Barbanel, Cheryl S. MD, MBA, MPH; Winkelman, James W. MD; Fischer, George A. PhD; King, Andrew J. MD
The Department of Transportation (DOT) routinely screens the urine of workers for metabolites of marijuana, cocaine, opiates, and for amphetamines and phencyclidine. 1 Specimen validity testing is the evaluation of the specimen to determine if it is authentic, unaltered human urine. The purpose of validity testing of urine samples is to detect tampering that would invalidate drug tests. The testing determines whether certain adulterants or foreign substances were added, if the urine is diluted, or if the specimen was substituted with another substance. Adulteration of a specimen is the addition of a substance into urine that could mask or destroy a drug or drug metabolite and render it undetectable. Substituting a specimen is the provision of some other liquid in place of a urine specimen. 2,3
The DOT Regulations (49 Code of Federal Regulations Part 40) defines a urine specimen as substituted if it has a creatinine concentration of ≤5 mg/dL and a specific gravity of ≤1.001 or ≥1.020. On February 14, 2000, the Health and Human Services Substance Abuse and Mental Health Services Administration (SAMHSA) published an analysis of the scientific literature, 4,5 on which their definition of a substituted specimen was based. 6,7 This analysis included 51 references and over 40 research studies that examined clinical data in a variety of serious conditions associated with overhydration, including water-loading studies in healthy subjects. Studies included individuals with pathologic conditions associated with dilute urine, including psychogenic polydipsia, water intoxication, diabetes insipidus, nephrogenic diabetes, and iatrogenic diabetes. None of the specimens met the criteria for a substituted specimen.
The SAMHSA analysis was criticized on the basis of the limited number of paired measurements of both creatinine and specific gravity from single urine specimens and because the samples were primarily from men. The SAMHSA analysis was also criticized because it did not specifically cover certain employee subgroups. It was claimed that certain individuals whose urine drug tests were reported as substituted were not drug users but rather had consumed large quantities of water over a long work period and were often small-framed minority-group women. It was suggested that a combination of these factors, low muscle mass and dilution from high fluid intake, could have resulted in the natural, innocent production of urine that met the substitution criteria. 2
Substituted specimen values under challenge are those at the low end of the criteria, so that is the focus of our study. To resolve the controversy over the substituted specimen criteria, we asked: Is it possible for normal subjects, or even patients with pathologic conditions, to produce a urine specimen that has a creatinine concentration of ≤5 mg/dL and a specific gravity of ≤1.001? Could certain clinical situations result in a patient’s meeting the lower limits of the substituted criteria range? How does the percentage of urine samples meeting the substituted specimen criteria in a hospital laboratory database compare with the percentage meeting the substituted specimen criteria in laboratories performing urine drug tests under the DOT regulation? Could racial or ethnic differences account for creatinine and specific gravity measurements that meet the substituted specimen criteria?
We reviewed the results of 803,130 random urine specimens measured for creatinine and/or specific gravity, from a large tertiary care academic medical center laboratory, from February 1989 to June 2001. In this database, we identified 13,467 urine specimens on which both specific gravity and creatinine measurements were performed. There was no workplace drug testing performed in this laboratory and therefore no motivation for any patient to engage in substitution. Nevertheless, because these patients were tested for a variety of reasons (eg, screening, illness), we would expect a greater probability of meeting the DOT substituted specimen criteria compared with a healthy working population.
Results of all random urine specimens that underwent either specific gravity or creatinine testing at the Brigham and Women’s Hospital in Boston, from February 1989 to June 2001, were downloaded into a database for further analysis. The downloaded data consisted of 816,597 files that included patient specimen numbers, test results, age, sex, race, and keys to patient diagnosis. All patient identification data were carefully encoded so that patient confidentiality would be absolutely ensured. Each test result created an individual file. The above database was screened for urine specimens on which both specific gravity and creatinine measurements were performed, yielding 13,467 matches. There were 816,597 files with individual results for urine creatinine and specific gravity from a total of 803,130 random urine samples tested, which yielded 13,467 paired specimens, tested for both creatinine and specific gravity from single urine samples. Each specimen subjected to “paired” testing created one additional file (ie, 803,130 + 13,467 = 816,597).
