies that have measured b2M. This suggestion is supported by the observation that the excretion of (RBP) (data not shown), which indicates similar damage and is of similar sensitivity, has also not been increased in those studies in which both parameters have been measured (Viau et al. 1987a; Lauwerys et al. 1983; Vyskocil et al. 1989, 1991). Results may also be compared with those obtained for NAG activity in urine (see above, Fig. 3a, b). The NAG is a marker of renal tubular injury that is considered to be less sensitive than b2M or RBP. In three studies, both b2M and NAG were measured. An increased NAG activity in urine of solvent-exposed workers was found in the studies by Vyskocil et al. (1989, 1991), and no increase was observed by Mutti et al. (1992). Thus, no increase in b2M excretion and an increased NAG activity in the same group of workers are not readily explicable. However, although both b2M excretion and NAG activity are markers of tubular dysfunction, they do not indicate the same damage: While an increase in b2M in urine indicates a reduced tubular reabsorption ability, an increased NAG activity indicates an enhanced release of lysosomal enzyme from tubular cells. It may also be argued that the well-known acid lability of b2M could have led to a destruction of the enzyme in urine before the sample could be analyzed. If this is true, it should be reflected by low mean values. However, as can be seen from Fig. 2a and b, the values reported in many studies are not exceedingly low; in fact, they are in the normal ranges reported or even at the upper end of the normal range for healthy individuals. Therefore, it is unlikely that degradation of b2M played a major role in the outcome of the studies presented.

483

The authors (Vyskocil et al. 1989) speculate that in this case the increase in NAG may be a result of stimulated exocytosis or increased renal activity rather than a toxic effect. Furthermore, as NAG is a marker of cellular damage, it may reflect rather acute than chronic effects, which is supported by the observed increase only at the end of the workweek by Vyskocil et al. (1989). Some case reports indicate renal damage of tubular origin, and others indicate glomerular damage. In casecontrol studies, different types of glomerulonephritis have been found to be associated with solvent exposure, pointing to glomerular damage. In the cross-sectional studies, indications for both locations have been found. Increased albumin levels in urine point to defects in glomerular filtration of macromolecules. Tubular defects, either on reabsorption (indicated by increases in b2M) or toxic effects with release of enzymes such as NAG from tubular cells, are less pronounced than glomerular effects. Glomerulonephritis is mainly immunologically mediated, either through antibodies reacting with autoantigens of the kidney or by the deposition of immune complexes within renal structures. The solvent-induced reaction with autoantigens is supported by the finding of antiglomerular basement membrane antibodies in workers exposed to hydrocarbons (Stevenson et al. 1995). This would be consistent with the hypothesis that peak exposures may lead to autoimmune glomerular disease only in some susceptible individuals. There are indications, that at least at an early stage, or, if only minor effects are found, the disease or the occurrence of biomarkers in the urine is reversible, if exposure to solvents is discontinued (Yaqoob et al. 1993b; Harrison et al. 1986; Ravnskov 1986).

Conclusion The results of the evaluation of the literature suggest that albumin excretion exceeding the upper limit of the normal range occurs more frequently in solvent-exposed groups of workers than in the corresponding control groups. This has been observed in various groups exposed to different solvents (toluene, styrene, petroleum hydrocarbons, tetrachloroethene, other haloalkenes) at different workplaces. However, possible effects cannot be ascribed to defined solvents or solvent mixtures. As renal damage from solvent exposure may remain clinically silent for many years due to the large functional reserve capacity of the kidney, it is necessary to apply sensitive, reliable early indicators (‘‘biomarkers of effect’’) to detect early effects and prevent further damage. Albumin has been found to be a sensitive biomarker that should be applied in standard occupational examinations of workers exposed to solvents. The determination of albumin in urine can be done by routine laboratory methods. Samples may be obtained from morning second urination. Strenuous physical activity shortly before

sampling should be avoided since it may lead to an increase of albumin excretion. For the interpretation of the results, other well-known individual risk factors, which are assoiciated with albuminuria have to be taken into account, e.g. diabetic nephropathy and hypertension (Boege 1998; Mangelsdorf et al. 2005). Since conventional test strips for screening purposes only detect an albumin excretion, which exceeds the normal rate by a factor of about ten, albumin detection must be carried out by using more sensitive tests. Semiquantitative test strips of high sensitivity which are based on an improved indicator dye method or immunochemical procedures may be used for rapid screening purposes.

