Neurobehavioral and Respiratory Findings in Jet Engine Repair Workers: A Comparison of Exposed and Unexposed Volunteers

Environmental Research Section A 80, 244—252 (1999) Article ID enrs.1998.3898, available online at http://www.idealibrary.com on

Neurobehavioral and Respiratory Findings in Jet Engine Repair Workers: A Comparison of Exposed and Unexposed Volunteers Kaye H. Kilburn Environmental Sciences Laboratory, University of Southern California, School of Medicine, 2025 Zonal Avenue, CSC 201, Los Angeles, California 90033; and Neuro-Test, Inc., P.O. Box 5374, Pasadena, California 91107 Received January 26, 1998

containing higher concentrations of manganese and chromium than mild steel. A special instance of stainless steel welding indoors (Kilburn and Warshaw, 1989) is the repair of the fans and intake cylinders of jet engines. Manganese has well-known neurotoxicity (Cotzias et al., 1971). Productive cough, shortness of breath on exertion, and chest tightness with wheezing were complaints of jet engine repair workers at Tinker Air Force Base in Oklahoma City. Most worked as welders, grinders, and nozzle sanders and wore respirators but did not use air supplies. Air monitoring data showed that concentrations of chromium, nickel, cobalt, and strontium chromate were elevated in some workers’ personal samples, but area samples were above PELs (permissible exposure limits) only for strontium chromate, trichlorofluoroethane, trichloroethane, trichoroethylene, and methanol (Table 1). Levels of aluminum, beryllium, copper, molybdenum, titanium, and zinc were below PELs. Exposures other than welding included vibration from grinding stainless steel, noise, acetylene, radiation from nickel—thorium, and chlorinated and fluorinated solvents used to clean metals. Metal aircraft parts were cleaned by immersion in heated solvents including freon R-112 and R-113 (trichlorotrifluoroethanes), so effects of solvent exposures were evaluated. Solvents used frequently besides freons included toluene, tetrachloroethylene (PERC), trichloroethylene (TCE), 1,1,1-trichloroethane, ethanol, methanol, ethyl acetate, and acidic sodium chromates. Adhesives that contained 1,1,1-trichloroethane were also used. The neurotoxicity of these metal-cleaning solvents has been studied for over 20 years (Hanninen et al., 1976; Maizlish et al., 1985; Arlien-Soborg, 1992). The workers sought evaluation of their respiratory problems and engaged a law firm. This initial study plan focused on pulmonary effects of metal dusts although they measured below

Workers repairing jet engines had respiratory, rheumatic, and neurobehavioral symptoms. They had welded and ground stainless steel parts using hard metal tools and cleaned metal with chlorinated and fluorinated organic solvents. We compared 154 workers and 112 unexposed subjects, all volunteers of similar ages and with similar educational levels, for abnormalities on chest radiographs, spirometric measurements, and questionnaires. Also appraised were performance of reaction time, balance, blink reflex latency, color discrimination, Culture Fair, vocabulary, slotted pegboard, trail making A and B, profile of mood states (POMS), and frequencies of 35 symptoms. Compared to unexposed subjects, workers had significantly more respiratory symptoms but no differences in pulmonary function. They had significantly prolonged simple and choice reaction time (P < 0.0001), and abnormal balance with eyes open and eyes closed (P < 0.0001), and abnormal color discrimination. Blink reflex latency was abnormal in both exposed workers and in local unexposed compared to other reference groups. Focus of the inquiry on lung disease helped ensure that for neurobehavioral tests confounding factors were minimal and known biases were small. We tentatively attribute the neurobehavioral impairments and increased symptom frequencies to chlorinated solvent exposure. Excessive respiratory symptoms are attributed to welding stainless steel combined with cigarette smoking. Specifically, manganese exposure may have affected the respiratory and the central nervous systems.  1999 Academic Press Key Words: balance; blink; reaction time; trichloroethylene; 1,1,1-trichloroethane-chlorinated solvents.

INTRODUCTION

Respiratory symptoms and impaired function have been associated with welding of stainless steel 244 0013-9351/99 $30.00 Copyright  1999 by Academic Press All rights of reproduction in any form reserved.

