Neurobehavioral effects of formaldehyde and solvents on histology technicians: Repeated testing...

ENVIRONMENTAL RESEARCH 58, 134--146 (1992)

Neurobehavioral Effects of Formaldehyde and Solvents on Histology Technicians: Repeated Testing across Time

KAYE H . KILBURN AND RAPHAEL H. WARSHAW

Environmental Sciences Laboratory, School of Medicine, University of Southern California, 2025 Zonal Avenue, RB-1 Room 226, Los Angeles, California 90033; and Workers Disease Detection

Services, Inc., 441 West Allen Avenue, Suite 114, San Dimas, California 91773

Received June 19, 1991

Neurobehavioral functions were studied by periodic testing of 318 histology technicians and by a single session testing 494 of such technicians from 1982 through 1986. Tests included immediate recall of stories, of drawings, and of number series from the Wechsler Memory Scale, block designs from the Wechsler Adult Intelligence Scale (WAIS), slotted pegboard, trail making A and B, embedded figures, number writing on the fingers, visual simple and two-choice reaction times, balance (speed of body sway), and the profile of mood state (POMS) score. Mean test scores of four initial "first test" groups were not different from 1982 to 1986. Immediate recall of stories and drawings improved on the first retest as did the block design score. Other test scores showed insignificant variation over the 4 years, including memory for numbers forward and backward, trails A and B, number writing on fingers, radius of body sway, and choice reaction time. Only placing of pegs in the pegboard took longer. POMS score was lower in the second year in naive subjects and decreased from 39 to 13 in those retested. Variations in results of neurobehavioral tests given at intervals across 4 years were small. No cumulative effects of occupational exposures or of aging were found. However, small decreases may have been offset by increases due to test familiarity or learning. © 1992 Academic Press, Inc.

INTRODUCTION

Behavioral toxicology had emerged as a subspecialty by 1974 (Mello, 1975) and experimental animal studies were thought to hold promise for prediction while clinical studies were limited, of necessity, to descriptions, after the fact. Since that time an urgent need to detect and describe early consequences of toxic chemical exposure, particularly on the central nervous system (CNS) (Kilburn, 1989), has emerged as exposures to industrial chemicals which were generally thought to be occupational have occurred in communities, especially around Su- perfund sites. This was underscored in 1990 when 1207 Superfund sites were identified in the United States (USEPA, 1990). The effects of long-term residential exposures to chemicals might be predicted from effects of chronic occupational exposure. Thus formaldehyde, a well-recognized irritant and a possible human carcinogen (Federal Panel on Formaldehyde, 1982), deserves study. Animal studies in the 1980s showed that its inhalation caused nasal and pulmonary cancers (Swenberg et al., 1980). Chronic low dose occupational exposure produced symptoms of airway irritation. Broader inquiry during studies of respiratory complaints found neurological symptoms (Kilburn et al., 1985a, 1985b) and subsequently evidence of impaired memory, equilibrium, and dexterity (Kilburn et al., 1987). Also there was evidence of impaired pulmonary function (Kilburn et al., 1989).

134

N E U R O B E H A V I O R A L TESTS ON HISTOLOGY T E C H N I C I A N S 135

Histology technicians are exposed to formaldehyde and to combinations of solvents, including xylene, ethanol, and chloroform. Ethanol, xylene, paint solvents, and carbon disulfide have been identified as producing CNS impairment (Hanni- nen et al., 1976; Elofsson et al., 1980; Baker et al., 1985; Morrow et al., 1990). Solvent combinations often had greater effects than single agents (Hanninen et al., 1976; Baker et al., 1985). Thus histology technicians exposed to combinations of solvents and to formaldehyde and who showed evidence of increased symptoms and impairment on cross-sectional studies related to hours per day of exposure raised the issue of long-term effects (Kilburn et al., 1985a, 1985b, 1987). To respond to this question we tested groups of them every 1 or 2 years to determine whether neurobehavioral function deteriorated from year to year. To do this we had to examine the effects of retesting a year or more apart on test performance.

METHODS

Access to histology technicians to recruit for testing was provided by yearly symposia, workshops, and convention meetings of the National Society for His- totechnology. Newsletters before the meeting, personal contact with members of the health and safety committee, and presentation of the study's aims and pro- cedures orally and by bulletins during the meetings encouraged members to be examined. Those who were examined represented in age, seniority, or experience and sex the attendees at the annual meetings in 1982, 1983, 1985, 1986, and 1987. They were slightly older and more experienced than members of local chapters studied previously (Kilburn et al., 1985a, 1985b). It is reasonable to assume that they were more concerned about health and may have had more symptoms related to their work than their unexamined peers, Because the study was aimed at discovery of effects, in essence an analysis of "worst case," these possible positive biases were not considered adverse to the study plan.