References Albrecht W, Boiano J, Smith R (1987) IgA glomerulonephritis in a plumber working with solvent-based pipe cement. Ind Health 25:157–158 Askergren A (1981) Organic solvents and kidney function. A methodological and epidemiologic study. Arbete Och Ha¨lsa 5:1 Askergren A, Allgen LG, Karlsson C, Lundberg I, Nyberg E (1981) Studies on kidney function in subjects exposed to organic solvents. I. Excretion of albumin and beta-2-microglobulin in the urine. Acta Med Scand 209:479–483 Beirne G (1972) Goodpasture’s syndrome and exposure to solvents. J Am Med Assoc 222:1555 Beirne G, Brennan J (1972) Glomerulonephritis associated with hydrocarbon solvents ediated by antiglomerular basement membrane antibody. Arch Environ Health 25:365–369 Bernard A, Lauwerys R (1991) Proteinuria: changes and mechanisms in toxic nephropathies. CRC Crit Rev Toxicol 21:373– 405 Boege F (1998) Urinary proteins. In: Thomas L (ed) Clinical laboratory diagnostics. Frankfurt, TH-Books Verlagsgesellschaft, pp 382–400 Boogaard P, Caubo M (1994) Increased albumin excretion in industrial workers due to the shift work rather than to prolonged exposure to low concentrations of chlorinated hydrocarbons. Occup Environ Med 51(9):638–641 Boogaard P, Rocchi PJ, Van Sittert N (1993) Effects of exposure to low concentrations of chlorinated hydrocarbons on the kidney and liver of industrial workers. Brit J Ind Med 50:331–339 Bosch X, Campistol J, Montoliu J, Revert L (1988) Myelofibrosis and focal segmental glomerulosclerosis associated with toluene poisoning. Hum Toxicol 7:357–361 Brogren C, Christensen J, Rasmussen K (1986) Occupational exposure to chlorinated organic solvents and its effect of the renal excretion of N-acetyl-b-D-glucosaminidase. Arch Toxicol Suppl 9:460–464 Brouwer E, Evelo CA, Verplanke AW, Van Welie RH, De Wolff F (1991) Biological effect monitoring of occupational exposure to 1,3-dichloropropene: effects on liver and renal function and on glutathione conjugation. Br J Ind Med 48:167–172 Carlier B, Schroeder E, Mahieu P (1980) A rapidly and spontaneously reversible Goodpasture’s syndrome after carbon tetrachloride inhalation. Acta Clin Belg 35:193–198 Chen R, Seaton A (1996) A meta-analysis of mortality among workers exposed to organic solvents. J Soc Occup Med 46(5):337–344 Churchill D, Fine A, Gault M (1983) Association between hydrocarbon exposure and glomerulonephritis. An appraisal of the evidence. Nephron 33:169–172 Daniell W, Couser W, Rosenstock L (1988) Occupational solvent exposure and glomerulonephritis. A case report and review of the literature, cited in EHC (1996). J Am Med Assoc 259:2280– 2283