ADVERSE EFFECTS OF CHLORINATED SOLVENTS

TABLE 1 Metal Concentrations in Air (mg/m3) 8-H Time Weighted Average by Personal and Area Samplers for Six Workers in the Welding and Grinding Shop, 18 April 1993 Metal

Personal

Aluminum Beryllium Chromium Cobalt Copper Iron Manganese Molybdemum Nickel Titanium Zinc Strontium chromate

Area

0.09—0.10 (0.012a 0.008—4.16 (0.012—19 0.06 0.035—4.43 0.002—0.16 (0.013 0.02—12.25 0.48—1.2 0.02—0.05 0.006—0.29

1,1,1-Tricholorethane Tricholoroethylene Methanol Trichlorofluoroethane FC-113

(0.08 (0.005a 0.009 (0.005 (0.005 0.012 (0.002 (0.05 0.018 0.48 0.018 0.06

'1200 4800 13000 70,000

8-h PEL 10 0.002 0.5 0.05 1.0 5 5 10 1 10 10 0.0005 45 450 260 7600

Note. Underline indicates exceeded PEL. a Detection limit above PEL.

the 8-h PELs. No measurements of volatile organic chemicals were known to the AFGE union or discovered during subsequent investigation, which implied that they had not been measured. However, the union officers knew that large amounts of TCE and other chlorinated solvents had been consigned to Tinker AFB. Furthermore, Tinker AFB had been designated as a Superfund site because of solvents in groundwater of the Gerber-Wellington aquifer beneath the AFB (National Priorities List, 1990) This made it possible that TCE, PERC, and related chlorinated solvents contributed to workers respiratory complaints, rheumatic symptoms, depression, and possible neurobehavioral impairment. METHODS

One hundred fifty-four active workers from among nearly 500 who worked in the jet engine section were studied in July 1993. Most of those tested came from areas where metals were cleaned with the solvents listed. They included 51 women, 103 men, 27 African—Americans, and 127 Caucasians. We asked workers to contact an unexposed subject of their age and sex. These subjects were interviewed and accepted if they met the matched workers group and had not worked at Tinker AFB nor had been exposed to chemicals. For analysis, the groups were matched, not paired as individuals. The 112 local unexposed subjects included 46 women and 66 men. Twenty-

245

one were African—American and 91 were Caucasian. The distant unexposed comparison for blink latencies was needed because the Oklahoma ‘‘unexposed’’ subjects blink latencies were found to be prolonged (Kilburn et al., 1998). These 68 women and 49 men had been recruited from election rolls of an Arizona community and were tested in July 1993 for blink and other functions using identical methods by the same medical team and in the same season of the same year as Tinker subjects. Testers at Tinker were blinded to the exposure status of subjects. These neurophysiological and neuropsychological function tests had differentiated subjects exposed to chemicals from unexposed ones in previous studies (Kilburn and Warshaw 1993a,b, 1995; Kilburn et al. 1998b). The blink methods and technicians were identical to those used to study national unexposed groups. The profile of mood states (POMS) (McNair and Lorr 1967; McNair et al., 1971/1981) and questionnaires, including a respiratory one (DLD-78, Ferris, 1978), were completed by all subjects. Alcohol was measured in expired air with a breathalyzer. Spirometry measurements met ATS criteria (ATS Statement, 1987; Miller et al., 1986). All values were adjusted for height, sex, and years of cigarette smoking and were expressed as percentage predicted. Standard posteroanterior and lateral chest X rays were obtained and interpreted using the International Labour Office (1980) criteria for grading opacities and measured for lung volume by standard methods (Kilburn et al., 1992a; Harris et al., 1971). Neurophysiological tests. Simple reaction time and visual two-choice reaction (CRT) were timed using a computer clock from the appearance of a 6cm-tall letter A for simple and A or S for two choice on a video screen to its cancellation by striking the A or S on a key pad (Miller et al., 1989). Order and interval between stimuli were directed by a pseudorandom computer program for runs of 20 stimuli. One run was recorded for simple and three for choice reaction time. The score was the median of the last seven response times in each run, CRT ,  CRT2, CRT3, etc. Body balance while standing erect with feet together was tracked from a soundgenerating stylus on a headband by two microphones mounted 30 cm lateral to the emitter and 38 cm apart, processed in a computer, and expressed as mean speed of sway in centimeters per second (Kilburn et al., 1992b). The blink reflex latency was measured with surface electromyographic electrodes (EMG) from the lateral orbicularis oculi muscles