All of the specimens in this study were random urine samples. Creatinine was measured by the modified Jaffe method, and specific gravity was measured by automated refractometry. 8,9 Instrument calibration for both tests occurs on a daily basis, with two levels of quality control on each shift.
The following instruments measured creatinine: Technicon DAX 96 (Bayer Corp, Diagnosties Division, Tarrytown, NY), Dade RXL (Dade Behring, Inc, Deerfield, IL), and Olympus AU2700 (Tokyo, Japan). Specific gravity was measured by three generations of Bayer Diagnostics (Tarrytown, NY) “Clinitek” instruments, with a manual refractometer as backup.
Urine specific gravity results were presented by the refractometer to three decimal places. Urine creatinine results were presented by the instrument in whole numbers, so that a result of 5.4 mg/dL would be rounded off to 5.0 mg/dL, and a result of 5.5 mg/dL would be recorded as 6.0 mg/dL.
The demographic distribution by age, sex, and race of the 13,467 paired urine specimens is shown in Table 1. There were 6933 female and 6534 male subjects and 9289 whites, 2340 blacks, 821 Hispanics, 207 Asians, 202 other, and 608 whose racial identity was not provided. “Other” included, for example, Native Americans and patients of mixed race. The data were further divided into age groups for the working population. Mean, standard deviation, and highest and lowest values for each analyte were computed for each group. For any specimens meeting either the specific gravity or creatinine values for the DOT substituted specimen criteria, medical information was obtained for the age, sex, race, and medical diagnosis or health status.
Comparisons of male and female subjects were performed using a two-sample t test, whereas comparisons among different races and ethnicity were done using analysis of variance with the Tukey-Kramer method for adjustment of multiple comparisons. Results with P value < 0.05 were considered significant.
We were able to review 13,467 paired laboratory results for creatinine and specific gravity from the same specimen. These data were from a large academic medical center’s clinical laboratory database after a search for results that met either of the threshold values that define the criteria for a substituted specimen. Not one paired specimen met the lower limit specific gravity (≤1.001) and creatinine (≤5 mg/dL). Eleven paired specimens met the DOT criteria for substitution with creatinine (≤5mg/dL) and the higher specific gravity criteria of ≥1.020.
The mean urinary creatinine concentration was 89 mg/dL ± 71, the median was 71 mg/dL, and the range was 1 to 1130 mg/dL. Urine specific gravity measurements had a mean of 1.016 ± 0.007, a median of 1.015, and a range of 1.001 to 1.120. Ninety-one results for either urine creatinine or specific gravity in the total 13,467 paired specimen database that included results from all age groups, from newborn to >80 years of age, met one of the criteria set by the DOT for creatinine and the lower limit of specific gravity for a substituted specimen. The DOT criteria met was heavily weighted on the side of urinary creatinine, with 87 creatinine measurements from 83 patients ≤5 mg/dL and only 4 specific gravity measurements ≤1.001. Ten percent of the specimens exceeding one of the criteria occurred in newborns, and about 29% occurred in those 70 years or older, leaving roughly 60% in the working-age group. The demographics of patients whose specimens exceeded one criterion are provided in Table 1.
The number of paired specimens in the working population age group, which we defined as 18 to 65 years of age, was 8107 (Table 2). In the 18- to 65-year age group, 4085 specimens were from female donors and 4059 were from male donors. The mean for specific gravity in this working-age group was 1.016 ± 0.0075, with a range of 1.001 to 1.084. The mean creatinine concentration in this age group was 91 mg/dL ± 75.9, with a low of 1 mg/dL and a high of 1130 mg/dL.
The mean creatinine value for all female subjects from the 18- to 65-year age group was 82 mg/dL ± 69.1, with a range of 1 to 674 mg/dL (Table 2 and Fig. 1, A). Urine specific gravity measurements had a mean of 1.016 ± 0.0075, with a low of 1.001 and a high of 1.084 (Fig. 1, B). Race comparisons of mean specific gravity in women in the 18- to 65-year age range results were: Asian, 1.012; black and white, 1.016; and Hispanic, 1.017. Asian women had significantly lower specific gravity than all other groups (P = 0.0001). Mean creatinine values in women by race were: Asian, 57 ± 48 mg/dL; white, 73 ± 62 mg/dL; Hispanic, 97 ± 77 mg/dL; and black 100 ± 80 mg/dL (Fig. 2, above). Note that the values for Asian women were significantly lower than for whites (P < 0.05), blacks, and Hispanics (both P = 0.0001). Furthermore, the results for whites were significantly lower than for blacks (P = 0.0001) or for Hispanic women (P = 0.0001).
