This article was downloaded by: [Michigan State University] On: 23 February 2015, At: 06:59 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Occupational and Environmental Hygiene Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uoeh20
Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures a
a
b
b
ac
Jean-François Sauvé , Martine Lévesque , Mélanie Huard , Daniel Drolet , Jérôme Lavoué , a
b
Robert Tardif & Ginette Truchon a
Department of Environmental and Occupational Health, Université de Montréal, Montréal, Québec, Canada b
Institut de recherche Robert-Sauvé en santé et en sécurité du travail, Montréal, Québec, Canada c
Université de Montréal Hospital Research Center (CRCHUM), Montréal, Québec, Canada Accepted author version posted online: 05 Sep 2014.Published online: 31 Dec 2014.
Click for updates To cite this article: Jean-François Sauvé, Martine Lévesque, Mélanie Huard, Daniel Drolet, Jérôme Lavoué, Robert Tardif & Ginette Truchon (2015) Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures, Journal of Occupational and Environmental Hygiene, 12:2, 123-129, DOI: 10.1080/15459624.2014.955179 To link to this article: http://dx.doi.org/10.1080/15459624.2014.955179
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Occupational and Environmental Hygiene, 12: 123–129 ISSN: 1545-9624 print / 1545-9632 online c 2015 JOEH, LLC Copyright DOI: 10.1080/15459624.2014.955179
Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures 1 ´ 1 Martine Levesque, ´ ´ Jean-Franc¸ois Sauve, Melanie Huard,2 Daniel Drolet,2 ´ ome ˆ ´ 1,3 Robert Tardif,1 and Ginette Truchon2 Jer Lavoue, 1
Downloaded by [Michigan State University] at 06:59 23 February 2015
Department of Environmental and Occupational Health, Universite´ de Montreal, ´ Montreal, ´ Quebec, ´ Canada 2 Institut de recherche Robert-Sauve´ en sante´ et en securit e´ du travail, Montreal, Canada ´ ´ Quebec, ´ 3 Universite´ de Montreal Canada ´ Hospital Research Center (CRCHUM), Montreal, ´ Quebec, ´
Reference values for the biological monitoring of occupational exposures are generally normalized on the basis of creatinine (CR) concentration or specific gravity (SG) to account for fluctuations in urine dilution. For instance, the American Conference of Governmental Industrial Hygienists (ACGIH) uses a reference value of 1g/L for CR. The comparison of urinary concentrations of biomarkers between studies requires the adjustment of results based on a reference CR and/or SG value, although studies have suggested that age, sex, muscle mass, and time of the day can exert nonnegligible influences on CR excretion, while SG appears to be less affected. The objective of this study was to propose reference values for urinary CR and SG based on the results of samples sent for analysis by occupational health practitioners to the laboratory of the Occupational Health and Safety Research Institute of Qu´ebec (IRSST). We analyzed a database containing 20,395 urinary sample results collected between 1985 and 2010. Linear mixed-effects models with worker as a random effect were used to estimate the influence of sex and collection period on urinary CR and SG. Median CR concentrations were 25–30% higher in men (1.6 g/L or 14.4 mmol/L) than in women (1.2 g/L or 10.2 mmol/L). Four percent of the samples for men and 12% for women were below the acceptable threshold for CR (4.4 mmol/L). For SG, 5% of samples for men and 12% for women were below the threshold of 1.010. The difference in SG levels between sexes was lower than for CR, with a median of 1.024 for men compared to 1.020 for women. Our results suggest that the normalization of reference values based on a standard CR value of 1 g/L as proposed by the ACGIH is a conservative approach. According to the literature, CR excretion is more influenced by physiological parameters than SG. We therefore suggest that correction based on SG should be favored in future studies involving the proposal of reference values for the biological monitoring of occupational exposures.
Address correspondence to: Ginette Truchon, Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail, 505 boul. de Maisonneuve Ouest, Montr´eal, Qu´ebec, H3A 3C2, Canada; e-mail: [email protected]
R
Keywords
biological monitoring, creatinine, occupational exposure, specific gravity
INTRODUCTION
S
pot urine samples are usually favored for the biological monitoring of occupational exposures since 24-hr urine sample collection is rarely possible in the work environment.(1) However, spot urine concentration of biomarkers can vary widely due to fluctuation of the urinary flow rate.(2) To overcome this problem, creatinine (CR) or specific gravity (SG) normalization is used. These two common adjustment methods minimize variation in biomarker concentrations arising from the effects of hydration.(3) Effectively, since variation in urinary flow rate due to hydration is more important than the variations affecting CR or SG excretion the correction of biological indicator concentrations measured in spot samples remains essential.(4,5) Biological exposure indices (BEI) are assessment values used in occupational medicine to assess chemical exposure and health risk to workers.(6,7) They are based on published studies with bio-indicator concentrations usually expressed as CR and/or SG adjusted values. In some cases, such as for biomarkers excreted by tubular diffusion, no adjustment is required.(8) CR normalization is the approach traditionally used to adjust urinary concentrations to the level of urine dilution.(9,10) The rationale of using CR comes from the assumption of a constant CR excretion rate. In theory, CR normalization should only be used to adjust urinary concentrations of biomarkers
Journal of Occupational and Environmental Hygiene
February 2015
123
Downloaded by [Michigan State University] at 06:59 23 February 2015
that share the same processes of renal elimination as CR, i.e., through glomerular filtration without tubular reabsorption.(8,11) However, CR excretion rate is not constant and could be influenced by factors such as age, sex, and muscle mass.(8) CR normalization is applied simply by dividing the concentration of the biological parameter by the CR concentration of the urine sample.(8,12) Recent studies have suggested that the influence of sex, age, and muscle mass is much lower on the normalization on the basis of SG than of CR.(10,12–17) The accuracy of SG normalization can vary according to the molecular size or weight of the solutes found in the urine. For the same mass of solutes, the increase in SG is larger when the solutes consist mainly of high molecular weight molecules (e.g., glucose, proteins) as opposed to molecules of lower molecular weight (e.g., sodium, urea).(18,19) For instance, Chadha et al.(20) reported an increase in SG—as measured by refractometry—of 0.003 associated with a proteinuria of 10 g/L, and an increase of 0.002 associated with a glycosuria of 10 g/L. Hence, adjusting by SG would be less appropriate for workers affected by diabetes mellitus or with severe kidney diseases.(18) The correction of concentrations by SG for various urinary biomarkers is performed using Eq. (1), based on a reference SG (SGref ) value: Ccorr =
Ci (SGref − 1) (SG − 1)
(1)
Ccorr. = Corrected concentration Ci = Measured concentration of the biological indicator SGref = Reference specific gravity value, usually between 1.016 and 1.024 SG = Specific gravity of the urine analyzed Since the establishment and reevaluation of BEIs are generally based on data from multiple studies, there currently exist questions within the scientific community about which reference CR or SG values should be used for the conversion of the urinary concentrations when different studies have to be compared and when their results are expressed in different units (corrected for CR or SG).(12,13,21) The aim of this study is to propose reference values for urinary CR and SG by using the results of a database obtained from samples sent for analysis by occupational health practitioners to the laboratory of the Occupational Health and Safety Research Institute of Qu´ebec (IRSST). METHODS Measurement Database The occupational health practitioners of the Qu´ebec prevention network regularly take blood and urinary samples for the biological monitoring of workers’ exposure. Between 30 and 40 different bio-indicators are analyzed on a regular basis by the IRSST. From 1985 onwards, the analytical results (measured CR, SG, and biomarker concentrations) were entered in a computerized database (Laboratory Information Management System - LIMS) along with ancillary information 124
such as sampling period and sample reception date at the laboratory. Urinary CR was analyzed by a spectrometric method based on Jaffe’s reaction with a detection limit of 0.06 mmol/L.(22) Urinary SG was determined by refractometry (Density meter DMA 38, Anton-Paar, Saint-Laurent, Canada). The IRSST considers a urine sample as acceptable (neither too dilute nor concentrated) when the CR and SG levels fall within the range of 4.4–26.5 mmol/L and 1.010–1.030 g/mL, respectively.(23) For this study, the analytical results from urine samples in the LIMS for the 1985–2010 period were extracted and compiled in an Excel spreadsheet. The CR and SG levels for each urinary sample were accompanied by information on the sample reception date and year, collection period, name of the worker, and a unique identifying number for each worker. The age of the workers at the time of sample collection was not available in the database. The sex of the workers was also not originally entered in the database and was imputed (by M.L.) on the basis of the workers’ names for this project. In the event that the sex of a worker could not be ascertained with reasonable confidence, the results pertaining to this worker were excluded from further analyses. In addition, when multiple biomarkers were assessed from a single urine sample, the result for each biomarker was entered separately in the database during the data entry process, with the associated SG and CR values being identical. For such cases, a single SG and CR result was retained in this analysis and the remaining entries corresponding to the same urine sample were excluded.