246

KAYE H. KILBURN

bilaterally (Feldman et al., 1988) after tapping of each supraorbital notch with a light hammer, which triggered a recording computer. Ten firings of R-1 and the volley of second waves, R-2, were averaged to find the mean response in milliseconds and failures were recorded (Warshaw et al., 1993; Kilburn et al., 1998). Unexposed subjects were tested for blink by the same examiners using identical methods and the same apparatus in the same year. Color discrimination was measured with the desaturated 15 hue test (Lanthony, 1978) and scored by the method of Bowman (1982). Neuropsychological tests. Culture Fair battery 2A tested nonverbal, nonarithmetic intelligence by the selection of designs for similarity, difference, completion, and pattern recognition and transfer (Cattell et al., 1941; Cattell, 1951). A vocabulary test of 46 words was administered. Culture Fair and vocabulary tests were given to groups of subjects (Jackson, 1985). Two tests, slotted pegboard and trail making A and B, which measure dexterity, coordination, decision making, and discrimination from the Halstead—Reitan battery (Reitan 1966; Matthews et al., 1970), were administered individually. Subjects self-assessed their moods during the preceding week using the profile of mood states (POMS) (McNair and Lorr 1967; McNair et al., 1971/1981). The POMS score was the sum of scores for anxiety, depression, anger, extreme fatigue, and confusion minus the vigor score. Questionnaires (Kilburn and Warshaw, 1993a,b) inquired about demographic data, occupational exposures, and respiratory, rheumatic (Levin et al., 1984), neuropsychiatric, and other disorders and the frequency of 35 key symptoms for irritative, respiratory, neurologic, and vegetative status. Each subject marked the frequency of each symptom on a scale from 1 as never to 11, daily. All questionnaire responses were checked for completeness using a computer, and any missing data were obtained. Chronic bronchitis was diagnosed when cough and phlegm were present for most days of the week for 3 months in 2 succeeding years. All scores and computed data for sway, blink, and reaction time were entered into a IBM-compatible microcomputer and descriptive and analytical computations including analysis of variance (ANOVA) and linear regression modeling with Stata Statistical Software (STATA Corp., College Station, TX). To examine influences on the differences between the groups, the test scores were appropriately transformed to the natural logarithm or reciprocal when this improved the symmetry of the distribution.

When coefficients obtained by stepwise linear regression for age, educational level, gender, ethnicity, height, and weight were significant, test scores were adjusted for them. Minimal statistical significance was P(0.05. RESULTS

All results are means for groups unless otherwise identified. The 154 workers’ mean age of 42.7 years was not significantly different from the mean of 41.9 years of the 112 unexposed subjects (Table 2). Their educational attainments were 12.4 and 12.9 years, respectively, statistically different but of small magnitude. School grade point averages were not different (3.4 and 3.6, C#). Family incomes were not different for the groups. One exposed man was intoxicated with alcohol (90.08 ll/dl) and was dropped from analysis. Otherwise, alcohol and drug use profiles were not different in exposed and unexposed subjects. Respiratory symptoms and pulmonary function. There were large differences in all respiratory symptoms between the groups (Table 2). Fifty-eight percent of the exposed subjects produced phlegm compared to 11.6% of unexposed and two-thirds had

TABLE 2 Age, Educational Level, Smoking (Ever) and Respiratory Symptoms for 154 Exposed Workers and 112 Unexposed Subjects Oklahoma

Age/year Educational Attainment Smoking Ever (%) Duration (years) Phlegm Chronic bronchitis by Criteria Shortness of breath At rest Walking Climbing stairs Wheezing Shortness of breath with wheezing Normal breathing between attacks Abnormal radiographs (%)

Exposed (154)

Unexposed (112)

Mean SD

Mean SD

P value

42.7 12.4

41.9 12.9

0.453 0.042

8.2 1.4

8.8 2.1

72% 61% 14.4 12.9 12.3 14.4 58.4 11.6 64.9 26.8

0.091 0.374 0.0005 0.0005

39.0 53.6 72.2 50.5 38.8 77.1 12.3

0.0005 0.0005 0.0005 0.0005 0.0005 0.001 0.01

4.5 6.2 29.4 8.9 7.1 41.7 5.1

Note: Irregular opacities and/or pleural changes consistent with asbestosis percentage of group.

247

ADVERSE EFFECTS OF CHLORINATED SOLVENTS

TABLE 3 Comparison of Pulmonary Function Tests in 154 Exposed Workers and 112 Unexposed Subjects Exposed (n"154)

Unexposed (n"112)

Status and number

Mean SD

Mean SD

P value

FVC (% predicted) FEV1 (% predicted) FEF25—75 (% predicted) FEF75—85 (% predicted) RV/ TLC TLC (% predicted)

99.5 97.8 97.2 87.3 43.7 115.8

103.0 101.6 100.9 93.4 43.5 117.4

0.034 0.066 0.399 0.712 0.831 0.319

13.7 16.4 37.6 48.9 9.3 12.9

12.9 16.0 32.5 43.0 8.1 12.5

chronic bronchitis by epidemiological criteria (Ferris, 1978). Wheezing and shortness of breath at rest, during walking, and climbing stairs were all significantly more frequent in the exposed subjects. The 266 chest radiographs showed small, irregular opacities characteristic of asbestosis by the criteria of the International Labour Office (1980) in 12.3% of exposed and 5.1% of unexposed groups, a significant

difference but no evidence of other fibrotic or alveolar filling diseases was found. Forced vital capacity (FVC) was significantly lower in exposed subjects and their difference for FEV1 bordered on significant when compared as percentage of predicted (Table 3). However, despite previous adjustments for the effects of smoking, differences disappeared when only subjects who had never smoked were compared. Flows FEV1, FEF25—75, and FEF75 were normal in both groups after adjusting for height, age, and duration of cigarette smoking. Total lung capacity was elevated in both groups, as were the RV/TLC ratios, but there were no differences between the exposed and unexposed. Neurophysiological measurements. Simple and choice reaction times were significantly slower in the exposed subjects by 60 ms (P(0.0001) and 96 ms (P(0.0005) (Table 4). Sway speed with eyes open was significantly different (0.92 vs 0.76 cm/s (P( 0.0005)) as was sway speed with eyes closed (1.60 vs 1.30 cm/s (P(0.0005)). Color discrimination was poorer in the exposed group (P(0.022). Blink reflex R-1 latencies after