In men in the 18- to 65-year age group, the mean creatinine value was 99 mg/dL ± 81, a value significantly higher than that seen in the women (82 ± 69;Table 2 and Fig. 1, A). Creatinine results ranged from 1 mg/dL to 1130 mg/dL. Urine specific gravity measurements had a mean of 1.017 ± 0.0074 and a range of 1.001 to 1.066, which was significantly higher than those in the women (Fig. 1, B). Comparisons of mean specific gravity varied little in men based on race in the 18- to 65-year age group, ranging from 1.016 to 1.018. Mean specific gravity for white men was 1.016; for black men, 1.017; and for Hispanic and Asian men, 1.018. Mean creatinine values in men by race were: Asian 99 ± 71 mg/dL; white, 92 ± 72 mg/dL; Hispanic, 109 ± 80 mg/dL; and black 123 ± 110.7 mg/dL (Fig. 2, below). Blacks had significantly higher creatinine concentrations compared with whites (P = 0.0001) and Hispanics (P < 0.05) but not compared with Asian male subjects (P = 0.2).
Of the 8107 specimens in the working-age group, 43 (0.53%) individual creatinine and 4 (0.05%) individual specific gravity measurements were at the low end of the substituted range (Table 2). No specimens met both of these criteria. Of the 43 individuals with creatinine measurements <5 mg/dL, 25 (0.62%) were in specimens from female subjects and 18 (0.44%) were from male subjects, suggesting that women are at slightly higher risk of meeting the criteria. Of the 25 creatinine measurements in the substituted specimen range in female subjects, 22 (0.91%) were in specimens from white women, 1 (0.11%) from a black woman, 2 (0.62%) from Hispanic women, and none from Asian women. The races of the 18 male subjects with creatinine concentrations in the substituted range were: 16 (0.59%) white, 0 black (0%), 1 Hispanic (0.32%), 0 Asian (0%), and 1 other (1.09%). Of the four specimens with specific gravity ≤1.001, two were from female subjects, one Asian and one white, and two were from white male subjects.
The medical diagnoses and demographics of the patients with low specific gravity or creatinine concentrations are presented in Table 3. Note that of the patients who met the urinary creatinine criteria, 10 were newborns and 29 died during that hospitalization. As expected, those with urinary creatinine concentrations ≤5 mg/dL were those most likely to have a very low muscle mass, such as newborns. All the rest of the patients were severely ill with diagnoses that would make them very unlikely to be able to participate in the workforce. Their diseases belonged to certain major categories, all of which are associated with muscle atrophy, including malignancy, renal failure, cardiovascular disease, metabolic disease, and other. The number of patients meeting the specific gravity criteria was too small to make a general statement other than that all cases were associated with overtly ill or highly debilitated persons.
Table 3, section C, identifies 11 individuals who met the DOT’s criteria for a substituted specimen with a creatinine concentration of ≤5 mg/dL and a specific gravity of ≥1.020. These individuals were identified by first searching for creatinine concentrations ≤5 mg/dL and then screening for those with creatinine concentrations ≥1.020. All of these individuals were extremely ill. Six of these 11 patients died during the admission. Eight were 69 years or more in age. Eight of the 11 patients had a diagnosis of cancer. The diagnoses of the three without cancer, who all were among those who died during the admission, included pneumonia; renal insufficiency, congestive heart failure, and diabetes; and diabetes mellitus and urinary tract infection in a person aged 101. Of the three in the working-age group, two died during the admission and one had a diagnosis of metastatic cervical carcinoma, dehydration, and urinary tract infection.