Statistical Analysis The arithmetic mean (AM), geometric mean (GM), arithmetic standard deviation (ASD), geometric standard deviation (GSD), and selected percentiles of the distribution of CR levels and SG were computed over all the samples of the database, as well as stratified by sex and sampling period. The relationship between the CR concentrations and SG for each sample for which both of these parameters were available was also evaluated. Hexagonal binning(24,25) was used to represent this relationship graphically due to the large number of measurements in the database and the significant overlap of the data points. Conceptually similar to a bivariate histogram, this approach can be seen as superimposing a hexagonal grid on a scatterplot, with the color or shade of each hexagonal cell being proportional to the number of data points contained within the cell. Multivariate statistical modeling was used to estimate the combined influence of sex and sampling period on the CR concentrations and SG. Linear mixed-effect models, with worker as a random effect, were used to account for the potential correlation of CR and SG for a same worker. These models allow for the evaluation of the within-worker (σw2 ) and between-worker (σb2 ) variance components. These two parameters can then be used to estimate the correlation between two repeated measures (ρ) of CR or SG for a given
Journal of Occupational and Environmental Hygiene
February 2015
worker according to Eq. (2).(26,27)
Downloaded by [Michigan State University] at 06:59 23 February 2015
ρ=
σb2 σb2 + σw2
(2)
For both CR and SG, five models were evaluated, each containing a unique combination of the sex and sampling period variables as well as their interaction. The interaction term was included to assess whether the effect of the collection period differed between men and women. The Akaike information criterion (AIC) was then computed for each of the five different models, with the model associated with the lowest AIC value being selected as the final model. The models were fitted using the nlme package of the R statistical software (R Foundation for Statistical Computing, Vienna, Austria). Log-transformation was applied to the CR concentrations prior to modeling due to the positive skew of their distribution, while SG levels followed an approximately normal distribution and no log-transformation was used. The value of 1 was subtracted from the SG results to reflect how this parameter is treated in Eq. (1). Relative exposure indices (RIE)(28) were computed from the model coefficients to estimate the influence of sex and collection period on CR concentrations as a percentage of increase or decrease compared to a reference category. The combination of men/pre-shift collection period was selected as the reference category, taken as 100%. For the SG results, predictions were made for each of the six sex/sampling period combinations. The predictions were then divided by the prediction of the reference combination (men/pre-shift) to express the combined influence of sex and sampling period as percentages, as it was for CR.
RESULTS Descriptive Statistics of the Measurement Database Following the exclusion of duplicate entries and records for which the sex of the worker could not be ascertained with confidence, the database contained a total of 20,395 samples over a 25-year period, from a high of 4983 measurements in 1996 to a low of 121 samples for 1985. CR concentration was not available for 199 samples (1%), and for 199 samples (1%) for SG. Only 30 samples had no information reported for both parameters. A total of 9193 workers were represented in the database, with 3944 (43%) workers providing two samples or more. The majority (88%) of the samples came from male workers, while 50% (n = 10,200) of the samples were collected at the end of the work shift compared to 17% (n = 2488) for the pre-shift collection period, with the remaining 33% collected at other times during the day. The AM, ASD, GM, GSD, and selected percentiles of the CR and SG levels by sex and collection period are presented in Table I. None of the creatinine results were reported as being lower than the limit of detection. The median CR concentrations were lower for women (10.2 mmol/L) than for men (14.4 mmol/L). Regarding the acceptability of the CR
FIGURE 1. Relationship between creatinine concentrations and specific gravity levels (n = 20,027)
levels in the sample (Table II), relative to the IRSST’s criteria (between 4.4 mmol/L and 26.5 mmol/L), 4% of the samples for men (n = 642) were below the acceptable range compared to 12% for women (n = 300). Conversely, 5% of the samples for men (n = 846) were considered too concentrated, compared to 1% for women (n = 36). In contrast to the sex-specific GMs, the distribution of CR levels by collection period was much more homogenous, with a GM of 12.4 mmol/L (pre-shift) and 12.9 mmol/L (post-shift). The SG results exhibited a pattern similar to the one found for CR, with values for women being generally lower than those of men, with the women’s AM of 1.019, compared to 1.023 for men (AM being used due to the approximately normal distribution of the SG values in the database). Compared to the IRSST’s acceptability criteria for SG (Table II), 5% of samples for men (n = 808) and 12% for women (n = 300) were below the threshold of 1.010, while 6% of samples for men (n = 1086) and 4% for women (n = 89) were too concentrated relative the upper limit of 1.030. The distribution by collection period was also similar, with a pre-shift AM of 1.023 and a post-shift AM of 1.022. The variability in the SG measurements was also very small, with an ASD of 0.006 over all the reported measurements. A Pearson correlation coefficient of 0.83 was observed between the log-transformed CR concentrations and the SG results. The graphical relationship between these two variables, using the hexagonal binning approach, is presented in Figure 1. Statistical Modeling The coefficients for the fixed and random effects of the CR and SG models, selected on the basis of the AIC, are presented in Table III. For the CR model, all three variables evaluated—sex, period, and their interaction—were retained in the final model, whereas the interaction was dropped from the model for SG. The estimated correlation between two results
Journal of Occupational and Environmental Hygiene
February 2015
125
TABLE I. Descriptive Statistics of the Creatinine and Specific Gravity Levels in the Database Creatinine (mmol/L)
Downloaded by [Michigan State University] at 06:59 23 February 2015
Sex Parameter
Total Men
1st percentile 2.5th percentile 5th percentile 10th percentile 25th percentile 50th percentile 75th percentile 90th percentile 95th percentile 97.5th percentile 99th percentile Arithmetic mean Geometric mean Arithmetic SD Geometric SD Number of samplesA
2.20 4.50 3.20 6.50 10.00 14.00 18.30 22.90 26.00 29.00 32.80 14.49 12.84 6.47 1.71 20196
AExcludes
2.60 5.10 3.70 7.20 10.70 14.40 18.70 23.10 26.20 29.20 33.03 14.95 13.43 6.37 1.65 17873
Specific gravity
Sampling period
Sex
Women Pre-shift Post-shift Other Total Men 1.10 2.40 1.80 3.70 6.40 10.20 14.40 18.90 22.59 25.00 28.26 10.94 9.09 6.08 1.71 2323
2.40 4.50 3.30 6.23 9.40 13.50 18.00 22.80 25.90 28.64 31.22 14.09 12.44 6.47 1.72 3384
2.20 4.60 3.20 6.60 10.30 14.10 18.20 22.60 25.60 28.70 32.30 14.50 12.92 6.34 1.69 10166
2.00 4.30 3.10 6.40 10.00 14.10 18.70 23.00 26.28 29.00 33.56 14.67 12.93 6.65 1.74 6646
1.005 1.007 1.009 1.013 1.019 1.023 1.027 1.029 1.031 1.032 1.034 1.022 1.022 0.006 1.006 20196
1.005 1.007 1.010 1.014 1.020 1.024 1.027 1.029 1.031 1.032 1.034 1.023 1.023 0.006 1.006 17811
Sampling period
Women Pre-shift Post-shift Other 1.003 1.005 1.006 1.008 1.014 1.020 1.024 1.028 1.030 1.031 1.034 1.019 1.019 0.007 1.006 2385
1.005 1.007 1.009 1.013 1.020 1.024 1.027 1.030 1.031 1.031 1.034 1.023 1.023 0.006 1.006 3383
1.005 1.007 1.009 1.012 1.019 1.023 1.027 1.029 1.031 1.032 1.034 1.022 1.022 0.006 1.006 10161
1.005 1.007 1.009 1.012 1.017 1.022 1.025 1.028 1.030 1.032 1.033 1.021 1.021 0.006 1.006 6652
samples with missing creatinine and/or specific gravity.