TABLE 4 Age, Educational Level, Alveolar Air Alcohol Concentration, and Neurophysiological Function in Exposed Workers and Unexposed Subjects Local Exposed Mean Age/year Educational attainment Alv. air alcohol (ppm) Neurophysiological Simple reaction time (ms) Choice reaction time 2 (ms) Balance, sway speed (cm/s) Eyes open 2 Eyes closed 2 Color vision score Blink Reflex Supra orbital Right (ms) Left (ms) Neuropsychological Cognitive function Culture Fair A Vocabulary Perceptual Motor Speed Pegboard Dominant (s) Trail making A Trail making B POMS score

42.7 12.4 0.0008 349 628

Unexposed

SD 8.2 1.4 0.0002 145 165

Mean 41.9 12.9 0.0024 289 532

SD 8.8 2.1 0.0006 87 95

P value 0.453 0.042 0.163 0.0005 0.0005

0.92 1.60 12.2

0.36 0.76 1.5

0.76 1.30 11.8

0.16 0.38 1.1

0.0005 0.0005 0.022

13.8 13.8

1.8 1.8

13.5 13.8

1.8 1.8

0.353 0.974

27.8 20.3

6.6 8.3

27.0 19.0

6.8 8.4

0.381 0.214

75.0 37.5 79.9 83.0

18.4 14.0 28.4 42.4

72.0 33.2 79.7 14.5

18.2 11.3 35.8 28.2

0.193 0.008 0.951 0.0005

248

KAYE H. KILBURN

TABLE 5 Comparison of Blink Reflex Latency after Supraorbital Tap of Workers, Oklahoma City Unexposed and Distant Arizona Unexposed Subjects Unexposed

Race % Afro/Cauc Sex (%) Women/men

Tinker (154)

Oklahoma (112)

Arizona (117)

17.5

18.8

0

33

35.7

59

Age years

Mean sd 42.7$8.2

Mean sd 41.9$8.8

Mean sd 42.4$15.4

Blink reflex latency (ms) Right Left

13.8$1.8 13.8$1.8

13.5$1.8 13.8$1.8

12.8$2.1 12.9$2.1

P1a

P2b

P3c

0.35 0.97

0.0003 0.00005

0.0141 0.001

a

P1 values for comparison of exposed to OK unexposed. P2 values for comparison of exposed to AZ unexposed. c P3 values for comparison of OK unexposed to AZ unexposed. b

supraorbital stimulation were 13.8 ms on right and left and not different from those of unexposed Oklahoma City subjects. However, these latencies were statistically significantly abnormal compared to a unexposed group outside this region (Table 5). When the mean differences between groups were adjusted for age and education, they remained statistically significant for balance, reaction time, trail making A, and POMS score. The difference in vocabulary became significant and the difference in color score became insignificant in this analysis. The lack of difference between Oklahoma exposed and unexposed groups for blink was also verified. Ethnicity and gender were balanced between local groups so that these factors could be ignored for comparison of exposed workers and unexposed subjects. Cognitive function tested by Culture Fair was not different between the groups but vocabulary was lower (adjusted difference was 2.50 standard deviation 0.95 P(0.007) in exposed compared to unexposed subjects. Perceptual motor function was also not different for slotted pegboard and trail making B. Trail making A scores were different, but were considered less important, as this test is a training exercise for trail making B. Subjective findings. Profile of mood states scores for exposed subjects was 83 (mean) compared to a 14.5 mean for unexposed subjects (Table 4). Anger, tension, confusion, depression, and fatigue were all significantly elevated and vigor was only half as high in exposed subjects as in the unexposed ones.

The interaction of POMS scores with neurobehavioral test scores in separate correlation matrices for exposed subjects and for unexposed ones showed small coefficients that were insignificant ((0.16) in the unexposed group. Exposed subject coefficients were 0.19—0.26, whereas 0.70 was the threshold for statistical significance. Thus, there were no important interactions of POMS score with neurobehavioral functions. Symptom frequencies were higher for exposed subjects than for unexposed for all 35 complaints (Table 6). The magnitudes of these differences varied but they were statistically significant for all areas: neurobehavioral, irritative-respiratory, and general-vegetative symptoms. The 10 rheumatic complaints which constitute the American Rheumatism Association’s (ARA) criteria for lupus erythematosus were significantly more frequent in the exposed than in the unexposed except for seizures. Nearly 72% of the exposed groups had numb fingers, 44% had had pleurisy (pain on breathing), 38% had been told they had anemia, and 32% had had rheumatic pain. Eighteen percent of exposed subjects and 0% of unexposed subjects had five or more symptoms and those with four symptoms were 34% vs 0.9% (P(0.0000). As four or more ARA symptoms are considered a presumptive diagnosis of lupus erythematosus, this observation should be pursued (Levin et al., 1984). Sources of confounding. The prevalence of prior respiratory disorders and medical diseases and the