None of the 13,467 paired consecutive laboratory results for creatinine and specific gravity from a large academic medical center’s clinical laboratory database for the years 1989 to June 2001 met the DOT’s lower limit of specific gravity (≤1.001) and creatinine (≤5 mg/dL) criteria for a substituted specimen. These data strongly support the proposition that the criteria established by the DOT for the lower limit of specific gravity and for creatinine required for a substituted specimen are reasonable. The data must be interpreted with some caution in this retrospective study because we did not know the specific indications used by the ordering physicians. Indeed, ordering both tests simultaneously heightens the chance that these persons presented with a renal or electrolyte disorder. Nevertheless, as no patient fulfilled the criteria, it is extremely unlikely that urine produced by normal working individuals would achieve the DOT criteria for a substituted specimen with a specific gravity of ≤1.001 and a creatinine concentration of ≤5 mg/dL.
Eleven of 13,467 paired specimens met both of the criteria for a DOT substituted specimen by having a creatinine concentration of ≤5 mg/dL and meeting the upper limit for specific gravity of ≥1.020. In the drug-testing population, this is usually the result of substituting urine with another substance besides water. Severely ill persons produced these 11 specimens. Severe cachexia causing low muscle mass likely resulted in very low creatinine concentrations in the urine, whereas dehydration produced high specific gravity. None of these patients was likely to have been in the workforce at the time they produced their urine specimens.
In the working-age group (18 to 65 years), there were 8107 specimens, with 4085 from female subjects. Mean concentrations of creatinine (Fig. 2, above) and urine specific gravity were lower in Asian women as compared with other racial and ethnic groups. This variation among ethnic or racial groups was likely due to differences in lean body mass and water and meat intake. The mean urinary creatinine concentrations for women in the four racial groups identified ranged from 57 mg/dL to 100 mg/dL, with large standard deviations suggesting wide-ranging differences in lean body mass and water intake among individuals within a group. Although the mean creatinine concentration for Asian women was the lowest, in fact, no Asian women had a concentration ≤5 mg/dL. The group with the greatest percentage meeting the creatinine criteria for substitution was white. Meeting the specific gravity criteria was so rare that no inferences can be made regarding the role of race or ethnic groups. Men had a mean creatinine value of 99 ± 81 mg/dL as compared with women, with a mean creatinine concentration of 82 ± 69 mg/dL (P = 0.0001) (Fig. 1, A). Race and ethnicity in men were also influential on urinary creatinine concentrations, with black male subjects having higher values. These differences are mostly driven by variance in lean body mass, though they may also reflect differences in diet (water and meat). Specific gravity varied little among the male groups.
Data from a recent study demonstrated that healthy individuals did not achieve the federal criteria for a substituted urine specimen, even after intentional water loading. 10 The DOT sponsored a water-loading study to address concerns about the criteria for validity testing. Fifty-four participants, 13 male and 43 female, completed the study, producing 500-paired specimens tested for creatinine and specific gravity. Each subject drank at least 80 ounces of fluid within 6 hours. Only five men and seven women were even able to drink over 1 gallon of fluid by the end of the 6-hour test period. None of their specimens met the criteria for a substituted specimen.
Creatinine excretion rates have been measured for both men and women aged 20 to 60 years of age (Table 4). 11,12 Increasing age is associated with a decline in renal creatinine excretion rate, mostly due to normal reduction in lean body mass with aging. Likewise, the larger muscle mass of men accounts for the greater creatinine production and urinary excretion as compared with women. To provide insight into the magnitude of water ingestion required to achieve a urinary creatinine concentration as low as 5 mg/dL, we computed creatinine production rates for a healthy 60-year-old women weighing only 40 kg (88 lb). Her advanced age, gender, and size would be expected to lead to a creatinine excretion rate in the lowest range, thereby increasing the likelihood of achieving a level of ≤5 mg/dL.