for a given worker was moderate for both CR (ρ = 0.33) and SG (ρ = 0.38). The coefficient of determination (R2) for the fixed effects was 5.5% for CR and 4.1% for SG. Regarding the influence of sex and period on the CR levels (Table IV), compared to the reference combination of men/preshift, the estimated pre-shift levels for women were 29% lower (95% CI 27%–30%). The difference between pre-shift and post-shift levels were negligible for both sexes, with a 1% decrease for men (95% CI 0%–2%) and a 1% increase for women (95% CI 0–3%) relative to their respective pre-shift categories. For the measures of SG (following the subtraction
of 1), the pre-shift levels for women were 15.3% lower compared to the men’s (95% CI 13.7%–16.9%). The postshift levels were increased by 5.2% compared to the pre-shift collection period (95% 4.6%–5.9%); the effect was the same for the two sexes due to interaction between sex and collection period being absent in the final model. DISCUSSION
T
his study reports urinary CR and SG results from 20,395 urine samples collected by occupational health and safety
TABLE II. Number of Creatinine and Specific Gravity Outside the Acceptability Guidelines of the IRSST, ACGIH, and World Health Organization (WHO), Stratified by Sex R
Creatinine IRSST/ACGIH 26.5 mmol/L WHO 300 mg/dLB Specific gravity IRSST/ACGIH/WHO 1.03 g/mL ACorresponds BCorresponds
126
Men N(%)
Women N(%)
642 (4%) 846 (5%)
300 (13%) 36 (2%)
198 (1%) 846 (5%)
136 (6%) 36 (2%)
808 (5%) 1086 (6%)
300 (13%) 89 (4%)
to 2.65 mmol/L. to 26.5 mmol/L.
Journal of Occupational and Environmental Hygiene
February 2015
TABLE III. Coefficients of the Models for Creatinine and Specific Gravity Specific gravity (−1)A
Creatinine (ln(mmol/L)) β
Downloaded by [Michigan State University] at 06:59 23 February 2015
Fixed effects
β
Std. Err.
Intercept Sex Men Women Sampling period Pre-shift
2.581
0.011
0.0216
Post-shift Other Sex:period interactionB Women/post-shift Women/other Random effects σB σW ρ R2 (fixed effects) C
−0.009 0.038
0.011 0.013
0.0011 0.0010
0.025 −0.077
0.030 0.032
— —
Reference −0.339
Std. Err. 0.0001 Reference
−0.0033
0.024 Reference
0.0002 Reference
0.303 0.430 0.331 5.5%
0.0001 0.0001
0.004 0.005 0.375 4.1%
AThe
value of 1 was subtracted from the specific gravity measurements prior to model fitting. interaction was not retained in the final model for specific gravity. CThe R2 was computed on a model without the worker random effect. BThis
practitioners and sent for analysis to the IRSST’s laboratory for the years 1985 to 2010. Our results show that the median CR concentrations were 25 to 30% higher in men (1.6 g/L or 14.4 mmol/L) than in women (1.2 g/L or 10.2 mmol/L). The effect of the collection period on CR levels was much more subtle. Other factors such as age, meat consumption, and the importance of muscle mass may also influence the levels of CR.(10) However, these factors could not be explored in our study as they were not available in the database, which was originally developed for purposes of administrative record-keeping rather than for the investigation of potential determinants of CR and SG excretion. The values reported here are in the same range as those found by other authors for adult populations. For example, Cocker et al.(21) reported mean (13 mmol/L) and median
TABLE IV.
(12 mmol/L) CR concentrations in men that were higher compared to women (mean: 9 mmol/L; median: 10 mmol/L). Bader et al.(29) reported mean CR concentrations of 12.9 mmol/L for men (median: 12.1 mmol/L) and 9.9 mmol/L for women (median: 8.8 mmol/L), while mean concentrations observed by Carrieri et al.(13) were 15.7 mmol/L and 13.0 mmol/L for men and women, respectively. Taken together, these results suggest that median CR concentrations of 14.4 mmol/L (1.6 g/L) for men and 10.2 mmol/L (1.2 g/L) for women seem to be representative of a workers’ population. These results are in agreement with those of Alessio et al.(30) and Sieniawska et al.(31) who suggested that different biological monitoring values should be considered for men and women when CR adjustment is used to correct spot urine concentrations for urine output fluctuation. Regarding the correlation of
Relative Effects of Sex and Collection Period on Creatinine and Specific Gravity Levels Specific gravity (−1)A
Creatinine Period
Men
Women
Men
Women
Pre-shift Post-shift Other
ReferenceB 99 (98– 100)C 104 (103– 105)
71 (70– 73) 72 (71– 74) 68 (67– 70)
Reference 105.2 (104.6– 105.9) 104.6 (103.7– 105.5)
84.7 (83.1– 86.3) 89.9 (88.4– 91.4) 89.3 (87.7– 90.8)
AThe
value of 1 was subtracted from the specific gravity measurements prior to model fitting. level taken as 100%. CRelative effect (%) and 95% confidence interval. BReference
Journal of Occupational and Environmental Hygiene
February 2015
127
Downloaded by [Michigan State University] at 06:59 23 February 2015
CR measurements for a given worker, Cocker et al.(21) reported σ w and σ b values of 0.49 and 0.35, respectively, corresponding to an intra-worker correlation (ρ) of 0.34 which is virtually identical to the value of 0.33 observed in our study. The comparison of biological monitoring results from different studies, expressed in different units (corrected for CR or SG), requires the use of conversion factors based on a tentative average urinary CR concentration. For instance, ACGIH applies a mean CR concentration of 1 g/L.(32) Indeed, the new BEI proposed by the ACGIH (33) for urinary chromium (increase during shift) is 10 μg/L, based on the former BEI of 10 μg/g CR converted using a mean CR concentration of 1 g/L. Our results, as well as data from other studies suggest that the conversion of BEI units should use CR values in the order of 1.6 g/L for men and 1.2 g/L for women. Using these values, the corrected BEI of urinary chromium would be 16 and 12 μg/L, respectively, for men and women. Considering that the mean CR concentration in a population of workers is greater than 1 g/L, the use of a mean CR value of 1 g/L to convert BEIs expressed in g/g CR into g/L corresponds to the use of a conservative approach. The characterization of urinary SG appears to be less investigated compared to CR, especially in populations of workers. The AM of SG levels for men in our study was 1.023 (SD = 0.006, n = 17,811) which was higher than for women (1.019 ± 0.007, n = 2385). When accounting for the correction of biomarker results for SG in Eq. (1), these correspond to levels on average 15% higher in men than in women, and 5% higher at the end of the shift compared to pre-shift. Carrieri et al.(13) did not find any meaningful difference in the SG values measured at the beginning of the work shift compared to the end of the shift, with post-shift values of 1.023 ± 0.005 for men (n = 346) and 1.020 ± 0.005 for women (n = 133). Suwazono et al.(10) reported significantly lower average SG values for women (1.020 ± 0.008, n = 434) compared to men (1.023 ± 0.007, n = 289) for morning urine collection. The values for post-shift urine collection reported in the same study were 1.015 ± 0.006 for women and 1.018 ± 0.006 for men. Our results as well as data from other studies suggest that mean SG values for men and women workers are, respectively, in the order of 1.023 and 1.020. Hence, as for CR, SG values for women are generally lower than those for men. While trends observed for sex were similar, according to the literature, the influence of sex, age, and muscle mass appears to be lower for SG than for CR.(10,12–17) In addition, SG represents an interesting alternative as its determination is simpler and less cost-intensive than for urinary CR.(17,18,34) For these reasons, some authors suggested that SG correction would be preferable to CR adjustment.(16,18) Based on results observed in our study as well as on information available from the literature, we suggest that correction based on SG should be favored in future studies involving the proposal of reference values for purposes of biological monitoring of occupational exposures. Such an approach would avoid, where possible, the transformation of 128
