Cadmium concentrations in human kidney in the UK: 1978–1993 Thomas D. B. Lyon,a Elizabeth Aughey,b Roy Scottc and Gordon S. Fella aDepartment of Clinical Biochemistry, Glasgow Royal Infirmary University Trust, Glasgow, UK G4 0SF bDivision of Veterinary Anatomy, University of Glasgow, Veterinary School, Garscube, Stockiemuir Road, Glasgow, UK G61 1QH cDepartment of Urology, Glasgow Royal Infirmary University Trust, Glasgow, UK G4 0SF Received 18th February 1999, Accepted 3rd May 1999 Almost 2700 samples of human renal cortex have been collected from throughout the UK over a 16 year period from 1978 to 1993.The mean Cd concentration was 19 mg g-1 and the median 16 mg g-1.Smokers were, on average, about 5 mg g-1 higher than non-smokers. Cd increased from low concentration in the young to a maximum of 23 mg g-1 in middle age followed by a decrease in old age.Subjects who had died of renal disease had lower Cd concentrations. Geographical variations in the UK are small and the concentrations appear to be static over the 16 year period. Some 3.9% of the population had Cd concentrations >50 mg g-1, the critical level at which b2-microglobulin appears in urine.Aim of investigation Experimental procedure Pathologists from various parts of the UK agreed to provide Cadmium (Cd) is a toxic metal with a wide range of industrial samples of renal cortex (1–2 g wet weight) from the lower applications, which is now dispersed into the environment and pole of the left kidney taken at autopsy. The samples were can be found at low concentrations in most soils, plants and stored at -20 °C, in plastic containers without addition of animal species.Non-industrial human exposure is primarily preservative, and then forwarded frozen in dry cold to our from food and tobacco smoke. Occupational exposure to laboratory. The co-operating centres were Aberdeen, Bath, diVerent types of Cd compounds in fumes and dusts results in Belfast, Bristol, Chesterfield, Dumfries, Edinburgh, Glasgow, significant renal damage and bone disease.Monitoring of Kettering, Leicester, Leeds, London, Newcastle, SheYeld, industrial workers has yielded much information as to bio- Southampton and Watford. Information about age, sex, chemical markers of eVect.1 Cadmium is a persistent toxin underlying cause of death and smoking habit was requested. with a biological half-life of over 20 years.The main target organ is the kidney, which contains about 1/3 of the whole body content. Concerns as to the eVects on the health of the Analysis and quality control general population from environmental pollution were raised The samples were thawed and the wet weight established initially by the Japanese experience of severe bone disease before oven drying overnight at 80 °C.The cadmium concenrelated to Cd release into agricultural land in the waste water trations are reported on a wet weight basis as this is the most from zinc refineries and its uptake by the rice crop.1 In the widely used convention (multiply by 5 to convert to dry UK a geochemical soil survey revealed that high amounts of weight).The dried tissue was digested in 50 ml Pyrex tubes Cd and other metal residues were present in land now used with 7 ml of AristarR grade nitric acid in a heating block for for crop production. This led to an extensive investigation2 60 min at 110 °C and the digest made up to 25 ml with which concluded that there was no evidence of an adverse deionized water.The digests were stored in sealed plastic health eVect to the local population. bottles at 4–10 °C. More recently, Lauwerys et al.3 (Cadimbel Study) have The cadmium concentration of the digest was measured by reported on the renal eVects of Cd on Belgian populations flame atomic absorption spectrometry (FAAS) using a Perkinliving in areas where heavily polluted soils have been identified.Elmer 3030 with air–acetylene flame using the cadmium They have concluded that some of the general population of 228.8 nm line. Belgium have absorbed suYcient Cd to cause slight renal The accuracy of this method was established by dysfunction. They base this on the finding that urine Cd Cunningham5 by comparison with a flame fluorescence excretion above 2 mg l-1 was associated with an increased technique and replicate analysis (n=4) of the standard referurinary excretion of b-microglobulin and other biochemical ence material (National Bureau of Standards 1577, Bovine markers of renal tubular damage.They have calculated that Liver). Cadmium in bovine liver reference material was deterthis urinary output of Cd is equivalent to a renal cortex Cd mined to be 0.26±0.04 mg g-1 (the certified value is content of 50 mg g-1 wet weight. 