Using the normal values of creatinine excretion (Table 4), the 60-year-old, 40-kg woman will excrete 15 mg/kg/day, for a total of 600 mg/day (15 mg/kg/day × 40 kg = 600 mg/day). If she excretes 600 mg of creatinine, to produce a urine creatinine concentration of 5 mg/dL in urine, how much urine would she need to produce (X) (or how much water must she drink) in a day? The calculations are: 600 mg/X = 5 mg/dL, X = 120 dL/day, 120 dL/day = 12 L per 24 hours! The average urine volume is between 750 and 2500 mL/24 hours. The approximate 24-hour fluid volume intake in women required to produce a urine concentration of 5 mg/dL is calculated on the average excretion rates based on age in Table 4, which also provides other examples of men and women of various ages and sizes.
It is noteworthy that the amount of water ingestion needed to fulfill the creatinine criteria exceeds the capacity of the kidney to excrete water and, thus, would be associated with development of progressive hyponatremia. If the woman in the above example were on a low-solute diet (low-salt, low-protein, anorexic), her vulnerability to hyponatremia would be dramatically increased. The pathologic condition associated with this degree of water ingestion and normal renal water excretion is termed psychogenic polydipsia. Typically, these persons are schizophrenic with overt behavioral abnormalities. Progressive hyponatremia is associated with the development of gastrointestinal symptoms and, ultimately, neurologic deterioration (lethargy, seizures).
With regard to the urine specific gravity cutoff, the most diluted urine a person can generate is approximately 50 mosmol/kg, which translates into a urine specific gravity slightly over 1.001. It should be noted that the specific gravity compares the weight of the fluid with that of water, which is given a value of 1.000. A specific gravity of 1.001 is equivalent to approximately 35 to 40 mosmol/kg, which is below maximal dilution. 13 Many references have been identified in the National Laboratory Certification Program Document 4 that corroborate this lower limit of urinary dilution. A normal person must be drinking inordinate amounts of water to achieve a specific gravity in this range. Even with a rapid and large-volume infusion of water, one cannot decrease urine osmolality to <50 mosmol/kg; however, plasma sodium concentration can be lowered abruptly with massive water ingestion. Note that advancing renal disease is typically associated with impairment of urinary diluting capacity, ultimately resulting in isosthenuria, which is reflected by a urine specific gravity of approximately 1.010.
Theoretically, to fail to meet the specific gravity and creatinine criteria simultaneously would require extreme muscle wasting (cachexia) and ultrahigh water intake. Assuming reasonable bladder capacity, an individual would need to urinate 30 to 40 times a day, including while sleeping, a state that would make working anywhere difficult. In our opinion, the confluence of factors required to reach these criteria without altering the urine are extremely rare. Furthermore, the risk of encountering this extraordinarily unusual individual must be weighed against the formidable risk of not detecting a substituted specimen. The data in this study offer corroborative support of this contention by showing that in even in the most acutely ill patient population, no person met the lower limit of specific gravity (≤1.001) and creatinine required for substitution by the DOT criteria.
The Medical Review Officer reviews all specimen results reported as substituted and interviews the employee to determine if there is a legitimate medical explanation for the result. If the Medical Review Officer believes that the employee’s explanation may present a reasonable basis for concluding the existence of a legitimate medical explanation, he or she will offer the employee an opportunity for an independent medical evaluation to demonstrate that the laboratory findings are physiologically possible and are related to a medical condition. 3
Working persons who are identified to have substituted urine are undoubtedly polyuric. The standard workup for polyuria caused by a urinary concentrating defect involves a water restriction test in which urine and plasma osmolality are monitored over time in the absence of water. If the person’s urine sample fails to become concentrated, antidiuretic hormone is infused to clarify whether there is a central or nephrogenic source of diabetes insipidus. The evaluation of the person with a urine specimen identified as “substituted” asks a different, although related, question: How dilute can this person make his or her urine? As previously discussed, the literature contains many water-loading trials but no standardized protocols. The individual who achieves DOT substituted urine criteria with urinary creatinine ≤5 mg/dL and specific gravity of ≤1.001 presumably achieved the criteria through oral ingestion of water. Thus, one potential workup would require the person to ingest water ad libitum under close observation for a period of 6 hours while the change in urinary parameters is monitored hourly. For the safety of the person, plasma osmolality would be monitored simultaneously to preclude the development of hyponatremia due to overwhelming water intake.