units of BEIs according to an estimation on mean CR or SG values, which could sometimes lead to significant bias. However, should the transformation of units be done for BEIs, this should be applied based on the average CR and SG values relevant to a population of workers, such as those reported in this study. FUNDING
F
inancial support from the Institut de recherche RobertSauv´e en sant´e et en s´ecurit´e du travail (grant no. 20100059) is gratefully acknowledged. JFS was supported by the Fonds de recherche du Qu´ebec—Sant´e (FRQS) and the IRSST. REFERENCES
1. Sorahan, T., D. Pang, N. Esmen, and S. Sadhra: Urinary concentrations of toxic substances: An assessment of alternative approaches to adjusting for specific gravity. J. Occup. Environ. Hyg. 5(11):721–723 (2008). 2. Akerstrom, M., T. Lundh, L. Barregard, and G. Sallsten: Sampling of urinary cadmium: Differences between 24-h urine and overnight spot urine sampling, and impact of adjustment for dilution. Int. Arch. Occup. Environ. Health 85(2):189–196 (2012). 3. Cone, E.J., Y.H. Caplan, F. Moser, T. Robert, M.K. Shelby, and D.L. Black: Normalization of urinary drug concentrations with specific gravity and creatinine. J. Anal. Toxicol. 33(1):1–7 (2009). 4. Fortin, M.-C., G. Carrier, and M. Bouchard: Concentrations versus amounts of biomarkers in urine: A comparison of approaches to assess pyrethroid exposure. Environ. Health 7(1):1–13 (2008). 5. Garde, A.H., Å.M. Hansen, J. Kristiansen, and L.E. Knudsen: Comparison of uncertainties related to standardization of urine samples with volume and creatinine concentration. Ann. Occup. Hyg. 48(2):171–179 (2004). 6. American Conference of Governmental Industrial Hygienists (ACGIHR ): Documentation of the Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2001. 7. Lauwerys, R., V. Haufroid, P. Hoet, and D. Lison: Toxicologie industrielle et intoxications professionnelles (Industrial toxicology and occupational intoxications). Paris: Elsevier Masson, 2007. [in French] 8. Boeniger, M.F., L.K. Lowry, and J. Rosenberg: Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: A review. Am. Ind. Hyg. Assoc. J. 54(10):615–627 (1993). 9. Barr, D.B., L.C. Wilder, S.P. Caudill, A.J. Gonzalez, L.L. Needham, and J.L. Pirkle: Urinary creatinine concentrations in the US population: Implications for urinary biologic monitoring measurements. Environ. Health Perspect. 113(2):192 (2005). 10. Suwazono, Y., A. Åkesson, T. Alfven, L. J¨arup, and M. Vahter: Creatinine versus specific gravity-adjusted urinary cadmium concentrations. Biomarkers 10(2–3):117–126 (2005). 11. Waikar, S.S., V.S. Sabbisetti, and J.V. Bonventre: Normalization of urinary biomarkers to creatinine during changes in glomerular filtration rate. Kidney Int. 78(5):486–494 (2010). 12. Heavner, D.L., W.T. Morgan, S.B. Sears, J.D. Richardson, G.D. Byrd, and M.W. Ogden: Effect of creatinine and specific gravity normalization techniques on xenobiotic biomarkers in smokers’ spot and 24-h urines. J. Pharm. Biomed. Anal. 40(4):928–942 (2006). 13. Carrieri, M., A. Trevisan, and G.B. Bartolucci: Adjustment to concentration-dilution of spot urine samples: Correlation between specific gravity and creatinine. Int. Arch. Occup. Environ. Health 74(1):63–67 (2001). 14. Moriguchi, J., T. Ezaki, T. Tsukahara, et al.: Decreases in urine specific gravity and urinary creatinine in elderly women. Int. Arch. Occup. Environ. Health 78(6):438–445 (2005).
Journal of Occupational and Environmental Hygiene
February 2015
Downloaded by [Michigan State University] at 06:59 23 February 2015
15. Nermell, B., A.-L. Lindberg, M. Rahman, et al.: Urinary arsenic concentration adjustment factors and malnutrition. Environ. Res. 106(2):212–218 (2008). 16. Dai, X., K. Kutschke, and N. Priest: Rapid urinary output normalization method using specific gravity. Health Phys. 101(2):140–143 (2011). 17. Pearson, M.A., C. Lu, B.J. Schmotzer, L.A. Waller, and A.M. Riederer: Evaluation of physiological measures for correcting variation in urinary output: Implications for assessing environmental chemical exposure in children. J. Expo. Sci. Environ. Epidemiol. 19(3):336–342 (2009). 18. Miller, R.C., E. Brindle, D.J. Holman, et al.: Comparison of specific gravity and creatinine for normalizing urinary reproductive hormone concentrations. Clin. Chem. 50(5):924–932 (2004). 19. Watson, A.: Urine specific gravity in practice. Aust. Vet. J. 76(6):392 (1998). 20. Chadha, V., U. Garg, and U.S. Alon: Measurement of urinary concentration: A critical appraisal of methodologies. Pediatr. Nephrol. 16(4):374–382 (2001). 21. Cocker, J., H. Mason, N. Warren, and R. Cotton: Creatinine adjustment of biological monitoring results. Occup. Med. 61(5):349–353 (2011). 22. Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail (IRSST): D´etermination de la cr´eatinine urinaire (Determination of urinary creatinine). Analytical method 163-2. Montr´eal, Quebec, Canada: IRSST, 1996. [in French] 23. Gagn´e, S.: Guide de pr´el`evement des e´ chantillons biologiques (Biological samples collection guide). Technical Report T-25. Montr´eal, Quebec, Canada: Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail (IRSST), 2013. [in French] 24. Carr, D.B., R.J. Littlefield, W.L. Nicholson, and J.S. Littlefield: Scatterplot matrix techniques for large N. J. Am. Stat. Assoc. 82(398):424–436 (1987).
25. Minnotte, M.C., S.R. Sain, and D. Scott: Multivariate Visualization by Density Estimation. In Handbook of Data Visualization, C.-H. Chen, W. H¨ardle, and A. Unwin (eds.). London: Springer, 2008. pp. 389–413. 26. Peretz, C., A. Goren, T. Smid, and H. Kromhout: Application of mixedeffects models for exposure assessment. Ann. Occup. Hyg. 46(1):69–77 (2002). 27. Rappaport, S.M., and L.L. Kupper: Quantitative Exposure Assessment. El Cerrito, Calif.: Stephen Rappaport, 2008. 28. Lavou´e, J., R. Vincent, and M. G´erin: Statistical modelling of formaldehyde occupational exposure levels in French industries, 1986–2003. Ann. Occup. Hyg. 50(3):305–321 (2006). 29. Bader, M., P. Messerer, and W. Will: Urinary creatinine concentrations in an industrial workforce and comparison with reference values of the general population. Int. Arch. Occup. Environ. Health 86(6):673–680 (2013). 30. Alessio, L., A. Berlin, A. Dell’Orto, F. Toffoletto, and I. Ghezzi: Reliability of urinary creatinine as a parameter used to adjust values of urinary biological indicators. Int. Arch. Occup. Environ. Health 55(2):99–106 (1985). 31. Sieniawska, C.E., L.C. Jung, R. Olufadi, and V. Walker: Twenty-fourhour urinary trace element excretion: Reference intervals and interpretive issues. Ann. Clin. Biochem. 49(4):341–351 (2012). 32. American Conference of Governmental Industrial Hygienists (ACGIHR ): Documentation of the Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2009. 33. American Conference of Governmental Industrial Hygienists (ACGIH): Threshold Limit Values and Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2010. 34. Haddow, J.E., G.J. Knight, G.E. Palomaki, L.M. Neveux, and B.A. Chilmonczyk: Replacing creatinine measurements with specific gravity values to adjust urine cotinine concentrations. Clin. Chem. 40(4):562–564 (1994).