0.27±0.04 mg g-1). The Bovine Liver SRM has a relatively We now report on the determination of some 2700 renal low cadmium concentration and was not suitable for use as a cortical Cd contents of human kidney obtained at autopsy quality control material in this study. Samples of freeze dried from sixteen locations throughout the UK and conducted over horse kidney with an assigned value of 224 mg g-1 were a 16 year period.The data presented here is an inclusive obtained from the Karolinska Institute and analysed with each extension of that described by Scott et al.,4 which covered the batch of human kidney. Subsequently, IAEA H-8 horse kidney with an assigned value of 184–193 mg g-1 dry weight (95% period 1978–1986.J. Environ. Monit., 1999, 1, 227–231 227Table 1 Distribution of subjects by age (in decades) and gender Age M F All 1 1 1 2 2 3 3 6 3 15 11 26 4 17 20 37 5 51 37 88 6 135 78 213 7 279 231 510 8 371 313 684 9 213 316 529 All 1085 1010 2095 Not known 620 Fig. 1 Quality control chart for IAEA Horse Kidney from 1986 to 1993. decades. The frequency distribution for cadmium in renal cortex was skewed.However, the distribution of the log transformed data was normal and so all statistical tests were confidence interval ) was used from 1986 onwards. The quality performed using the log-normal data. The error bars on all control chart showing the mean value for the IAEA horse figures are 95% confidence intervals. DiVerences between sev- kidney from 1986 to 1993 is shown in Fig. 1. The average eral group means were tested for by one-way analysis of mean over the period was 193±7.9 (s). There was no temporal variance. The Tukey pairwise test was used for ‘follow on’ trend and no results were out of bounds (±3 s). analysis. The association between cadmium concentration and During the course of this study inductively coupled plasma age, sex and smoking habit was examined by multiple mass spectrometry (ICP-MS) became available and an analytregression analysis.ical comparison was made between the FAAS method and All statistical calculations were performed using MinitabR ICP-MS. The results are shown in Fig. 2. The diVerences 10 Xtra. between the values obtained by FAAS and ICP-MS for a number of kidney samples are plotted against the mean result as suggested by Bland and Altman.6 The data are normally Results and discussion distributed about zero diVerence, indicating that there is no Cadmium concentration bias between the two methods.The limit of agreement is +3.2 to -4.4 mg g-1. The frequency distribution of cadmium concentrations in renal cortex is shown in Fig. 3. The distribution has a mean of Population characteristics 19.1 mg g-1 (n=2659) and is skewed. The median is 16 and the 95% confidence interval is 15–16 mg g-1. There is a con- Kidney samples (n=2659) from the general population, siderable variance, with a range of 0.1–163 mg g-1 and inter- including three known coppersmiths, were collected over a 16 year period from 1978 to 1993.The majority of samples were from subjects who had died of natural causes, although some did come from accident victims. The distribution of subjects by age (in decades) and sex is shown in Table 1. In the group of subjects who smoked (or were ex-smokers) a higher proportion were men (60%) compared with women (40%). Subject information is incomplete and there is missing data on gender (n=451, 17%), age (n=620, 23%) and smoking habit (n= 1708, 64%).Statistical analysis For statistical analysis former smokers were grouped along with the smokers and the ages at death were grouped into Fig. 2 Plot of the diVerence (ICP-MS-FAAS) against the mean values for a set of 23 kidney cortexes. The dashed