Quest Diagnostics Incorporated performed more than 450,000 drug tests from January to June 2001, more than 800,000 drug tests in 2000, and more than 650,000 in 1999 for the federally mandated safety-sensitive workforce, reporting a steady decline in the drug-positive rate among this group of workers from 3.5% in 1997 to 3.0% in the first 6 months of 2001. The percentage of all drug tests that were substituted in the federally mandated safety-sensitive workforce in 1999 and 2000 was 0.03%, and 0.02% in the first 6 months of 2001. According to Quest Diagnostics’ data, in this workforce population during 1999, 0.96% of all positive drug tests (ie, drug-positive, substituted, and adulterated) resulted from substitution, and the percentage decreased to 0.77% in 2000 and to 0.68% in the first 6 months of 2001 (Table 5). These data may indicate a deterrent effect of validity testing.
We conclude that a large population of persons with pathologic states failed to meet the lower limit for specific gravity (≤1.001) and creatinine (≤5 mg/dL) criteria for a DOT substituted urine specimen. These data lend strong credence to the belief that normal individuals are extremely unlikely to achieve these criteria, especially in the workplace.
We thank Lori Barbanel, Michael Moskowitz, Yolanda Rodriguez, and Helen Williams, RNC COHN-S, for their administrative and editorial support; and Phillip Stone, PhD, and Janice Weinberg, ScD, for biostatistical support.
1. US Code of Federal Regulations. Department of Transportation: Procedures for Transportation Workplace Drug and Alcohol Testing Programs; Final Rule. Washington, DC: DOT; Dec 19, 2000. Title 49, Part 40: 79539.
2. US Code of Federal Regulations. Department of Transportation: Procedures for Transportation Workplace Drug and Alcohol Testing Programs; Final Rule. Washington, DC: DOT; Dec 19, 2000. Title 49, Part 40 Preamble: 79478–79482.
3. US Code of Federal Regulations. Department of Transportation: Procedures for Transportation Workplace Drug and Alcohol Testing Programs; Final Rule. Washington, DC: DOT; Dec 19, 2000. Title 49, Part 40: 79519–79520.
4. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, National Laboratory Certification Program. State of the Science—Update # 1: Urine Specimen Validity Testing: Evaluation of the Scientific Data Used to Define a Urine Specimen as Substituted. Rockville, MD: DHHS, SAMHSA, NLCP; Feb 14, 2000.
5. Cook JD, Caplan YH, LoDico CP, Bush DM. The characterization of human urine for specimen validity determination in workplace drug testing: a review. J Anal Toxicol. 2000; 24: 579–588.
6. Program Document (PD) 035, September 28, 1998, Notice to the National Laboratory Certification Program (NLCP) Inspectors and HHS Certified Laboratories, “Guidance for Reporting Specimen Validity Test Results.” Available at: http://workplace.samhsa.gov/drugtesting
7. Program Document (PD) 037, July 28, 1999, Notice to the National Laboratory Certification Program (NLCP) Inspectors and HHS Certified Laboratories, “General Guidance/Criteria for Specimen Validity Testing.” Available at: http://work
8. Jaffe M. Ueber den Niederschlag welchen Pikrinsaure in normalen Harn erzeugt und uber eine neue Reaction des Kreatinins. Z Physiol Chem. 1886; 10: 391–400.
9. Rubini MD, Wolf AV. Refractometric determination of total solids and water of serum and urine. J Biol Chem. 1957; 225: 868–876.
10. Edgell KC, Glass LR, Caplan YH. Paired measurements of creatinine and specific gravity after water loading. Paper presented at: Society of Forensic Toxicologists, Inc; Oct 4, 2000; Milwaukee, WI. Available at: http://www.dot.gov/ost/dapc/main/waterloading.pdf.
11. Davies DF, Shock NW. Age changes in glomerular filtration rate, effective renal plasma flow, tubular excretory capacity in adult males. J Clin Invest. 1950; 29: 496–507.
12. Walser M. Creatinine excretion as a measure of protein nutrition in adults of varying age. J Parenter Enteral Nutr. 1987; 11 (suppl): 735–785.
13. Rose BD, Post TW. Clinical Physiology of Acid-Base and Electrolyte Disorders. 5th ed. New York: McGraw Hill; 2001: 412,692.