Effectiveness of federally required medical laboratory screening in the detection of chemical liver injury.

The increasing concern of industrialized societies over the potential health hazard of synthetic chemicals in the occupational environment has led to government requirements for medical laboratory screening of workers. The specific tests for such screening programs are most often selected on the basis of medical experience which utilized them in symptomatic or hospitalized populations. Required screening tests for hepatic injury including cancer in vinyl chloride workers has been systematically and prospectively studied in an industrial population working with synthetic rubber and plastics. Approximately 1300 employees were studied over a five-year period. A cohort of 969 male employees, for the purposes of analysis, were divided into a “standard” and “nonstandard” population based upon the absence or presence of significant medical disease (including liver disease). A subcohort of 120 individuals was further identified based on availabiliity of liver biopsy. Evaluation of federally required studies included alkaline phosphatase (AP), γ-glutamyl transpeptidase (GGTP), alanine aminotranserase (ALT, SGPT), aspartic aminotransferase (AST, SGOT) and bilirubin (BR). Also studied were indocyanine green clearance (ICG) and radioisotopic liver spleen scans (L-S scans). The GGTP provided the highest positive predicted value as a screening test for identifying “nonstandard” individuals (individuals with all types of medical disease) followed by ICG, AST, ALT, L-S scan, AP, and BR. In the identification of asymptomatic liver disease the GGTP had the least specificity due to a high false positive rate, while the AP provided the highest specificity. The ICG clearance however, provided the best combination of positive predictive value and sum of specificity and sensitivity. The AP provided additional increase in specificity as a follow-up study. There was no evidence that any of the other federally required tests added any additional benefit and did add significant increase in the false positive rate. These studies support the need for evaluating screening tests as to their sensitivity, specificity and positive predictive value, in asymptomatic individuals, before they are made established requirements.


Introduction
Industrialized societies throughout the world have become increasingly concerned over the potential health hazard of synthetic chemicals in the occupational environment. Governmental regulations have increased the number and types of medical laboratory screening required for a large variety of halogenated hydrocarbons as well as other potential environmental hazards. The primary objective of these screening programs is to reduce disability, morbidity and mortality in workers, especially as related to serious low-grade health hazards. In general, screening programs are instituted because of the presence in the work environment of a suspected or proven environmental toxin or carcinogen, which has the potential of producing low-grade injury over long periods of exposure.
Most screening studies are directed toward the detection of abnormalities in certain body systems. The specific tests are frequently selected on the basis of medical experience which utilized them in symptomatic or hospitalized populations. Prior experiences utilizing nonspecific multiphasic health surveillance screening and maintenance have not proven to be cost effective except under certain limited October 1981 117 conditions (1). The cost effectiveness of such tests, however, in the determination of medical screening requirements, has played a limited role due to the potential seriousness of these occupational agents. Little attention has been paid as to whether the effectiveness of federally required screening provides the best or, more importantly, a necessary benefit when applied to asymptomatic and otherwise healthy worker populations.
The discovery in 1974 of hepatic toxicity and cancer formation in vinyl chloride workers provided the opportunity to systematically and prospectively study the effectiveness of federally required and federally recommended medical screening procedures for the detection of chemical liver injury, including cancer development (2). Table 1 lists the federally required medical screening procedures since 1974 for environments utilizing vinyl chloride or polyvinyl chloride. Table 2 lists the federally recommended studies for these same environments. This paper will present a preliminary assessment of the effectiveness of these federally required studies in the accurate detection and identification of chemically induced liver injury due to halogenated hydrocarbons, especially vinyl chloride.