ADVERSE EFFECTS OF CHLORINATED SOLVENTS

TABLE 6 Symptom Frequencies as Means for Exposed and Unexposed Subjectsa 154 exposed 112 unexposed Skin itching Finger nail changes Chest tightness Palpitations Burning in chest Shortness of breath Dry cough Cough with mucus Cough with blood Dry mouth Throat irritation Eye irritation Reduced sense of smell Headache Nausea Dizziness Lightheadedly Unusual exhilaration Balance disturbance Loss of consciousness Extreme fatigue Somnolence Cannot fall asleep Wakes frequently Sleeps only a few hours Irritability Lack of concentration Recent memory loss Long-term memory loss Mood instability Decreased libido Decreased alcohol tolerance Indigestion Loss of appetite Stomach bloats/swells a

6.1 4.1 4.9 4.1 4.4 5.8 4.3 5.8 2.2 5.9 5.5 5.5 6.0 6.5 4.6 4.8 5.0 2.3 5.0 1.8 7.8 6.4 4.8 6.1 5.6 6.7 7.0 7.5 6.0 5.7 5.9 3.4 6.2 3.6 5.4

2.8 1.5 2.1 2.1 2.1 2.4 2.3 2.8 1.2 2.8 2.6 2.4 1.9 4.0 2.3 2.2 2.5 1.5 2.1 1.3 2.9 2.3 2.6 2.8 2.9 3.2 2.8 2.9 2.1 2.0 2.5 2.3 3.4 2.4 2.6

P 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.034 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.008 0.000 0.000 0.000

Scale: 1, never; 11, always.

use of medication showed only small differences between the exposed and unexposed subjects, but angina pectoris was more common in the exposed. In contrast, use of alcohol, overdose with alcohol or with drugs, and use of marijuana and lysergic acid diethylamide (LSD) were low and equal in the groups. Comparison of the frequencies of other factors which might influence neurobehavioral performance such as head injury, unconsciousness, hours of anesthesia, heart bypass surgery, neurological illnesses, drug overdose, and use of marijuana, and LSD revealed that they were infrequent and not different between the groups. Past psychiatric illness in 13.6% of exposed vs 4.5% (P(0.003) of unexposed and the use of tranquilizers in 30.5% of exposed vs 15.1% of unexposed were different

249

(P(0.004), as was the diagnosis of depression, which was 12.3% of exposed and 3.6% of unexposed (P(0.0001). Use of tranquilizers as a possible confounder of neurobehavioral tests showed by regression analysis that there were no significant differences between mean test values of the current tranquilizer users or ever users versus nonusers in either group of subjects. Similarly, a history of depression had no significant effect on neurobehavioral test scores and symptoms. Exposure resulting from 15 occupations, including metal refining and aerospace, and working with solvents, adhesives, asbestos, and vibrating tools, were significantly more frequent in the exposed group. Because they reflected regular work assignments at Tinker, they were not confounders. Of other exposures, only one, to herbicides, was more prevalent— at 4.5% in exposed than in unexposed 0.6%—but excluding the herbicide and the vibrating tool subgroups did not affect the results. Numb fingers and rheumatic complaints were attributed to the use of vibrating tools. The blink reflex latency abnormality in the workers was consistent with their occupational exposure to chlorinated solvents, especially to TCE (Feldman et al., 1985, 1988; Kilburn and Warshaw, 1993a). However, the same abnormality in the unexposed group was not expected and was not related to occupational exposure. As a preliminary step to understand this finding, we divided all Oklahoma subjects by their city of residence (Fig. 1). Thus, the mean R-1 latencies after right supraorbital tap in exposed and unexposed subjects’ in 10 communities surrounding Tinker AFB were then compared by analysis of variance (ANOVA). In Oklahoma City and Shawnee exposed subjects’ R-1’s were slower than unexposed while in Del City immediately west of Tinker the unexposed, subjects were significantly slower than exposed subjects. Only in Moore did unexposed and exposed subjects match with entirely normal latencies of 12.5 ms. These were statistically significantly lower values than those found in Oklahoma City, Del City, and Midwest City. DISCUSSION

This investigation focused on the persistent respiratory complaints of welders and helpers. Because the exposure to metals measured by area samples was below the PELs, and many welders complained of impaired memory and loss of concentration, it was decided to test the performance of the central nervous system as well as lung functions. The initial

250

FIG. 1.