Exposure to chemicals is frequently difficult to estimate for other than current dose levels for an industrial work force. Our study population was scattered across the United States and its exposure began up to 37 years before our study. To provide some ranges, contemporary exposures to formaldehyde and to xylene, toluene, and chloroform were measured by area sampling of workplace air in 10 representative histology laboratories in Los Angeles in 1983 by the regional NIOSH laboratory (Kilburn et al., 1985b). Formaldehyde levels in workplace air varied from 0.2 to 5 ppm. Xylene was from 8.9 to 12.6 ppm. Toluene was from 8.9 to 12.6 ppm and chloroform from 2 to 19.1 ppm. In staining and cover slipping operations maximal exposure solvent levels were measured during monitoring, cleaning, and refilling of automatic tissue processors: xylene was from 3.2 to 10.2 ppm and toluene was from 8.9 to 12.6 ppm. Formaldehyde levels were highest during selection of tissue samples for processing.

Self-administered questionnaires (Kilburn et al., 1985a, 1985b) for demographic data, occupational and residential exposures, and medical histories, including current health complaints, were checked for completion by trained interviewers after the testing. Complaint frequencies were rated from almost never to daily for 33 symptoms by each subject. Self-rating exposure scales were completed for

136 KILBURN AND W A R S H A W

hours of grossing (cutting fixed tissue), automatic processor monitoring, processor cleaning, cover slipping, and mixing of solvent solutions.

Neurophysiological and neuropsychological testing was done by appointment on all subjects whose exposure status was unknown to the testors. The neurobehavioral test battery was slightly modified from that used in earlier studies of histology technicians (Kilburn et al., 1987) and took 2 to 3 hr. Subjects were tested for memory, cognition, pattern recognition, dexterity, decision making, motor speed, and balance using established tests in October or November of 1982, 1983, 1985, and 1986.

In 1989 144 histology technicians were tested using neurophysiological assays for speed of sway, simple and choice reaction times, color vision, strength of grip, proportion of body weight as fat (Danniger, Body Composition Analyzer TV 1-10), blink reflex (R-l), Culture Fair A, and POMS. Although some of the subjects had been tested with the original battery, these were new tests for this group, except for reaction time and sway, and facilitated the search for relationships between various test results and age, educational level, and intelligence as measured by Culture Fair and affective status as assessed by POMS in a sample of this well- characterized group.

Simple reaction time (SRT) was measured by timing the interval from the appearance of an " A " on a video screen of a personal computer to its cancellation by the subject by pressing the corresponding key with the index finger of the dominant hand. The letters were generated by the computer using a pseudoran- dom numbers array and a letter appeared at between 2 and 6 sec. The median time in milliseconds for 7 responses following 13 sequences for learning was recorded as SRT (Miller et al., 1989). Choice reaction time (CRT) was measured as the interval from appearance to cancellation of an " A " or an " S " using the same apparatus. The median of the sequence of the final 7 of 20 responses from the third consecutive trial was scored as CRT. A change in computer keyboards between 1983 and 1985 increased mean CRT systematically.

Body balance was measured with the subject standing erect with feet together. It tests the vestibular apparatus, cerebellar function, proprioception, motor function, and visual effector links. Deviations from postural fixation while standing were mapped using a sound generating stylus mounted on the head and two microphones. This sonic ranging system recorded the sequential positions of the head and stored this path in a computer. The x and y axis moments of the path were plotted for the subject standing for 60 sec with eyes open and then for 60 sec with eyes closed. Results were expressed as mean speed of sway in cm/sec (Kilburn et al., 1991a).

The blink reflex was measured with surface electromyographic recording (EMG) from lateral orbicularis oculi muscles bilaterally (Kilburn et al., 1991b) after mechanical stimulation of the glabella (midline of lower front bone) by tap- ping with a lightweight hammer containing a phonographic cartridge which trig- gered a recording computer. Ten responses were recorded using special EMG amplifiers which operated with usable signal-to-noise ratios in ordinary rooms. Each firing of R-1 and the volley of second waves, R-2, were recorded. Ten R-l 's and R-2's were averaged to find the mean response. Failures on right and left were

NEUROBEHAVIORAL TESTS ON HISTOLOGY TECHNICIANS 137

also recorded. Speed and latency of eye closure were also recorded with an infrared sensor of our design and recorded by a computer.