Materials and Methods
The industrial population studied consisted of approximately 1200-1400 employees of a chemical plant whose two major products were synthetic rubber and plastics. The industrial plant had been in operation for over 35 years and had a predominance of male employees (96%), approximately 80-87% of the work force being white, 11-12% black, less than 1% of other racial origins. Turnover of the plant was approximately 10 to 15% per year with 65-70% of the work force having worked five years or more at the plant. Employee ages ranged from 18-65, with a mean of 52 years.
A cohort consisting of 969 male employees who worked continually from June 1, 1976 to May 31, 1977 was, for purposes of this analysis, divided into a "standard" and a "nonstandard" population. These designations were given on the basis of a review of  all present standard medical data on each employee, including the federally required studies. Other screening studies of the medical surveillance programs were not utilized in the classification of overall medical status because, at that time, their clinical usefulness was unknown or controversial. All studies were performed on an annual basis; those individuals with ten years or more of employment were examined and screened semiannually.
Compliance with medical screening studies during the five-year study period showed a continuous participation in the history and physical examinations by over 75% of the work force, laboratory tests and chest x-rays by 86%, and liver-spleen scans by 85%. Seventeen percent failed to undergo at least one history and physical examination, 9% did not have any of the radiological studies, and only 4% failed to have laboratory studies during this period. Approximately 40-50% of these individuals who did not undergo an examination claimed to have been examined by their private physician. A subcohort of 120 individuals was further identified based on the availability of a liver biopsy performed for medical reasons, both related and not related to their work.
The term "standard" is used for those individuals who, based upon the best medical opinion, demonstrated no evidence of any significant medical disease, occupational or nonoccupational in origin. The "nonstandard" population included all others not included in the standard population.
The subcohort population was divided into those individuals with and without histological evidence of liver injury and further subdivided into those with and without histological features characteristic of chemical injury.
All employees had individual work histories. These consisted of a standardized job classification for all jobs within the plant since its opening and a Environmental Health Perspectives rank ordering of exposure for 22 different suspected or potentially hazardous heptatoxic chemicals used within the work place (3)(4)(5). The agents were rank ordered on the basis of the intensity of exposure for each of the job classifications for each of the years that the plant was in operation. From this detailed work history, a cumulative exposure ran month ration (CERM) was determined for each employee for each of the 22 chemicals. All histological material was classified as to the presence or absence of liver disease, and to whether the abnormalities were consistent with chemical or nonchemical injury. This classification was conducted double blindly by three experienced physicians, two pathologists, and a hepatologist (6)

Results
Although 50 or more biochemical 4 were performed during this stud; paper will limit itself to the eva federally required studies, the indc clearance (ICG) study at the 0.5 mg and radioisotopic liver and spleen ,/kg dose (7,8) The positive predictive values of these screening scan (9). The tests in identifying medical disease (including liver disease) in this asymptomatic working population are shown in Figure 1. and sum values shown in Figure 3 in the biopsied subpopulation. Here again, y-glutamyl transpeptidase and ICG clearance (0.5 mg dose) show the greatest sensitivity for identifying individuals with liver disease. However, GGPT had the least specificity, reflecting its high incidence of false positives. Specificity increased with the use of AST, ALT, and ICG clearance. The alkaline phosphatase provided the highest specificity, suggesting that mild or low grade chronic hepatic injury due to environmental agents may be activating hepatic AP synthesis in the absence of biliary tract obstruction or cholestasis. The ICG clearances, even at the low dose (0.5 mg/kg), clearly remains the test with the best combined sensitive and specific screening study for detection of individuals with subclinical hepatic disease. This subcohort biopsied group was further examined on the basis of the histological interpretation of their liver biopsies and their work exposure to vinyl chloride. All biopsied individuals were subdivided into three groups: 19 with histological evidence consistent with chemical liver injury; 30 with histological evidence of liver disease, nonchemical liver injury; and 29 with normal liver biopsies. Each of the histological subgroups were further subdivided based on their vinyl chloride exposure, on a scale of 1 to 4 (Fig. 4).
The chemical liver injury group contained the highest percentage of individuals with the highest average rating (CERM) for vinyl chloride exposure. In contrast, with those with liver disease, nonchemical, and those with normal livers have a 120  more even distribution of individuals relative to their degrees of vinyl chloride exposure.
In our previous studies we noted that almost all individuals with histologically specific lesion of -vinyl chloride injury or angiosarcoma had a total average CERM rating of 3.5 or greater. The asterisk in Figure 3 indicates the percentage of individuals in each of the three histological groups with exposure ratings of 3.5 or greater. Again, the chemical liver injury group have the highest percentage of individuals with the high exposure ratings. This further supports previous work (4,10) identifying focal hepatocellular hyperplasia as the earliest histological characteristics ofchemical injury in liver disease.
A study of the frequency with which these tests are positive among those individuals with liver disease, based on their histological findings (chemical versus nonchemical), provides additional data supporting the clinical observation that an increased AP has a greater specificity for chronic liver injury. Figure 5 shows the ratio of the proportion of positive screening tests in those with histological chemical liver disease divided by the proportion of positive tests in those whose disease is not of chemical origin. All tests, independent of their sensitivity and specificity for liver injury, were more frequently abnormal in the presence of nonchemical, subclinical liver injury, except for AP. In contrast, AP was far more frequently abnormal in those individuals with chemical liver injury, which tended to be more chronic than acute and generally less severe.
Environmental Health Perspectives