lines indicate the limits of agreement. The solid lines delineate the 95% confidence Fig. 3 Descriptive statistics output from Minitab 10 Xtra for 2659 interval of the mean diVerence and as this includes 0 there is no bias between the methods. kidney cortex Cd concentrations (mg g-1) from the UK. 228 J. Environ. Monit., 1999, 1, 227–231quartile range 9-25 mg g-1. Three of the subjects who were known coppersmiths7 (and therefore exposed to cadmium) had high concentrations of 55, 142 and 163 mg g-1.The mean concentrations found are higher than those of 1–2 mg g-1 found in kidney from the nineteenth century and earlier in the twentieth century.8,9 The present mean is below the ‘critical level’ of 200 mg g-1 suggested by the FAO/WHO10 as the concentration at which the first signs of renal damage, that is selective proteinuria, will become apparent.The Cadimbel Study, however, has suggested that slight renal dysfunction will be seen in 10% of a general population when the urine excretion of Cd exceeds 2 mg l-1, which is observed when the renal cortex Cd is >50 mg g-1. Elinder11 has collated arithmetic means for renal cortex cadmium concentrations from a number of countries for subjects in the age range 40–59 years.We found an arithmetic mean of 23 mg g-1 for this age group in the UK, which is similar to the findings of for Finland, France, USA, Sweden Fig. 5 Frequency distribution of Cd concentration (mg g-1) in cortex and Austria (range 20–40 mg g-1). The range of arithmetic of smokers and non-smokers. For purposes of comparison the distri- means in Japan is much higher, 60–140 mg g-1.bution for the non-smokers has been scaled so as to match the area for the smokers distribution. Sex The geometric mean value for males (n=1202) was 14.8 mg g-1 from both Sweden and the USA,15 but these studies involved and for females (n=1089) 14.6 mg g-1. A ‘t ‘ test performed relatively small numbers (n=74 and n=54), respectively. Other on the lognormal data indicates no significant diVerence ( p= causes for the discrepancy are possible diVerences in the 0.48) between males and females.Neither were there any cadmium content of cigarettes and in diVerences in the selection diVerences between males and females across the age groups. of subjects. These possibilities and also the extent of the smoking habit on cadmium cortex concentrations have been Age discussed elswhere.4 The association of cadmium concentrations with age for 2122 Cause of death individuals is shown in Fig. 4. Cadmium concentrations rise from the small amounts found in the youngest children to a The seven underlying causes of death were used to group the maximum at age 40–60 years. This trend has been noted by data and a one-way analysis of variance on the log-normalised Elinder.11 Infants are known to have low cadmium concen- cadmium concentrations was performed.The analysis indicates trations in their tissues12,13 and it seems reasonable to infer that there are significant diVerences ( p<0.05) between the that cadmium accumulates in kidney throughout life to a group means. The means and confidence intervals are shown maximum of 23 mg g-1 at 50–60 years of age.The subsequent in Fig. 6. The renal disease group is markedly lower, geometric fall in old age cannot be explained by changes in the environ- mean 8 mg g-1, than the other groups, but the size of this mental exposure of the elderly.14 However, it is known that group is small (n=13). Nevertheless, Tukey’s pairwise comthe mass of the kidney decreases in old age with an associated parisons suggest that the diVerence is significant.The mean decrease in renal function. values of all of the other groups are all outwith the confidence interval for the renal group. Likewise, the miscellaneous group Smoking habits is lower than the pulmonary, cancer, heart and cerebrovascular groups. It has been suggested that cadmium may be a causative The subjects were divided into smokers (n=662, including factor in hypertension,16 cardiovascular disease17 and in vari- pipe, cigar and former smokers) and non-smokers (n=372).ous cancers;18–21 however, no evidence of elevated cadmium The results are shown as a histogram in Fig. 