For finger writing numbers 3 through 6 were written in random order four times on each finger pad of both hands while the subject's eyes were closed. Failures to recognize numbers were recorded, which reflected impaired peripheral cutaneous sensation and cortical interpretation, depending upon alertness and concentration (Reitan, 1966).

Immediate memory or recall was measured by three tests from Wechsler's Memory Scale (Wechsler, 1945). First was the immediate recall of words and phrases of two brief stories. Visual memory was tested by having the subject draw three diagrams immediately after they were presented for 10 sec. Digit span measured immediate recall after hearing sequences of four to nine numbers read forward and recalled forward and of three to eight numbers read forward but recalled backward.

Cognitive function. Culture Fair battery 2A, a nonverbal nonarithmetical intelligence scale based on selection of designs for series (similarity), classification (difference), matrices (completion), and topology (pattern recognition and trans- fer), estimates overall intelligence (Cattell et at., 1941; Cattell, 1951).

The profile of mood states, POMS (McNair and Lorr, 1967; McNair et al., 1971/1981), measured affective disorders by having each subject rate on a 0 to 4 scale of frequency 65 adjectives or phrases which characterize tension-anxiety, anger, depression, vigor, fatigue, and confusion, thus self-judging his feelings or emotional status during the preceding week.

Culture Fair and POMS were administered to groups of 12-24 subjects. Neu- rophysiological testing (balance, blink, and reaction time) was done individually by experienced investigators or technicians. Neuropsychological testing (memory scale, WAIS subtests, Halstead-Reitan tests, and others) was done individually and in sequence for each subject.

In construction of block designs from the Wechsler Adult Intelligence Scale (WAIS) (Wechsler, 1971) the subject assembled blocks to match diagrams which tests construction, interpretative, and integrative capacities. Time needed to com- plete each of nine designs was recorded and the composite score calculated. Digit symbol also from the WAIS consists of writing symbols beneath numbers and the number of correct responses in 90 sec is scored. Embedded figures tested dis- criminatory perception in recognizing 10 simple objects presented in each of four overprinted arrays. The number of objects recognized in 30 and 60 sec was scored (Valciukas and Singer, 1979).

Four tests which measure dexterity, coordination, decision making, and peripheral sensation from the Halsted-Reitan battery (Reitan, 1966) were administered in sequence. Eye-hand coordination or dexterity was measured as the time required to place 25 keyed pegs in a slotted pegboard, as fast as possible, first with the dominant and then the nondominant hand (Matthews et al., 1970). In trail making A the time needed to connect 25 numbered dots in consecutive order with a continuous pencil line measured perceptual motor speed. Normal subjects required 25 to 30 sec. Trail making B added interference and decision making as

138 KILBURN AND WARSHAW

numbers and letters are alternately connected in the order 1, A, 2, B, etc. Normal subjects required 60 to 65 sec, twice the time needed for trails A (Reitan, 1958).

All data, including those from the dedicated acquisition systems for sway, blink, and reaction time, were entered into a Compaq computer and descriptive and analytical computations made using Stata statistical software (Computing Resource Center, Los Angeles, CA). Significance of means of differences between groups were compared by Student's t test. Age regressions and other interactions were modeled by stepwise linear regression. There were 350 naive subjects, 19 subjects were studied at each of four sites, and 299 were retested once or twice. In 1989, 144 subjects were tested with an additional test battery for which they were all naive.

RESULTS

Testing 350 naive subjects provided an index of the reproducibility of selection and testing. Age decreased slightly but educational attainment remained above 14 years (Table 1) in the four groups of naive subjects. These four groups of subjects had remarkably similar performances on verbal recall, visual recall, recall of

TABLE 1 DESCRIPTION AND NEUROBEHAVIORAL TESTS (GRouP MEANS) AND STANDARD DEVIATIONS FOR

350 NAIVE SUBJECTS

Year

1982 1983 1985 1986

Mean SD Mean SD Mean SD Mean SD

Age (years) Education (years) Memory (Wechsler Memory Scale)

Story 1 immediate recall 13.2 3.1 Story 2 immediate recall 8.4 3.2 Visual immediate recall 10.1 2.1 Digits forward 6.6 1.2 Digits backward 4.9 1.2

Cognitive function Block score (WAIS) 30.2 7.6

Perceptual motor speed Slotted pegboard (sec) 66.9 16.2 Choice reaction times (msec) 577 118 Trails A (sec) 27.8 8.8 Trails B (sec) 63.8 28.6

Peripheral sensory Fingers right (errors) 2.5 2.4 Fingers left (errors) 2.2 2.1

Balance (sway) distance Eyes open (cm/sec) Eyes closed (cm/sec)