Journal of Occupational and Environmental Hygiene
February 2015
129
Journal of Occupational and Environmental Hygiene Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uoeh20
Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures a
a
b
b
ac
Jean-François Sauvé , Martine Lévesque , Mélanie Huard , Daniel Drolet , Jérôme Lavoué , a
b
Robert Tardif & Ginette Truchon a
Department of Environmental and Occupational Health, Université de Montréal, Montréal, Québec, Canada b
Institut de recherche Robert-Sauvé en santé et en sécurité du travail, Montréal, Québec, Canada c
Université de Montréal Hospital Research Center (CRCHUM), Montréal, Québec, Canada Accepted author version posted online: 05 Sep 2014.Published online: 31 Dec 2014.
Click for updates To cite this article: Jean-François Sauvé, Martine Lévesque, Mélanie Huard, Daniel Drolet, Jérôme Lavoué, Robert Tardif & Ginette Truchon (2015) Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures, Journal of Occupational and Environmental Hygiene, 12:2, 123-129, DOI: 10.1080/15459624.2014.955179 To link to this article: http://dx.doi.org/10.1080/15459624.2014.955179
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Occupational and Environmental Hygiene, 12: 123–129 ISSN: 1545-9624 print / 1545-9632 online c 2015 JOEH, LLC Copyright DOI: 10.1080/15459624.2014.955179
Creatinine and Specific Gravity Normalization in Biological Monitoring of Occupational Exposures 1 ´ 1 Martine Levesque, ´ ´ Jean-Franc¸ois Sauve, Melanie Huard,2 Daniel Drolet,2 ´ ome ˆ ´ 1,3 Robert Tardif,1 and Ginette Truchon2 Jer Lavoue, 1
Downloaded by [Michigan State University] at 06:59 23 February 2015
Department of Environmental and Occupational Health, Universite´ de Montreal, ´ Montreal, ´ Quebec, ´ Canada 2 Institut de recherche Robert-Sauve´ en sante´ et en securit e´ du travail, Montreal, Canada ´ ´ Quebec, ´ 3 Universite´ de Montreal Canada ´ Hospital Research Center (CRCHUM), Montreal, ´ Quebec, ´
Reference values for the biological monitoring of occupational exposures are generally normalized on the basis of creatinine (CR) concentration or specific gravity (SG) to account for fluctuations in urine dilution. For instance, the American Conference of Governmental Industrial Hygienists (ACGIH) uses a reference value of 1g/L for CR. The comparison of urinary concentrations of biomarkers between studies requires the adjustment of results based on a reference CR and/or SG value, although studies have suggested that age, sex, muscle mass, and time of the day can exert nonnegligible influences on CR excretion, while SG appears to be less affected. The objective of this study was to propose reference values for urinary CR and SG based on the results of samples sent for analysis by occupational health practitioners to the laboratory of the Occupational Health and Safety Research Institute of Qu´ebec (IRSST). We analyzed a database containing 20,395 urinary sample results collected between 1985 and 2010. Linear mixed-effects models with worker as a random effect were used to estimate the influence of sex and collection period on urinary CR and SG. Median CR concentrations were 25–30% higher in men (1.6 g/L or 14.4 mmol/L) than in women (1.2 g/L or 10.2 mmol/L). Four percent of the samples for men and 12% for women were below the acceptable threshold for CR (4.4 mmol/L). For SG, 5% of samples for men and 12% for women were below the threshold of 1.010. The difference in SG levels between sexes was lower than for CR, with a median of 1.024 for men compared to 1.020 for women. Our results suggest that the normalization of reference values based on a standard CR value of 1 g/L as proposed by the ACGIH is a conservative approach. According to the literature, CR excretion is more influenced by physiological parameters than SG. We therefore suggest that correction based on SG should be favored in future studies involving the proposal of reference values for the biological monitoring of occupational exposures.
Address correspondence to: Ginette Truchon, Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail, 505 boul. de Maisonneuve Ouest, Montr´eal, Qu´ebec, H3A 3C2, Canada; e-mail: [email protected]
R
Keywords
biological monitoring, creatinine, occupational exposure, specific gravity
INTRODUCTION
S
pot urine samples are usually favored for the biological monitoring of occupational exposures since 24-hr urine sample collection is rarely possible in the work environment.(1) However, spot urine concentration of biomarkers can vary widely due to fluctuation of the urinary flow rate.(2) To overcome this problem, creatinine (CR) or specific gravity (SG) normalization is used. These two common adjustment methods minimize variation in biomarker concentrations arising from the effects of hydration.(3) Effectively, since variation in urinary flow rate due to hydration is more important than the variations affecting CR or SG excretion the correction of biological indicator concentrations measured in spot samples remains essential.(4,5) Biological exposure indices (BEI) are assessment values used in occupational medicine to assess chemical exposure and health risk to workers.(6,7) They are based on published studies with bio-indicator concentrations usually expressed as CR and/or SG adjusted values. In some cases, such as for biomarkers excreted by tubular diffusion, no adjustment is required.(8) CR normalization is the approach traditionally used to adjust urinary concentrations to the level of urine dilution.(9,10) The rationale of using CR comes from the assumption of a constant CR excretion rate. In theory, CR normalization should only be used to adjust urinary concentrations of biomarkers
Journal of Occupational and Environmental Hygiene
February 2015
123
Downloaded by [Michigan State University] at 06:59 23 February 2015
that share the same processes of renal elimination as CR, i.e., through glomerular filtration without tubular reabsorption.(8,11) However, CR excretion rate is not constant and could be influenced by factors such as age, sex, and muscle mass.(8) CR normalization is applied simply by dividing the concentration of the biological parameter by the CR concentration of the urine sample.(8,12) Recent studies have suggested that the influence of sex, age, and muscle mass is much lower on the normalization on the basis of SG than of CR.(10,12–17) The accuracy of SG normalization can vary according to the molecular size or weight of the solutes found in the urine. For the same mass of solutes, the increase in SG is larger when the solutes consist mainly of high molecular weight molecules (e.g., glucose, proteins) as opposed to molecules of lower molecular weight (e.g., sodium, urea).(18,19) For instance, Chadha et al.(20) reported an increase in SG—as measured by refractometry—of 0.003 associated with a proteinuria of 10 g/L, and an increase of 0.002 associated with a glycosuria of 10 g/L. Hence, adjusting by SG would be less appropriate for workers affected by diabetes mellitus or with severe kidney diseases.(18) The correction of concentrations by SG for various urinary biomarkers is performed using Eq. (1), based on a reference SG (SGref ) value: Ccorr =
Ci (SGref − 1) (SG − 1)
(1)
Ccorr. = Corrected concentration Ci = Measured concentration of the biological indicator SGref = Reference specific gravity value, usually between 1.016 and 1.024 SG = Specific gravity of the urine analyzed Since the establishment and reevaluation of BEIs are generally based on data from multiple studies, there currently exist questions within the scientific community about which reference CR or SG values should be used for the conversion of the urinary concentrations when different studies have to be compared and when their results are expressed in different units (corrected for CR or SG).(12,13,21) The aim of this study is to propose reference values for urinary CR and SG by using the results of a database obtained from samples sent for analysis by occupational health practitioners to the laboratory of the Occupational Health and Safety Research Institute of Qu´ebec (IRSST). METHODS Measurement Database The occupational health practitioners of the Qu´ebec prevention network regularly take blood and urinary samples for the biological monitoring of workers’ exposure. Between 30 and 40 different bio-indicators are analyzed on a regular basis by the IRSST. From 1985 onwards, the analytical results (measured CR, SG, and biomarker concentrations) were entered in a computerized database (Laboratory Information Management System - LIMS) along with ancillary information 124
such as sampling period and sample reception date at the laboratory. Urinary CR was analyzed by a spectrometric method based on Jaffe’s reaction with a detection limit of 0.06 mmol/L.(22) Urinary SG was determined by refractometry (Density meter DMA 38, Anton-Paar, Saint-Laurent, Canada). The IRSST considers a urine sample as acceptable (neither too dilute nor concentrated) when the CR and SG levels fall within the range of 4.4–26.5 mmol/L and 1.010–1.030 g/mL, respectively.(23) For this study, the analytical results from urine samples in the LIMS for the 1985–2010 period were extracted and compiled in an Excel spreadsheet. The CR and SG levels for each urinary sample were accompanied by information on the sample reception date and year, collection period, name of the worker, and a unique identifying number for each worker. The age of the workers at the time of sample collection was not available in the database. The sex of the workers was also not originally entered in the database and was imputed (by M.L.) on the basis of the workers’ names for this project. In the event that the sex of a worker could not be ascertained with reasonable confidence, the results pertaining to this worker were excluded from further analyses. In addition, when multiple biomarkers were assessed from a single urine sample, the result for each biomarker was entered separately in the database during the data entry process, with the associated SG and CR values being identical. For such cases, a single SG and CR result was retained in this analysis and the remaining entries corresponding to the same urine sample were excluded.