484 D’Apice AF, Kincaid-Smith P, Becker G, Loughhead M, Freeman J, Sands J (1978) Goodpasture’s syndrome in identical twins. Ann Intern Med 88(1):61–62 Delzell E, Austin H, Cole P (1988) Epidemiologic sudies of the petroleum industry. Occup Med 3:455–474 Ehrenreich T (1977) Renal disease from exposure to solvents. Ann Clin Lab Sci 7:6–16 Ehrenreich T, Yunis S, Churg J (1974) Membranous nephropathy following exposure to volatile hydrocarbons. Lab Invest 30:373 Franchini I, Cavatorta A, Falzoi M, Lucertini S, Mutti A (1983) Early indicators of renal damage in workers exposed to organic solvents. Int Arch Occup Environ Health 52:1–9 Ge´rin M, Patrice S, Begin D, Goldberg M, Vyskocil A, Adib G, Drolet D, Viau C (1997) A study of ethylene glycol exposure and kidney function of aircraft de-icing workers. Int Arch Occup Environ Health 69(4):255–265 Glaeser J (1892) III. Zwei Fa¨lle von Nephritis acuta ex intoxicatione. Klin Med (Vienna) 21:388–390 Gruener N (1992) Early detection of changes in kidney function in workers exposed to solvents and heavy metals. Isr J Med Sci 28(8–9):605–607 Harrison D, Thomson D, Macdonald M (1986) Membranous glomerulonephritis. J Clin Pathol 39:167–171 Hashimoto D, Kelsey K, Seitz T, Feldman H, Yakes B, Christiani D (1991) The presence of urinary cellular sediment and albuminuria in newspaper pressworkers exposed to solvents. J Occup Environ Med 33:516–526 Haufroid V, Thirion F, Mertens P, Buchet J, Lison D (1997) Biological monitoring of workers exposed to low levels of 2butoxyethanol. Int Arch Occup Environ Health 70(4):232–236 Hotz P (1994) Occupational hydrocarbon exposure and chronic nephropathy. Toxicology 90:163–283 Hotz P, Boillat M (1989) Kohlenwasserstoffe und nichtneoplastische Nierenkrankheiten. Ther Umsch 46:801–808 Hotz P, Pilliod J, Bernard A, Berode M, Rey F, Mazzocato C, Guillemin M, Boillat M (1990) Hydrocarbon exposure, hypertension and kidney function tests. Int Arch Occup Environ Health 62:501–508 Hotz P, Pilliod J, Berode M, Rey F, Boillat M (1991) Glycosaminoglycans, albuminuria and hydrocarbon exposure. Nephron 58:184–191 Hotz P, Thielemans N, Bernard A, Gutzwiller F, Lauwerys R (1993) Serum laminin, hydrocarbon exposure, and glomerular damage. Br J Ind Med 50:1104–1110 Hotz P, Carbonnelle P, Tschopp A, Rousseau O, Bernard A (1997) Assessment for subclinical kidney damage in workers exposed to low concentrations of hydrocarbons. Int J Occup Environ Health 3:266–272 Huber W, Conradt C, Dangl C, Misol F, Brass H, Strauch M (2000) Organische Lo¨semittel und chronische Glomerulonephritis. Nieren-Hochdruckkr 29(2):41–53 Keogh A, Ibels L, Allen D, Isbester J (1984) Exacerbation of Goodpasture’s syndrome after inadvertent exposure to hydrocarbon fumes. Br Med J 288:188 Krusell, Nielsen H, Baelum J, Lundqvist G, Omland O, Vaeth M, Husted S, Mogensen C, Geday E (1985) Renal effects of chronic exposure to organic solvents: a clinical controlled trial. Acta Med Scand 218:323–327 Lauwerys R, Bernard A (1989) Preclinical detection of nephrotoxicity: describtion of the tests and appraisal of their health significance. Toxicol Lett 46:13–29 Lauwerys R, Herbrand J, Buchet J, Bernard A, Gaussin J (1983) Health surveillance of workers exposed to tetrachlorethylene in dry-cleaning shops. Int Arch Occup Environ Health 52(1):69– 77 Mangelsdorf I, Voss JU, Roller M, Brinkmann (2005) Albumin im Urin als Marker fu¨r die nephrotoxische Wirkungen von Lo¨semitteln. Manuscript (in preparation) Meyer B, Fischbein A, Rosenman K, Lerman Y, Drayer D, Reidenberg M (1984) Increased urinary enzyme excretion in workers exposed to nephrotoxic chemicals. Am J Med 76(6):989–998