KAYE H. KILBURN

The location of Tinker Air Force Base ( jet plane silhouette) is related to the residence cities of exposed and unexposed groups.

focus of the study on the lung problems meant that any self-selection bias by workers should be directed toward respiratory symptoms and functions, not those of the central nervous system. Thus, any influences on worker’s participation, performance, or the picking of suprahealthy unexposed to compare to themselves would be misdirected. Opportunities for bias toward central nervous system functions would be minimized by the study’s focus on the pulmonary system. Impairment of both lung and brain occurs from exposure to manganese (Cotzias et al., 1971), from cadmium and copper, and from solvents used as degreasers (Reid et al., 1973; Hanninen et al., 1976; Kilburn and Warshaw 1993a). The excesses of chronic bronchitis, shortness of breath, and other respiratory symptoms, together with the reduced vital capacities, are consistent with industrial bronchitis due to inhalation of welding fumes and of particles from grinding stainless steel. There was more radiographic evidence of disease due to exposure to asbestos (Kilburn and Warshaw, 1990), to hard-metal grinding tools (cobalt) (Sprince et al., 1988), or to manganese (steel welding, Davies, 1946) in workers at Tinker AFB than in the unexposed group, in keeping with their workplace exposures.

Jet engine repair workers had impaired balance, simple and choice reaction time, blink reflex latency, and color discrimination and excess frequencies of neurobehavioral symptoms. Excessive blink reflex latencies have been associated with TCE toxicity (Kilburn and Warshaw 1993a,b; Feldman et al., 1985, 1988). Also, disorders of mood shown by elevated POMS scores, psychiatric disorders, and principally depression (accompanied by the use of tranquilizer drugs) were significantly more common in the Tinker workers. Contamination of soil and water with solvents, particularly trichloroethylene, had led to Tinker AFB being designated as a group 9 Superfund site in 1987 (National Priorities List, 1990). The lung quickly absorbs TCE as it does other anesthetic gases because of a blood—gas partition coefficient of 9.92 at 37°C and a high lipid solubility (olive oil to gas) coefficient of 960 at 37°C (Davidson and Beliles, 1991). There is further rapid partition to the brain, consistent with its use as an anesthetic before 1965. Cranial neuropathies have followed inhaled anesthesia, and surface anesthesia has occurred after industrial exposure (Feldman et al., 1985). Chronic TCE exposure of 31 printers reduced transmission in sural and trigeminal nerves (Ruijten

ADVERSE EFFECTS OF CHLORINATED SOLVENTS

et al., 1995). Delayed blink reflex latency R-1 was first shown from neighborhood TCE exposure in Woburn, Massachussetts (Feldman et al., 1988), and Tucson, Arizona. In Tucson, there was also slowing of reaction time and increased speed of sway (balance), coupled with cognitive, recall, and perceptual motor impairment, especially of trail making A and B (Kilburn and Warshaw, 1993a). Excessive cardiac birth defects were found in the TCE exposed group at Tucson (Goldberg et al., 1990), reflecting another toxic effect. Chemical exposure. Measurements of concentrations of the suspected solvents, especially TCE in Tinker engine shops, were rare (Table 1). Such ‘‘one point in time’’ measurements have not correlated with neurobehavioral impairment in other subjects. Thus, exposures to TCE above 5 ppb in water for many years were associated with neurobehavioral impairment but were not correlated with well water measurements of TCE as peak levels, high years, average levels, cumulative doses nor to air estimates in Tucson (Kilburn and Warshaw, 1993a,b). Therefore, adverse effects that are most plausibly related to proximity to chemical releases frequently do not show dose relationships in environmental epidemiology. The plausibility is strengthened because neurobehavioral effects were connected to workplace chlorinated solvent exposures two decades ago and solvent mixtures had greater effects than did one agent (Hanninen et al., 1976; Maizlish et al., 1985). Bias. Tester bias was minimized because the exposure status of subjects at Tinker was withheld from the staff doing the examinations. That all the tested subjects were volunteers and some were clients in a lawsuit would not alter their performance of balance and blink, which are subconscious acts and unlikely to slow reaction time without causing excessive variability, which was not found. Also, because the study plan focused on the respiratory tract rather than the central nervous system, significant test bias affecting brain function is unlikely. Confounding factors. Potential confounding factors were searched for systematically with the questionnaire. The distributions of age and educational levels were nearly identical, and so had no effect. Only the expected occupational exposures from jet engine repair separated the workers from controls. Other factors, including alcohol and drug use, unconsciousness, and prior neurological illness, were unusual in both groups. Regional effects. Delayed blink reflex latency in local unexposed might argue against TCE exposure