Statistical Analysis The arithmetic mean (AM), geometric mean (GM), arithmetic standard deviation (ASD), geometric standard deviation (GSD), and selected percentiles of the distribution of CR levels and SG were computed over all the samples of the database, as well as stratified by sex and sampling period. The relationship between the CR concentrations and SG for each sample for which both of these parameters were available was also evaluated. Hexagonal binning(24,25) was used to represent this relationship graphically due to the large number of measurements in the database and the significant overlap of the data points. Conceptually similar to a bivariate histogram, this approach can be seen as superimposing a hexagonal grid on a scatterplot, with the color or shade of each hexagonal cell being proportional to the number of data points contained within the cell. Multivariate statistical modeling was used to estimate the combined influence of sex and sampling period on the CR concentrations and SG. Linear mixed-effect models, with worker as a random effect, were used to account for the potential correlation of CR and SG for a same worker. These models allow for the evaluation of the within-worker (σw2 ) and between-worker (σb2 ) variance components. These two parameters can then be used to estimate the correlation between two repeated measures (ρ) of CR or SG for a given
Journal of Occupational and Environmental Hygiene
February 2015
worker according to Eq. (2).(26,27)
Downloaded by [Michigan State University] at 06:59 23 February 2015
ρ=
σb2 σb2 + σw2
(2)
For both CR and SG, five models were evaluated, each containing a unique combination of the sex and sampling period variables as well as their interaction. The interaction term was included to assess whether the effect of the collection period differed between men and women. The Akaike information criterion (AIC) was then computed for each of the five different models, with the model associated with the lowest AIC value being selected as the final model. The models were fitted using the nlme package of the R statistical software (R Foundation for Statistical Computing, Vienna, Austria). Log-transformation was applied to the CR concentrations prior to modeling due to the positive skew of their distribution, while SG levels followed an approximately normal distribution and no log-transformation was used. The value of 1 was subtracted from the SG results to reflect how this parameter is treated in Eq. (1). Relative exposure indices (RIE)(28) were computed from the model coefficients to estimate the influence of sex and collection period on CR concentrations as a percentage of increase or decrease compared to a reference category. The combination of men/pre-shift collection period was selected as the reference category, taken as 100%. For the SG results, predictions were made for each of the six sex/sampling period combinations. The predictions were then divided by the prediction of the reference combination (men/pre-shift) to express the combined influence of sex and sampling period as percentages, as it was for CR.
RESULTS Descriptive Statistics of the Measurement Database Following the exclusion of duplicate entries and records for which the sex of the worker could not be ascertained with confidence, the database contained a total of 20,395 samples over a 25-year period, from a high of 4983 measurements in 1996 to a low of 121 samples for 1985. CR concentration was not available for 199 samples (1%), and for 199 samples (1%) for SG. Only 30 samples had no information reported for both parameters. A total of 9193 workers were represented in the database, with 3944 (43%) workers providing two samples or more. The majority (88%) of the samples came from male workers, while 50% (n = 10,200) of the samples were collected at the end of the work shift compared to 17% (n = 2488) for the pre-shift collection period, with the remaining 33% collected at other times during the day. The AM, ASD, GM, GSD, and selected percentiles of the CR and SG levels by sex and collection period are presented in Table I. None of the creatinine results were reported as being lower than the limit of detection. The median CR concentrations were lower for women (10.2 mmol/L) than for men (14.4 mmol/L). Regarding the acceptability of the CR
FIGURE 1. Relationship between creatinine concentrations and specific gravity levels (n = 20,027)
levels in the sample (Table II), relative to the IRSST’s criteria (between 4.4 mmol/L and 26.5 mmol/L), 4% of the samples for men (n = 642) were below the acceptable range compared to 12% for women (n = 300). Conversely, 5% of the samples for men (n = 846) were considered too concentrated, compared to 1% for women (n = 36). In contrast to the sex-specific GMs, the distribution of CR levels by collection period was much more homogenous, with a GM of 12.4 mmol/L (pre-shift) and 12.9 mmol/L (post-shift). The SG results exhibited a pattern similar to the one found for CR, with values for women being generally lower than those of men, with the women’s AM of 1.019, compared to 1.023 for men (AM being used due to the approximately normal distribution of the SG values in the database). Compared to the IRSST’s acceptability criteria for SG (Table II), 5% of samples for men (n = 808) and 12% for women (n = 300) were below the threshold of 1.010, while 6% of samples for men (n = 1086) and 4% for women (n = 89) were too concentrated relative the upper limit of 1.030. The distribution by collection period was also similar, with a pre-shift AM of 1.023 and a post-shift AM of 1.022. The variability in the SG measurements was also very small, with an ASD of 0.006 over all the reported measurements. A Pearson correlation coefficient of 0.83 was observed between the log-transformed CR concentrations and the SG results. The graphical relationship between these two variables, using the hexagonal binning approach, is presented in Figure 1. Statistical Modeling The coefficients for the fixed and random effects of the CR and SG models, selected on the basis of the AIC, are presented in Table III. For the CR model, all three variables evaluated—sex, period, and their interaction—were retained in the final model, whereas the interaction was dropped from the model for SG. The estimated correlation between two results
Journal of Occupational and Environmental Hygiene
February 2015
125
TABLE I. Descriptive Statistics of the Creatinine and Specific Gravity Levels in the Database Creatinine (mmol/L)
Downloaded by [Michigan State University] at 06:59 23 February 2015
Sex Parameter
Total Men
1st percentile 2.5th percentile 5th percentile 10th percentile 25th percentile 50th percentile 75th percentile 90th percentile 95th percentile 97.5th percentile 99th percentile Arithmetic mean Geometric mean Arithmetic SD Geometric SD Number of samplesA
2.20 4.50 3.20 6.50 10.00 14.00 18.30 22.90 26.00 29.00 32.80 14.49 12.84 6.47 1.71 20196
AExcludes
2.60 5.10 3.70 7.20 10.70 14.40 18.70 23.10 26.20 29.20 33.03 14.95 13.43 6.37 1.65 17873
Specific gravity
Sampling period
Sex
Women Pre-shift Post-shift Other Total Men 1.10 2.40 1.80 3.70 6.40 10.20 14.40 18.90 22.59 25.00 28.26 10.94 9.09 6.08 1.71 2323
2.40 4.50 3.30 6.23 9.40 13.50 18.00 22.80 25.90 28.64 31.22 14.09 12.44 6.47 1.72 3384
2.20 4.60 3.20 6.60 10.30 14.10 18.20 22.60 25.60 28.70 32.30 14.50 12.92 6.34 1.69 10166
2.00 4.30 3.10 6.40 10.00 14.10 18.70 23.00 26.28 29.00 33.56 14.67 12.93 6.65 1.74 6646
1.005 1.007 1.009 1.013 1.019 1.023 1.027 1.029 1.031 1.032 1.034 1.022 1.022 0.006 1.006 20196
1.005 1.007 1.010 1.014 1.020 1.024 1.027 1.029 1.031 1.032 1.034 1.023 1.023 0.006 1.006 17811
Sampling period
Women Pre-shift Post-shift Other 1.003 1.005 1.006 1.008 1.014 1.020 1.024 1.028 1.030 1.031 1.034 1.019 1.019 0.007 1.006 2385
1.005 1.007 1.009 1.013 1.020 1.024 1.027 1.030 1.031 1.031 1.034 1.023 1.023 0.006 1.006 3383
1.005 1.007 1.009 1.012 1.019 1.023 1.027 1.029 1.031 1.032 1.034 1.022 1.022 0.006 1.006 10161
1.005 1.007 1.009 1.012 1.017 1.022 1.025 1.028 1.030 1.032 1.033 1.021 1.021 0.006 1.006 6652
samples with missing creatinine and/or specific gravity.