5. The nonin tissue from those dying from cardiovascular disease, cerebro- smokers had a median renal cadmium concentration of vascular incident, cancer or bronchitis has been found.22–24 13 mg g-1 (geometric mean=12.6), which is significantly lower than that of the smokers 18 mg g-1 (geometric mean=16.4).The diVerence in geometric means, 3.8, is not as large as the diVerence between geometric means of ~10 mg g-1 reported Fig. 6 Variation of corticular Cd concentration (mg g-1) with Fig. 4 Distribution of Cd concentration (mg g-1) as a function of age (note log scale). underlying cause of death (note log scale). J. Environ. Monit., 1999, 1, 227–231 229from Uppsala with earlier published work for Stockholm and found that the cadmium concentrations had reduced to about 40% of the 1976 concentrations.25,26 This was an unexpected finding as environmental contamination in Sweden has risen over the same period.Unfortunately, an unequivocal interpretation of the Swedish results is not possible in the absence of intercomparative data for the two studies carried out there. Multiple regression analysis A multiple regression model using ‘age’, ‘smoking habit’ and ‘sex’ as explanatory variables was fitted to the data. Although the coeYcients for ‘age’ and ‘smoking’ were highly significant the model was poor and could only explain about 8% of the variation leaving 92% unexplained.Morgan and Sherlock27 demonstrated a strong relation between average dietary intake and kidney cadmium content so it is reasonable to infer that individual dietary intake and associated absorption and Fig. 7 Variation of corticular Cd concentration (mg g-1) with excretion of cadmium is the principal cause of the observed geographical locality (note log scale). variation.Our results suggest that the cadmium concentrations in heart Association of cadmium and zinc disease and in cancer, geometric mean 15 mg g-1, are slightly higher than the miscellaneous group, geometric mean Human autopsy studies have shown an association between 12.6 mg g-1, but the diVerence is not statistically significant.cadmium and zinc.28–30 We confirm a weak association and found that the regression of cadmium on zinc explained 24% Geographical variations within the UK of the variance. A regression of log (Cd) on log (Zn) explained 36% of the variance. The means and confidence intervals for the lognormalised data were calculated for the sixteen localities and the results are shown in Fig. 7. One-way analysis of variance indicates that Conclusions there are statistically significant diVerences ( p<0.003) between This was a large study involving nearly 2700 kidney cortex the diVerent localities. Only one pairwise diVerence was found analyses performed over a 16 year period. Good quality to be significant; Glasgow was slightly higher (geometric control of the analyses was maintained throughout the dur- mean=15.5) than Chesterfield (geometric mean=13.5).The ation of the project. The study confirms the results of smaller diVerence can possibly be explained by the age profile of the studies in other countries which demonstrated a relationship two groups; the proportion of subjects of 80+ years (i.e., with between kidney cortex cadmium and age and smoking habit.age lowered cadmium) is greater in Chesterfield (35%) than in It also shows that the UK levels are among the lowest found. Glasgow (21%). Also, a higher prevalence of smoking in Geographical variations within the UK have been shown to Glasgow is likely and may contribute to the diVerence (smokbe small.No change over the 16 years of the study was ing habit figures for Chesterfield were not available). In general apparent, suggesting cadmium uptake by the UK population though, geographical variations within the UK appear to is not increasing. However, if the critical threshold of 50 mg g-1 be small. in renal cortex, as advocated by Lauwerys et al.,3 is accepted, then a mean value of 19 mg g-1 is not suYciently removed Temporal variations from the critical threshold.The FAO/WHO10 suggested ‘criti- Since small geographical variations may exist the possibility cal level’ of 200 mg g-1 is a factor of ten greater than the