251

of workers, but chemical exposure of the unexposed is more likely because their blink reflex latencies exceeded those of other unexposed groups (Kilburn and Warshaw, 1993a,b; Kilburn et al., 1998) and matched those exposed groups (Feldman et al., 1985, 1988). As a working hypothesis, TCE and similar solvents that migrated from the Tinker Superfund site for over 40 years and from other industrial sources are thought to be the major contributors to TCE contamination of the Gerber-Wellington aquifer and thus groundwater exposed the communities where ‘‘unexposed’’ controls lived. The only other speculation is that delayed blink R-1 might be due to chronic exposure to low levels of hydrogen sulfide escaping from bore holes, oil wells, and refineries which have been present in Oklahoma City oil fields for 50 years. Neurobehavioral effects of exposure to low levels of H S have been found  (Kilburn and Warshaw, 1995). To help choose between the alternative explanations for delay of blink latency R-1 in unexposed subjects in the Oklahoma City area, TCE concentrations in water from the aquifer, in other groundwater, in the air levels of hydrogen sulfide in this zone, and in levels of other neurotoxic chemicals should be measured. Addition of these data to the locations of the residences of subjects with abnormalities of blink, and those of balance and of reaction time, in short overlaying the map with these concentration zones of TCE, of H2S, and any other neurotoxic chemicals, should help us to choose between the alternatives. ACKNOWLEDGMENTS Robert G. Fraser, M.D. FRCP(C), and Joseph Calhoun, M. D., provided chest radiographic interpretation using ILO criteria for pneumoconiosis. This study was supported by Plaintiff Law Firms and most of the exposed subjects were participants in a legal action.

REFERENCES Arlien-Soborg, P. (1992). ‘‘Solvent Neurotoxicity.’’ CRC Press, Boca Raton, FL. ATS Statement (1987). Standardization of spirometry—1987 update. Am. Rev. Respir. Dis. 136, 1285—1298. Bowman, K. J. (1982). A method for quantitative scoring of the Farnsworth panel D-15. Acta Ophth. 60, 907—916. Cattell, R. B., Feingold, S. N., and Sarason, S. B. (1941). A culture free intelligence test II evaluation of cultural influences on performance. J. Educ. Psychol. 32, 81—100. Cattell, R. B. (1951). Classical and standard score IQ standardization of the IPAT: Culture free intelligence scale 2. J. Consulting Psychol. 15, 154—159.

252

KAYE H. KILBURN

Cotzias, G. C., Papavasiliou, P. S., Ginos, J., Steck, A., and Duby, S. (1971). Metabolic modification of Parkinson’s disease and of chronic manganese poisoning. Annu. Rev. Med. 22, 305—326. Davidson, W. F., and Beliles, R. P. (1991). Considering the target organ toxicity of trichloroethylene in terms of metabolite toxicity and pharmacokinetics. Drug Metab. Rev. 23, 49, 3—5. Davies, T. A. L. (1946). Manganese pneumonitis. Brit. J. Ind. Med. 3, 111—135. Feldman, R. G., White, R. F., Currie J. N., Travers, P. H., and Lessell, S. (1985). Long term follow-up after single toxic exposure to trichloroethylene. Am. J. Ind. Med. 8, 119—126. Feldman, R. G., Chirico-Post, J., and Proctor, S. P. (1988). Blink reflex latency after exposure to trichloroethylene in well water. Arch. Environ. Health 43, 143—148. Ferris, B. G., Jr. (1978). Epidemiology standardization project. Am. Rev. Resp. Dis. 118, 7—54. Goldberg, S. J., Lebowity, M. D., Graves, E. J., and Hicks, S. (1990). An association of human congenital cardiac malformations and drinking water contaminants. J. Am. Coll. Cardiol. 16, 155—164. Hanninen, H., Eskelinen, L., Husman, K., and Nurminen, M. (1976). Behavioral effects of long-term exposure to a mixture of organic solvents. Scand. J. Work Environ. Health 4, 240—255.

Kilburn, K. H., Thornton, J. C., and Hanscom, B. (1998a). A field method for blink reflex latency (BRL R-1) and prediction equations for adults and children. Electromyography Clin. Neurophysiol. 38, 25—31. Kilburn, K. H., Thornton, J. C., and Hanscom, B. (1998b). Population-based prediction equations for neurobehavioral tests. Arch. Environ. Health 53, 257—263. Lanthony, P. (1978). The desaturated panel D-15. Doc. Ophthalmol. 46, 185—189. Levin, R. E., Weinstein, A., Peterson, M., Testa, M. A., and Rothfield, N. F. (1984). A comparison of the sensitivity of the 1971 and 1982 American Rheumatism Association criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 27, 530—538. Mackay, C. J., Campbell, L., Samuel, A. M., Alderman, K. J., Idzikowski, C., Wilson, H. K., and Gompertz. D. (1987). Behavioral changes during exposure to 1,1,1-trichloroethane: Timecourse and relationship to blood solvent levels. Am. J. Ind. Med. 11, 223—239. Maizlish, N. A., Langolf, G. D., Whitehead, L. W., Fine, L. J., Albers, J. W., Goldberg, J., and Smith, P. (1985). Behavioral evaluation of workers exposed to mixtures of organic solvents. Brit. J. Ind. Med. 42, 579—590.