for a given worker was moderate for both CR (ρ = 0.33) and SG (ρ = 0.38). The coefficient of determination (R2) for the fixed effects was 5.5% for CR and 4.1% for SG. Regarding the influence of sex and period on the CR levels (Table IV), compared to the reference combination of men/preshift, the estimated pre-shift levels for women were 29% lower (95% CI 27%–30%). The difference between pre-shift and post-shift levels were negligible for both sexes, with a 1% decrease for men (95% CI 0%–2%) and a 1% increase for women (95% CI 0–3%) relative to their respective pre-shift categories. For the measures of SG (following the subtraction
of 1), the pre-shift levels for women were 15.3% lower compared to the men’s (95% CI 13.7%–16.9%). The postshift levels were increased by 5.2% compared to the pre-shift collection period (95% 4.6%–5.9%); the effect was the same for the two sexes due to interaction between sex and collection period being absent in the final model. DISCUSSION
T
his study reports urinary CR and SG results from 20,395 urine samples collected by occupational health and safety
TABLE II. Number of Creatinine and Specific Gravity Outside the Acceptability Guidelines of the IRSST, ACGIH, and World Health Organization (WHO), Stratified by Sex R
Creatinine IRSST/ACGIH 26.5 mmol/L WHO 300 mg/dLB Specific gravity IRSST/ACGIH/WHO 1.03 g/mL ACorresponds BCorresponds
126
Men N(%)
Women N(%)
642 (4%) 846 (5%)
300 (13%) 36 (2%)
198 (1%) 846 (5%)
136 (6%) 36 (2%)
808 (5%) 1086 (6%)
300 (13%) 89 (4%)
to 2.65 mmol/L. to 26.5 mmol/L.
Journal of Occupational and Environmental Hygiene
February 2015
TABLE III. Coefficients of the Models for Creatinine and Specific Gravity Specific gravity (−1)A
Creatinine (ln(mmol/L)) β
Downloaded by [Michigan State University] at 06:59 23 February 2015
Fixed effects
β
Std. Err.
Intercept Sex Men Women Sampling period Pre-shift
2.581
0.011
0.0216
Post-shift Other Sex:period interactionB Women/post-shift Women/other Random effects σB σW ρ R2 (fixed effects) C
−0.009 0.038
0.011 0.013
0.0011 0.0010
0.025 −0.077
0.030 0.032
— —
Reference −0.339
Std. Err. 0.0001 Reference
−0.0033
0.024 Reference
0.0002 Reference
0.303 0.430 0.331 5.5%
0.0001 0.0001
0.004 0.005 0.375 4.1%
AThe
value of 1 was subtracted from the specific gravity measurements prior to model fitting. interaction was not retained in the final model for specific gravity. CThe R2 was computed on a model without the worker random effect. BThis
practitioners and sent for analysis to the IRSST’s laboratory for the years 1985 to 2010. Our results show that the median CR concentrations were 25 to 30% higher in men (1.6 g/L or 14.4 mmol/L) than in women (1.2 g/L or 10.2 mmol/L). The effect of the collection period on CR levels was much more subtle. Other factors such as age, meat consumption, and the importance of muscle mass may also influence the levels of CR.(10) However, these factors could not be explored in our study as they were not available in the database, which was originally developed for purposes of administrative record-keeping rather than for the investigation of potential determinants of CR and SG excretion. The values reported here are in the same range as those found by other authors for adult populations. For example, Cocker et al.(21) reported mean (13 mmol/L) and median
TABLE IV.
(12 mmol/L) CR concentrations in men that were higher compared to women (mean: 9 mmol/L; median: 10 mmol/L). Bader et al.(29) reported mean CR concentrations of 12.9 mmol/L for men (median: 12.1 mmol/L) and 9.9 mmol/L for women (median: 8.8 mmol/L), while mean concentrations observed by Carrieri et al.(13) were 15.7 mmol/L and 13.0 mmol/L for men and women, respectively. Taken together, these results suggest that median CR concentrations of 14.4 mmol/L (1.6 g/L) for men and 10.2 mmol/L (1.2 g/L) for women seem to be representative of a workers’ population. These results are in agreement with those of Alessio et al.(30) and Sieniawska et al.(31) who suggested that different biological monitoring values should be considered for men and women when CR adjustment is used to correct spot urine concentrations for urine output fluctuation. Regarding the correlation of
Relative Effects of Sex and Collection Period on Creatinine and Specific Gravity Levels Specific gravity (−1)A
Creatinine Period
Men
Women
Men
Women
Pre-shift Post-shift Other
ReferenceB 99 (98– 100)C 104 (103– 105)
71 (70– 73) 72 (71– 74) 68 (67– 70)
Reference 105.2 (104.6– 105.9) 104.6 (103.7– 105.5)
84.7 (83.1– 86.3) 89.9 (88.4– 91.4) 89.3 (87.7– 90.8)
AThe
value of 1 was subtracted from the specific gravity measurements prior to model fitting. level taken as 100%. CRelative effect (%) and 95% confidence interval. BReference
Journal of Occupational and Environmental Hygiene
February 2015
127
Downloaded by [Michigan State University] at 06:59 23 February 2015
CR measurements for a given worker, Cocker et al.(21) reported σ w and σ b values of 0.49 and 0.35, respectively, corresponding to an intra-worker correlation (ρ) of 0.34 which is virtually identical to the value of 0.33 observed in our study. The comparison of biological monitoring results from different studies, expressed in different units (corrected for CR or SG), requires the use of conversion factors based on a tentative average urinary CR concentration. For instance, ACGIH applies a mean CR concentration of 1 g/L.(32) Indeed, the new BEI proposed by the ACGIH (33) for urinary chromium (increase during shift) is 10 μg/L, based on the former BEI of 10 μg/g CR converted using a mean CR concentration of 1 g/L. Our results, as well as data from other studies suggest that the conversion of BEI units should use CR values in the order of 1.6 g/L for men and 1.2 g/L for women. Using these values, the corrected BEI of urinary chromium would be 16 and 12 μg/L, respectively, for men and women. Considering that the mean CR concentration in a population of workers is greater than 1 g/L, the use of a mean CR value of 1 g/L to convert BEIs expressed in g/g CR into g/L corresponds to the use of a conservative approach. The characterization of urinary SG appears to be less investigated compared to CR, especially in populations of workers. The AM of SG levels for men in our study was 1.023 (SD = 0.006, n = 17,811) which was higher than for women (1.019 ± 0.007, n = 2385). When accounting for the correction of biomarker results for SG in Eq. (1), these correspond to levels on average 15% higher in men than in women, and 5% higher at the end of the shift compared to pre-shift. Carrieri et al.(13) did not find any meaningful difference in the SG values measured at the beginning of the work shift compared to the end of the shift, with post-shift values of 1.023 ± 0.005 for men (n = 346) and 1.020 ± 0.005 for women (n = 133). Suwazono et al.(10) reported significantly lower average SG values for women (1.020 ± 0.008, n = 434) compared to men (1.023 ± 0.007, n = 289) for morning urine collection. The values for post-shift urine collection reported in the same study were 1.015 ± 0.006 for women and 1.018 ± 0.006 for men. Our results as well as data from other studies suggest that mean SG values for men and women workers are, respectively, in the order of 1.023 and 1.020. Hence, as for CR, SG values for women are generally lower than those for men. While trends observed for sex were similar, according to the literature, the influence of sex, age, and muscle mass appears to be lower for SG than for CR.(10,12–17) In addition, SG represents an interesting alternative as its determination is simpler and less cost-intensive than for urinary CR.(17,18,34) For these reasons, some authors suggested that SG correction would be preferable to CR adjustment.(16,18) Based on results observed in our study as well as on information available from the literature, we suggest that correction based on SG should be favored in future studies involving the proposal of reference values for purposes of biological monitoring of occupational exposures. Such an approach would avoid, where possible, the transformation of 128
