Early indicators of renal dysfunction in silicotic workers.

OBJECTIVES - The aim of the study was to determine whether silicosis is associated with renal alter ations detectable in urinary or blood-borne indicators of nephrotoxicity. M ETHODS - The study used a cross-sectional design. The subjects comprised 116male workers who had been exposed to silica for at least two years and had been diagnosed as having silicosis and 61 age-matched referents. The considered outcomemeasures were the concentrations of betafmicroglobulin and creatinine in serumand the urinary excretion of albumin, retinol-binding protein, ana beta-N-acetyl Dvglucosaminidase. RESULTS - Compared with the referents,the silicotic subjects excreted, on the average, slightly higher amounts of albumin, retinol-binding protein, and beta-N-acetyl-D-glucosaminidase. This increase did not correlate with the duration of exposure or the stage of silicosis and was not associated with an elevation in serum creatinine. The concentration of beta,-microglobulin in the serum of silicotic sub jects showed a tendency to rise that became significanf in the subgroup with pseudotumoral opaci ties. This effect. which did not correlate with markers of nephrotoxicity. is however more likely the consequence of silicosis-associated inflammatory reactions than of decreased renal filtration. CONCLUSIONS - The present study confirms that silicosis is associated with some infraclinical renal alterations. However. in the absence of a relationship with length of exposure or severity of silicosis, the implication of silica in their causation needs to be examined further.

An association between occ upational exposure to silica and chronic nephropathy has been suspe cted for many years. Animal studies have demonstrated that silica ca n produce a dose-related nephropath y that is primarily tubular and is ass ociated with inter stitial inflamma tory rea cti on s and fibrosis ( I, 2) . Several ca ses of nephropathy in workers exposed to silica ha ve been reported (3)(4)(5)(6)(7)(8)(9). The nephropathy has usually consis ted of a focal or seg mental, rapidly progressi ve glomerulonephritis without sig nific ant immune-complex deposits and with variable degrees of tubular degeneration. The nephropathy has been described even in the absence of silicosis ( 10). Abnormally elevated levels of silicon have been found in the kidneys of patients with silica-associated nephrop ath y. and the examination of biopsy materials has reve aled silic a deposits in epithelial cells and in subepitheli al and subendothelial areas of the basement membrane s. Altogether these findings sugg es t that silica ma y exe rt direct toxicity on th e kidneys although in some cases the nephrotoxicit y might also be med iated by the immune system (3,9). The hypothesis of chronic nephropathy in ass oci ation with 180 silica exposure is supported by several epidemiologic studies reporting an increased risk of end-stage renal disease or mortality from renal diseases in relatio n with silica exposure (10)(11)(12). Additional eviden ce has been pro vided by a recent cross-sect ion al study show ing an exce ss of albumin, alpha .vrnicrog lobulin, and beta-N-acetyl-D-glucosaminidase in the urin e of a small gro up of graphite workers ( 13). The objective of the present cross-section al stud y was to confirm the latter findings on a larger group of subjects suffering from silicosis and to examine the possible relations between ren al alterations and the duration of exp osure or severity of silicosis.

Subjects and methods
The study was ca rried out on 116 male wor kers with di agn osed silicos is compensated by the Bel gian Occupational Dise ase Fund. All of the workers had been ex posed to si lica for at least two ye ars in und erground mining. It was ascertained that none of them wa s suffering from diseases that could in volve the kidneys (hypertension, diabetes, nephrolithiasis, diabetes) or had been exposed to known nephrotoxic substances such as heavy met als or solvents. Th e mean duration of exposure to silica was 14.9 (ra nge 2-35) years. and on the average they had not been exposed for 23.2 years (table 1).
The referen ce group con sisted of 61 age-matched male subj ects selected from the ge neral population .
Each subject provi ded a sample of blood and urine and completed a questionnaire collecting information on the subjects' medical and employment history and their con sumption of tobacco, drugs, and alcohol. Chest radiographs were taken of all of the silicotic workers, and the fil ms were read by experienced pne umonologists of the Belgian Occ upational Disease Fund. Radiological opacities we re categorized as punctiform (p), micro or nodular (m or n), or pseudot umoral (A, B, or C), and thei r den sity was divided into three grades according to the fLO cla ssification of pneumoconiosis (14) . The concentratio ns of albumin, retinol-binding protein (R BP) and beta 2microglobulin (132-m) in serum or urine were determined by latex immunoassay (15 ). Creatinine was me asured by Jaffe's method, and the activity of beta -N-acetyl-D-g lucosaminidase (NAG) in urine was assayed by the fluo rimetric method of Tucker et al (16).
Stat istical tests were don e with Statview SE software (Abacus Concept, 1979). All of the variables (except age) were log-transformed before the ana lysis . The Kolmogorov -Smirnov test was app lied to check the normality of the distri butions, whereas the homogeneity of varia nces was ascertained by the Ftest. Differences between the exposed and reference subj ects were assessed by Student's t-test (twotailed), and the relations with radiological opacities were examined by a one -way ana lysis of variance followed by Scheffe's mult ip le comparison test. Pre valence s of abnormally elevated value s were calculated with the geometric mean plus two geometric standard deviations of the value s observed for the referents as the upper limit of normal. The pre valences of abnormal values amon g the silicotic workers and the referents were compared with the chisquare test with Yates' correction. Pearson correlation coefficients were calculated to examine the relations between renal parameters, whereas the determinant s of the indicators of nephrotoxicity were traced in a stepwise regression analy sis. Results were considered statistically significant at P < 0.05.
Scand J Work Enviro n Health 1994, vol 20, no 3