mean of any temporal trend was investigated using only data from value found here. If the ‘critical value’ is reduced to 50 mg g-1, the Glasgow area.The data are plotted in Fig. 8. No adjust- which is only twice the mean value, then about 3.9% of the ment was made for any age diVerences between year groups. population would exceed this threshold; any further exposure One-way analysis of variance indicates that there are diVer- of the UK population would therefore be undesirable. ences between the years (F=2.83, p=0.001) but there is no conclusive evidence of a trend with time.However, if the 1990 Acknowledgments data is omitted there is a suggestion of a slight rise. Friis et al. have compared recent (1996) kidney cortex concentrations This study was carried out under a contract (PECD/7/8/36) from the Department of the Environment. The authors thank Mrs. Mary Reilly and Mrs.Lana MacDonald of the Veterinary School, Glasgow, for expert technical assistance. The co-operation of those who supplied the kidneys from post mortem examinations and those who provided demographic information is greatly appreciated. References 1 Cadmium and Health: a Toxicological and Epidemiological Appraisal, eds. L. Friberg, C. G. Elinder, T. Kjellstrom and G. F. Nordberg, CRC Press, Boca Raton, Florida, USA, 1985, vol. 2. 2 J. C. Sherlock, G. A. Smart, B. Walters, W. H. Evans, Fig. 8 Temporal variation of corticular Cd concentration (mg g-1) for D. J. McWeeny and W. Cassidy, Sci. Tot. Environ., 1983, 29, 121. 3 R. Lauwerys, A. Bernard, J. P. Buchet, H. Roels, P. Bruaux, Glasgow (note log scale). 230 J. Environ. Monit., 1999, 1, 227–231F. Claeys, G. DucoVre, P.De Plaen, J. Staessen, A. Amery, 16 H. Schroeder, J. Chron. Dis., 1965, 18, 647. 17 G. L. Friberg, M. Piscator, G. F. Nordberg and T. Kjellstrom, R. Fagard, P. Lijnen, L. Thijs, D. Rondia, F. Sartor, A. Saint Cadmium in the Environment, CRC Press, Cleveland, Ohio, 2nd Remy and L. Nick, Acta Clin. Belg., 1991, 46, 219. edn, 1974, pp. 116–119. 4 R. Scott, E. Aughey, G. S.Fell and H. J. Quinn, Human Toxicol., 18 C. L. Potts, Ann. Occup. Hyg., 1965, 8, 55. 1987, 6, 111. 19 J. M. Morgan, Cancer, 1970, 25, 1394. 5 C. Cunningham, FIMLS thesis, Glasgow Royal Infimary, 1980. 20 L. N. Kolonel, Cancer, 1976, 37, 1782. 6 J. M. Bland and D. G. Altman, Lancet, 1986, i, 307. 21 T. Kjellstrom, L. Friberg and B. Rahnster, Environ. Health 7 R. Scott, P. J. Paterson, E.A. Mills, A. McKirdy, G. S. Fell, Perspect., 1979, 28, 199. J. M. Ottaway, F. E. R. Husain, O. P. Fitzgerald-Finch, 22 T. L. M. Syversen, T. K. Stray, G. B. Syversen and J. Ofstad, A. J. Yates and A. Lamont, Lancet, 1976, 21, 396. Scand. J. Clin. Invest., 1976, 36, 251. 8 C. G. Elinder and T. Kjellstrom, Ambio, 1977, 6, 270. 23 J. M. Morgan, Arch. Intern. Med., 1969, 123, 405. 9 G. A. Drasch, Sci. Tot. Environ., 1983, 26, 111. 24 G. P. Lewis, W. J. Jusko, L. L. Coughlin and S. Hartz, J. Chron. 10 FAO/WHO Evaluation of Certain Food Additives and the Dis., 1972, 25, 717. Contaminants Mercury, Lead and Cadmium. Sixteenth Report of 25 L. Friis, L. Petersson and C. Edling, Environ. Health Perspect., the Joint FAO/WHO Expert Committee on Food Additives, 1998, 106, 175. Geneva, Switzerland, 1974. 26 C. G. Elinder, T. Kjellstrom, L. Friberg, B. Lind and L. Linnman, 11 C. G. Elinder, in Cadmium and Health: A Toxicological and Arch. Environ. Health, 1976, 31, 292. Epidemiological Appraisal, eds. L. Friberg, C. G. Elinder, 27 H. Morgan and J. C. Sherlock, Food Add. Contam., 1984, 1, 45. T. Kjellstrom and G. F. Nordberg, CRC Press, Boca Raton, 28 C. G. Elinder, M. Piscator and L. Linnman, Environ. Res. 1977, Florida, USA, 1985, vol. 1, pp. 81–102. 13, 432. 12 M.Piscator, VDI Berichte, 1975, 203, 43. 29 D. I. Hammer, A. V. Colucci, V. Hasselblad, M. E. Williams and 13 S. B. Gross, D. W. Yeager and D. S. Middendorf, J. Toxicol. C. J. Pinkerton, Occup. Med., 1973, 12, 956. 30 M. Piscator and B. Lind, Arch. Environ. Health, 1972, 24, 426. Environ. Health, 1976, 2, 153. 14 C. C. Travis and A. G. Haddock, Environ. Res., 1980, 22, 46. 15 T. Kjellstrom, Environ. Health Perspect., 1979, 28, 169. Paper 9/01366K J. Environ. Monit., 1999, 1, 227–231 231