Harris, T. R., Pratt, P. C., and Kilburn, K. H. (1971). Total lung capacity measured by roentgenograms. Am. J. Med. 50, 756—763.

Matthews, C., Cleeland, C. S., and Hopper, C. L. (1970). Neurological patterns in multiple sclerosis. Dis. Nervous Syst. 31, 161—170.

International Labour Office (1980). ‘‘U/C International Classification of Radiographs of Pneumoconiosis.’’ In Occupational Safety and Health series. Geneva International Labour Office.

McNair, D. M., and Lorr, M. (1967). An analysis of mood in neurotics. J. Abnorm. Soc. Psychol. 69, 620—624.

Iregren, A. (1990). Psychological test performance in foundry workers exposed to low levels of manganese. Toxicol. Teratol. 12, 673—675. Jackson, D. N. (1985). ‘‘Multidimensional Aptitude Battery.’’ Sigma Assessment Systems Inc., Port Huron, MI. Kilburn, K. H., and Warshaw, R. H. (1989). Pulmonary functional impairment from years of arc welding. Am. J. Med. 87, 62—69. Kilburn, K. H., and Warshaw, R. H. (1990). Pulmonary functional impairment associated with pleural asbestos disease: circumscribed and diffuse thickening. Chest 98, 965—972. Kilburn, K. H., and Warshaw, R. H. (1992). Neurobehavioral effects of formaldehyde and solvents on histology technicians: repeated testing across time. Environ. Res. 57, 221—233. Kilburn, K. H., and Warshaw, R. H. (1993a). Effects on neurobehavioral performance of chronic exposure to chemically contaminated well water. Int. Toxicol. Ind. Health 9, 391—404. Kilburn, K. H., and Warshaw, R. H. (1993b). Neurobehavioral testing of subjects exposed residentially to groundwater contaminated from an aluminum die-casting plant and local unexposed. J. Toxicol. Environ. Health 39, 483—496. Kilburn, K. H., and Warshaw, R. H. (1995). Hydrogen sulfide and reduced sulfur gases adversely affect neurophysiological functions. Int. Toxicol. Ind. Occup. Health 11, 185—197. Kilburn, K. H., Warshaw, R. H., Thornton, J. C., Thornton, K., and Miller, A. (1992a). Predictive equations for total lung capacity and residual volume from radiographs: based on a probability sample of the Michigan population. Thorax 47, 519—523. Kilburn, K. H., Warshaw, R. H., and Hanscom, B. (1992b). Are hearing loss and balance dysfunction linked in construction iron workers? Brit. J. Ind. Med. 49, 138—141.

McNair, D. M., Lorr, M., and Droppleman, L. F. (1971/1981). ‘‘Profile of Mood States.’’ Educational and Industrial Testing Service, San Diego, CA. Miller, A., Thornton, J. C., Warshaw, R., Bernstein, J., Selikoff, J., and Teirstein, A. S. (1986). Mean and instantaneous expiratory flows, FVC, and FEV1: Prediction equations from a probability sample of Michigan, a large industrial state. Bull. Eur. Physiopathol. Resp. 22, 589—597. Miller, J. A., Cohen, G. S., Warshaw, R., Thornton, J. C., and Kilburn, K. H. (1989). Choice (CRT) and simple reaction times (SRT) in laboratory technicians: factors influencing reaction times and a predictive model. Am. J. Ind. Med. 15, 687—697. National Priorities List Fact Book (June 1986). Supplementary lists and supporting materials 1990. U.S. Environmental Protection Agency HW 7.3, Washington, DC. Reid, W. D., Ilett, K. F., Glick, J. M., and Krishna G. (1973). Metabolism and binding of aromatic hydrocarbons in the lung. Am. Rev. Resp. Dis. 107, 539—551. Reitan, R. M. (1966). A research program on the psychological effects of brain lesions in human beings. In ‘‘International Review of Research in Mental Retardation’’ (N. R. Ellis, Ed.) Academic Press, New York. Ruijten, M. W. M. M., Verberk, M. M., and Salle, S. J. A. (1995). Nerve function in workers with long term exposure to trichlorethyene. Brit. J. Ind. Med. 48, 87—92. Sprince, N. L., Oliver, L. C., Eisen, E. A., Greene, R. E., and Chamberlin, R. I. (1988). Cobalt exposure and lung disease in tungsten carbide production. Am. Rev. Respir. Dis. 138, 1220—1226. Warshaw, R. H., Hanscom, B., and Kilburn, K. H. (1993). A reproducible trigger for the measurement of trigeminal latencies elicited by a glabellar tap. Environ. Res. 60, 328—331.