Mutti A (1996) Organic solvents and the kidney. J Occup Health 38:162–169 Mutti A, Lucertini S, Falzoi M, Cavatorta A, Franchini I (1981) Organic solvents and chronic glomerulonephritis: a cross-sectional study with negative finding for alipathic and alicyclic C5– C7 hydrocarbons. J Appl Toxicol 1:224–226 Mutti A, Alinovi R, Bergamaschi E, Biagini C, Cavazzini S, Franchini I, Lauwerys R, Bernard A, Roels H, Gelpi E, Rosello J, Ramis I, Price R, Taylor S, De Broe M, Nuyts G, Stolte H, Fels L, Herbort C (1992) Nephropathies and exposure to perchloroethylene in dry-cleaners. Lancet 340:189–193 Nagaya T, Ishikawa N, Hata H (1989) Urinary total protein and beta-2-microglobulin in workers exposed to trichlorethylene. Environ Res 50:86–92 Nathan A, Toseland P (1979) Goodpasture’s syndrome and trichloroethane intoxication. Br J Clin Pharmacol 8:284–286 Ng T, Ong S, Lam W, Jones M, Cheung C, Ong C (1990) Urinary levels of proteins and metabolites in workers exposed to toluene. Int Arch Occup Environ Health 62:43–46 Normand J, Bernard A, Buchet J, Prost G, Lauwerys R (1989) Biome´trologie re´nale de l’exposition chronique aux solvants. Re´sultats pre´liminaires d’une enqueˆte e´pide´miologique transversale dans l’industrie de la peinture. Arch Mal Prof 50:244–245 Normand J, Bernard A, Buchet J, Roels H, Michaux J, De Wandeler V, Lauwerys R (1990) Recherche d’anomalies re´nales infracliniques chez les travailleurs de deux entreprises productrices de peintures. Arch Mal Prof Med Travail Sec Soc 51(4):261–266 Osterloh J, Feldmann B (1993) Urinary protein markers in pesticide applicators during a chlorinated hydrocarbon exposure. Environ Res 63:171–181 Pai P, Hindell P, Stevenson A, Mason H, Bell G (1996) Occupational hydrocarbon (HC) exposure and nephrotoxicity. J Am Soc Nephrol 7:1340 Pai P, Bone J, Bell G (1998) Hydrocarbon exposure and glomerulonephritis due to systemic vasculitis. Nephrol Dial Transplant 13(5):1321–1323 Phillips S, Petrone R, Hemstreet GI (1988) 8. A review of the nonneoplastic kidney effects of hydrocarbon exposure in humans. Occup Med 3:495–509 Piscator M (1989) Markers of tubular dysfunction. Toxicol Lett 46:197–204 Rasmussen K, Brogren C, Sabroe S (1993) Subclinical affection of liver and kidney function and solvent exposure. Int Arch Occup Environ Health 64:445–448 Ravnskov U (1986) Influence of hydrocarbon exposure on the course of glomerulonephritis. Nephron 42:156–160 Ravnskov U (2000) Hydrocarbons may worsen renal function in glomerulonephritis: a meta-analysis of the case-control studies. Am J Ind Med 37:599–606 Rheinhard (1887) Ein Fall von Terpentinintoxication in Folge Einathmens von Terpentinoel. Dtsch Med Wochenschr 13:256 Rocskay A, Robins T, Schork M, Echeverria D, Proctor S, White R (1993) Renal effects of naphtha exposure among automotive workers. J Occup Environ Med 35:617–622 Selde´n A, Hultberg B, Ulander A, Ahlborg G Jr (1993) Trichloroethylene exposure in vapour degreasing and the urinary excretion of N-acetyl-beta-D-glucosaminidase. Arch Toxicol 67:224–226 Solet D, Robins T (1991) Renal function in dry cleaning workers exposed to perchloroethylene. Am J Ind Med 20(5):601–614 Stengel B, Ce´ne´e S, Limasset J, Diebold F, Michard D, Druet P, He´mon D (1998) Immunologic and renal markers among photogravure printers exposed to toluene. Scand J Work Environ Health 24(4):276–284 Stevenson A, Yaqoob M, Mason H, Pai P, Bell G (1995) Biochemical markers of basement membrane disturbances and occupational exposure to hydrocarbons and mixed solvents. Q J Med 88:23–28 Verplanke A, Herber R (1998) Effects on the kidney of occupational exposure to styrene. Int Arch Occup Environ Health 71(1):47–52

485 Verplanke A, Leummens M, Herber R (1999) Occupational exposure to tetrachloroethene and its effects on the kidneys. J Occup Environ Med 41(1):11–16 Viau C, Bernard A, de Russis R, Ouled A, Maldague P, Lauwerys R (1987a) Evaluation of the nephrotoxic potential of styrene in man and in rat. J Appl Toxicol 7(5):313–316 Viau C, Bernard A, Lauwerys R, Buchet J, Quaeghebeur L, Cornu M, Phillips S, Mutti A, Lucertini S, Franchini I (1987b) A cross-sectional survey of kidney function in refinery employees. Am J Ind Med 11:177–187 Voss JU, Roller M, Mangelsdorf I (2003) Nephrotoxicity of organic solvents-evaluation of the literature. Research report BAUA project F5159. German federal institute for occupational safety and Health. Dortmund. http://www.baua.de/english/info/papers/endbericht_f5159.pdf Vyskocil A, Emminger S, Malir F, Fiala Z, Tusl M, Ettlerova E, Bernard A (1989) Lack of nephrotoxicity of styrene at current TLV level (50 ppm). Int Arch Occup Environ Health 61:409–411 Vyskocil A, Emminger S, Tejral J, Fiala Z, Ettlerova E, Cermanova´ A (1990) A Study on kidney function in female workers exposed to perchlorethylene. Hum Exp Toxicol 9:377–380

Vyskocil A, Popler A, Skutilova I, Ciharova M, Ettlerova E, Lauwerys R, Bernard A (1991) Urinary excretion of proteins and enzymes in workers exposed to hydrocarbons in a shoe factory. Int Arch Occup Environ Health 63(5):359–362 Whitworth J, Lawrence J, Meadows R (1974) Goodpasture’s syndrome. A review of nine cases and an evaluation of therapy. Aust N Z J Med 4:167–177 Yaqoob M, Bell G, Stevenson A, Mason H, Percy D (1993a) Renal impairment with chronic hydrocarbon exposure. Q J Med 86:165–174 Yaqoob M, Stevenson A, Mason, Bell G (1993b) Hydrocarbon exposure and tubular damage: additional factors in the progression of renal failure in primary glomerulonephritis. Q J Med 86:661–667 Zschiesche W, Triebig G (1990) Untersuchung von Nierenfunktionsparametern bei langjaehrig loesungsmittelexponierten Malern und Spritzlackierern. Arbeitsmed Sozialmed Pra¨ventivmed 25:259–263