Discussion
This preliminary systematic tive predictive values and specificity of federally and required tests for chemical v first scientific and biological b of medical screening tests for tional environments. Althou used medical tests have be experience to be effective as d symptomatically ill or hospital clinical work has been done ability to accurately separate from early latent or underlyii tomatic individuals. Tests whi false-positive rates (decrease( GGTP, interfere with the scre fication of the high risk work and cost required for repeat t productivity for the emplo3 increased anxiety, and by the the effectiveness of the testi employees and employer. D sensitivity and specificity of i asymptomatic individuals is recommendations are to be mz ment. This evaluation procee October 1981 studies are found to be abnormal, an AP should be done and a diagnostic work-up instituted to deter-| | | | | | mine the etiology (11).
The rationale for these recommendations is based on the actual study of chronic subacute chemical GGTP AP ICG injury in an asymptomatic population, not preselected 0.5 because of signs or symptoms. Therefore the test's ability to correctly differentiate disease from STS nondisease or one type of injury from another is chemical tests were abnor-more accurately determined. Chemical and envif hepatic injury expresses ronmental agents of low toxicity tend to produce njury, (LD) liver disease, repeated or persistent injury which accumulates over time. Tests which measure overall functional capacity quantitatively or semiquantitatively, rather than measuring acute low-grade injury over time are more likely to detect changes. For this reason, clearance or tolerance studies provide the best review of the posimeans for identifying latent hepatic disease, while the sensitivity and enzyme studies like ALT, GGPT, and SGOT ususome non-federally ally reflect acute cellular injury of higher grade or vorkers provides the degree and cannot accurately reflex accumulative asis for the selection damage until very late in the disease process. Tests Liver injury in occupa-which provide information concerning the progresgh these commonly sion or nonprogression of injury will be far more en found by clinical helpful to the practicing occupational physician. liagnostic tools in the They provide him/her with a better capability to Lized population, little discern between nonoccupational and occupational to determine their disease, and the best available reassurance for the biological variations worker of his or her safety while allowing the ng disease in asympgreatest possibility for continued productivity and ich provide very high employment. d specificity) such as Finally, these data provide a sound scientific hening process identi-basis upon which to modify federal requirements.
;er by the extra time The removal of specific testing requirements which, esting, the decreased with field experience, prove not to have any significant yer, the employees' positive predictive value, or effective sensitivity loss of confidence in and specificity will aid in reducing overall cost and ,ng program by both help maintain continued compliance by industry etermination of the and workers. There may be theoretical reasons to screening studies for maintain or continue some of the present federally essential if effective required screening studies, but these reasons should ade a federal require-be separately identified and not be confused with ss also provided the the purposes of the more effective test in the 121 detection of occupationally related liver injury.
The capabilities and limitations of any new tests or old tests for new screening purposes in detection of other potential occupational hazards, should be validated before making them a required screening procedure. This would lay the foundation for systematic determination of effectiveness of screening procedure against a proven standard. Unvalidated federal requirements provide the worker and employer with a false sense of security and safety by what is believed to be effective monitoring. More importantly, such a situation can lead to delay in effective correction of cause and disease prevention.

Conclusions
Biochemical screening for hepatic injury in asymptomatic chemical workers can be done most effectively by the use of liver specific clearance studies.
ICG clearance provides the best combination of positive predictive value and sensitivity and specificity for functional hepatic testing at the present time.
None of the present federally required studies provide any significant degree of sensitivity without marked reduction in specificity in asymptomatic individuals.
The ALT (SGOT) is the most useful among those Federal tests presently required and the alkaline phosphatase may provide additional specificity as a follow-up study in those individuals with positive ICG or ALT screening studies. There is no evidence that any of the other federal studies add any benefit and strong evidence that they significantly increase the false-positive results in well individuals.
All screening studies should undergo evaluation as to their positive predictive value and sensitivity and specificity in asymptomatic individuals before becoming permanent or established requirements.