Results
Compared with their age -ma tched referents, the subjects who had been exposed to silica had , on the average, significantly higher concentrations of albumin, RBP, and NAG in their urine (table I) . No significan t difference was fou nd in the mean va lues of creatinine and 132-m in serum, although the latter showed a slig ht tend enc y to be increased in silicotic workers (P =0.08). The prevalences of ab normally elevated values of the markers did not differ between the groups (results not shown).
The correlation analysis showed a significant degree of correlation between albumin , RBP, and NAG in the urine of both the reference and silicotic sub- The determinants of the urinary or serum indicators of nephrotoxicity in the silicotic subjects were examined by stepwise regression wit h age , duration of expo sure , size, and density of lung opa cities being used as the predictors. Apart from the associations of age with urinary NAG and serum 132-m, the only significant finding that emerged from thi s analysis was a positive correlatio n between the size of opacities and the serum levels of 132-m (r =0.3, P = 0.0007). To assess further the influence of opacity size , we compared the mean values of the different markers for subjects with punctiform, micro or nodular, or pseudotumoral opacities (table 2) . Whereas urinary markers and seru m creatinine on the average did not differ between the subgroups, the mean Table 1     , 1 mg or UI . g crealinine-' =0.113 mg or UI . g creatin ine-'. " Significantly different from the reference . "" Significantly different from the group with punctiform opacities.
values of serum~2-m increased in parallel with the size of the opacities.

Discussion
We found that workers with silicos is excreted slightly more albumin, RBP, and NAG in their urine than age-matched referents did . Th ese results confirm the recent observations of Ng et al ( 13), who reported an increased urinary excretion of albumin, NAG and alpha-rnicroglobulin in 33 silica workers in comparison with 16 referents, on a much larger number of subjects. Since alpha.vmicro globul in is a marker of pro xim al tubule dy sfunction comparabl e to RBP ( 17), the pattern of effects obser ved among our workers and amon g tho se examined by Ng et al ( 13) can be con sidered identical. Both studi es were also carried out on workers who had stopped exposure several years earlier; therefore ren al alterations associated with silica exposure may be irrever sible.
It is important to stre ss however that the silicotic workers examined by us were not distinguishable from their referents in the prevalences of abnormally elevated values of the renal markers. The difference between both groups was thus restricted to a very slight shift in the distribution of urinary markers. Thi s observation, combined with the fact that the seru m creatinine values were , in mo st cases , normal, testifi es to very discrete effects that do not impair renal function. An elevation of~2-m was found in serum, but this effect, which co rrelated with the severity of silicosis and not with markers of nephrotoxicity, is more likel y the con sequence of some silicosis-associated inflammatory cha nges than of a decline in renal filtration. It would be premature, however, to conclude from th ese obse rva tions that renal effects associated with silicos is ha ve no clinical signifi can ce . The possibility of a selection bias inherent to any cro ss-sectional study appears espe ciall y prob able here since the work er s were ex amined an average of 23 years after the cessation of exposure. Thi s possibility is also sugg es ted by the epidemio-182 logic evidence (10)(11)(12) of an excess of end-stage ren al diseases or of mortality from renal diseases in relation with silica exposure.
Although our results are consistent with the hypothesis of a nephrotoxic action of silica, they provide no correlational evidence of the implication of silica in their causa tion. Indeed, no association could be dem on strated between renal alteratio ns and the duration of exposure or the severity of silicos is. Thi s findin g might be due to the long delay bet ween removal from exposure and the examination (sur vival bias), and probably also to the inadequacy of the indic ator s used to evaluate the exposure to silica. But it might also sugg est that silicosis and renal alterations are not linked causally, but instead throu gh their common associations with a third factor. Thi s third factor co uld be , for instance, some kind of a predi sposition to immunologicall y based reacti on s that wo uld favor the development of silicosis but also , independently of the exposure to silica, the appearance of abnor malities in other organ s, including the kidn ey. The lack of a relationship with the stage of sili cosis is however in agreement with the view expressed in the literature (3,9) that silica-associated nephropathy is not simply a complication of silicosis , but might result from a nephrotoxic action of soluble sili ca absorbed by the lung s. Th is po ssibility, supported by animal data (1, 2), could be tested by study ing the do se-response effect rel ati ons between signs of renal dysfunction and indicato rs of the internal dose of silica in a cross -sec tional study on currentl y ex posed workers without silicosis .