units of BEIs according to an estimation on mean CR or SG values, which could sometimes lead to significant bias. However, should the transformation of units be done for BEIs, this should be applied based on the average CR and SG values relevant to a population of workers, such as those reported in this study. FUNDING
F
inancial support from the Institut de recherche RobertSauv´e en sant´e et en s´ecurit´e du travail (grant no. 20100059) is gratefully acknowledged. JFS was supported by the Fonds de recherche du Qu´ebec—Sant´e (FRQS) and the IRSST. REFERENCES
1. Sorahan, T., D. Pang, N. Esmen, and S. Sadhra: Urinary concentrations of toxic substances: An assessment of alternative approaches to adjusting for specific gravity. J. Occup. Environ. Hyg. 5(11):721–723 (2008). 2. Akerstrom, M., T. Lundh, L. Barregard, and G. Sallsten: Sampling of urinary cadmium: Differences between 24-h urine and overnight spot urine sampling, and impact of adjustment for dilution. Int. Arch. Occup. Environ. Health 85(2):189–196 (2012). 3. Cone, E.J., Y.H. Caplan, F. Moser, T. Robert, M.K. Shelby, and D.L. Black: Normalization of urinary drug concentrations with specific gravity and creatinine. J. Anal. Toxicol. 33(1):1–7 (2009). 4. Fortin, M.-C., G. Carrier, and M. Bouchard: Concentrations versus amounts of biomarkers in urine: A comparison of approaches to assess pyrethroid exposure. Environ. Health 7(1):1–13 (2008). 5. Garde, A.H., Å.M. Hansen, J. Kristiansen, and L.E. Knudsen: Comparison of uncertainties related to standardization of urine samples with volume and creatinine concentration. Ann. Occup. Hyg. 48(2):171–179 (2004). 6. American Conference of Governmental Industrial Hygienists (ACGIHR ): Documentation of the Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2001. 7. Lauwerys, R., V. Haufroid, P. Hoet, and D. Lison: Toxicologie industrielle et intoxications professionnelles (Industrial toxicology and occupational intoxications). Paris: Elsevier Masson, 2007. [in French] 8. Boeniger, M.F., L.K. Lowry, and J. Rosenberg: Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: A review. Am. Ind. Hyg. Assoc. J. 54(10):615–627 (1993). 9. Barr, D.B., L.C. Wilder, S.P. Caudill, A.J. Gonzalez, L.L. Needham, and J.L. Pirkle: Urinary creatinine concentrations in the US population: Implications for urinary biologic monitoring measurements. Environ. Health Perspect. 113(2):192 (2005). 10. Suwazono, Y., A. Åkesson, T. Alfven, L. J¨arup, and M. Vahter: Creatinine versus specific gravity-adjusted urinary cadmium concentrations. Biomarkers 10(2–3):117–126 (2005). 11. Waikar, S.S., V.S. Sabbisetti, and J.V. Bonventre: Normalization of urinary biomarkers to creatinine during changes in glomerular filtration rate. Kidney Int. 78(5):486–494 (2010). 12. Heavner, D.L., W.T. Morgan, S.B. Sears, J.D. Richardson, G.D. Byrd, and M.W. Ogden: Effect of creatinine and specific gravity normalization techniques on xenobiotic biomarkers in smokers’ spot and 24-h urines. J. Pharm. Biomed. Anal. 40(4):928–942 (2006). 13. Carrieri, M., A. Trevisan, and G.B. Bartolucci: Adjustment to concentration-dilution of spot urine samples: Correlation between specific gravity and creatinine. Int. Arch. Occup. Environ. Health 74(1):63–67 (2001). 14. Moriguchi, J., T. Ezaki, T. Tsukahara, et al.: Decreases in urine specific gravity and urinary creatinine in elderly women. Int. Arch. Occup. Environ. Health 78(6):438–445 (2005).
Journal of Occupational and Environmental Hygiene
February 2015
Downloaded by [Michigan State University] at 06:59 23 February 2015
15. Nermell, B., A.-L. Lindberg, M. Rahman, et al.: Urinary arsenic concentration adjustment factors and malnutrition. Environ. Res. 106(2):212–218 (2008). 16. Dai, X., K. Kutschke, and N. Priest: Rapid urinary output normalization method using specific gravity. Health Phys. 101(2):140–143 (2011). 17. Pearson, M.A., C. Lu, B.J. Schmotzer, L.A. Waller, and A.M. Riederer: Evaluation of physiological measures for correcting variation in urinary output: Implications for assessing environmental chemical exposure in children. J. Expo. Sci. Environ. Epidemiol. 19(3):336–342 (2009). 18. Miller, R.C., E. Brindle, D.J. Holman, et al.: Comparison of specific gravity and creatinine for normalizing urinary reproductive hormone concentrations. Clin. Chem. 50(5):924–932 (2004). 19. Watson, A.: Urine specific gravity in practice. Aust. Vet. J. 76(6):392 (1998). 20. Chadha, V., U. Garg, and U.S. Alon: Measurement of urinary concentration: A critical appraisal of methodologies. Pediatr. Nephrol. 16(4):374–382 (2001). 21. Cocker, J., H. Mason, N. Warren, and R. Cotton: Creatinine adjustment of biological monitoring results. Occup. Med. 61(5):349–353 (2011). 22. Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail (IRSST): D´etermination de la cr´eatinine urinaire (Determination of urinary creatinine). Analytical method 163-2. Montr´eal, Quebec, Canada: IRSST, 1996. [in French] 23. Gagn´e, S.: Guide de pr´el`evement des e´ chantillons biologiques (Biological samples collection guide). Technical Report T-25. Montr´eal, Quebec, Canada: Institut de recherche Robert-Sauv´e en sant´e et en s´ecurit´e du travail (IRSST), 2013. [in French] 24. Carr, D.B., R.J. Littlefield, W.L. Nicholson, and J.S. Littlefield: Scatterplot matrix techniques for large N. J. Am. Stat. Assoc. 82(398):424–436 (1987).
25. Minnotte, M.C., S.R. Sain, and D. Scott: Multivariate Visualization by Density Estimation. In Handbook of Data Visualization, C.-H. Chen, W. H¨ardle, and A. Unwin (eds.). London: Springer, 2008. pp. 389–413. 26. Peretz, C., A. Goren, T. Smid, and H. Kromhout: Application of mixedeffects models for exposure assessment. Ann. Occup. Hyg. 46(1):69–77 (2002). 27. Rappaport, S.M., and L.L. Kupper: Quantitative Exposure Assessment. El Cerrito, Calif.: Stephen Rappaport, 2008. 28. Lavou´e, J., R. Vincent, and M. G´erin: Statistical modelling of formaldehyde occupational exposure levels in French industries, 1986–2003. Ann. Occup. Hyg. 50(3):305–321 (2006). 29. Bader, M., P. Messerer, and W. Will: Urinary creatinine concentrations in an industrial workforce and comparison with reference values of the general population. Int. Arch. Occup. Environ. Health 86(6):673–680 (2013). 30. Alessio, L., A. Berlin, A. Dell’Orto, F. Toffoletto, and I. Ghezzi: Reliability of urinary creatinine as a parameter used to adjust values of urinary biological indicators. Int. Arch. Occup. Environ. Health 55(2):99–106 (1985). 31. Sieniawska, C.E., L.C. Jung, R. Olufadi, and V. Walker: Twenty-fourhour urinary trace element excretion: Reference intervals and interpretive issues. Ann. Clin. Biochem. 49(4):341–351 (2012). 32. American Conference of Governmental Industrial Hygienists (ACGIHR ): Documentation of the Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2009. 33. American Conference of Governmental Industrial Hygienists (ACGIH): Threshold Limit Values and Biological Exposure Indices. Cincinnati, Ohio: ACGIH, 2010. 34. Haddow, J.E., G.J. Knight, G.E. Palomaki, L.M. Neveux, and B.A. Chilmonczyk: Replacing creatinine measurements with specific gravity values to adjust urine cotinine concentrations. Clin. Chem. 40(4):562–564 (1994).
Journal of Occupational and Environmental Hygiene
February 2015
129