POMS score (affect)

(n = 104) (n = 110) (n = 53) (n = 83) 40.4 10 40.7 10 37.5 9.8 38 9.3 14.3 1.1 14.1 1.0 14.1 1.3 14.7 1.2

12.7 3.6 12.2 3.7 12.7 3.7 8.3 3.3 7.7 3.5 8.5 4.0

11.8 1.4 10.8 2.6 10.5 2.3 7.0 1.4 6.7 1.2 6.8 1.3 4.9 1.3 4.7 1.2 4.8 1.0

33.2 7.3 33.5 9.1 31.8 9.9

68.9 11.0 68.2 11.1 64.5 8.4 570 108 617 116 606 109

30.9 8.9 32.9 9.3 32.9 10.1 63.6 26.0 63.4 25.5 66.2 23.0

1.5 1.6 1.3 1.4 2.2 2.1 1.7 1.7 0.8 1.0 1.5 1.6

0.94 0.53 1.12 0.48 1.07 0.46 1.14 0.41

16.3 6.0 8.1 5.6


numbers, and block design scores. Similarly, performances on pegboard and trail making A and B did not vary significantly across this interval. A systematic increase in two-choice visual reaction time from 1983 to 1985 was due to the change in computer keyboard. Although finger writing appeared to have fewer errors after the first group the differences were random. Balance did not change from 1985 to 1986. Previous testing using the sharpened Romberg (Freghy and Graybill, 1968) showed a mean standing time of 25 to 32 sec for each year's group. POMS score was lower in 1986 compared to that in 1985 (8.1 vs 16.3).

Scores on all tests were normally distributed, except for choice reaction times which have been discussed previously (Miller et al., 1989) and for balance using the sharpened Romberg, a test rejected because of skewing. Attempts to convert the sharpened Romberg from a pass-fail outcome (Freghy and Graybill, 1986) to one with a continuous variable (time of stable standing) were unsuccessful. Even when seconds of standing heel to toe were used, about 20% of this population failed before 20 sec. This led to development of a measurement of sway speed during a minute of standing eyes open and then eyes closed (Neurotest, Inc., Corona, CA).

The 19 subjects who were retested yearly after 1982 were 6 years older than the group mean in 1982 but their years of education were identical (Table 2). Their mean scores for recall of verbal and visual information were exactly the same in 1982 as those in the naive subjects. Their verbal recall improved across the first year and thereafter remained constant. Visual recall and digits recalled did not change across any interval. Block design score improved across the first two intervals (total of 3 years) but then became constant and was above the scores for the naive subjects of those last 2 years.

Perceptual motor speed in pegboard, two-choice visual reaction time, and trail making A were remarkably consistent and similar in each group of naive subjects. There were small increases which were predicted by regression coefficients for age; however, trail making B showed no trend. Finger perception showed no significant changes. Balance speed was not significantly different from 1985 to 1986. POMS score (mean) was significantly lower on its second administration in 1986 versus that in 1985. This trend which was similar to that of the naive subjects is unexplained.

The 299 subjects who were tested two or three times across the 4-year interval were also older than the naive subjects (Table 3). They showed improvements from 1982 to 1983 in verbal and visual recall and block design score similar to those of the 19 retested subjects but there was no change in recall of digits. Perceptual motor speed and attention indices showed no sustained trends over the intervals nor did number recognition in the fingers. Balance, speed of sway, was identical in 1985 and 1986. POMS scores became lower on retest as was seen in the 19 retest groups and in the naive histotechs.

Age coefficients for the neurobehavioral tests were calculated by linear regression equations for 350 subjects at their first testing (naive) after modeling had eliminated those independent variables which had statistically insignificant coefficients. There were no sex differences in test scores so the sexes were combined for analysis. Variables dropped included educational level, formaldehyde expo-


TABLE 2 DESCRIPTION AND NEUROBEHAVIORAL TESTS (GRouP MEANS) AND STANDARD DEVIATIONS FOR 19

SUBJECTS RETESTED YEARLY

Year

1982 1983 1985 1986


Age (years) 46 10.6 47 10.6 49 10.6 50 10 Education (years) 14.4 0.9 14.4 0.9 14.4 0.9 14.4 0.9 Memory (Wechsler Memory Scale)

Story 1 immediate recall 13.0 3.6 15.1 2.7* 14.3 3.9 14.9 3.9 Story 2 immediate recall 8.8 2.9 10.8 3.1" 11.7 3.3 11.2 4.8 Visual immediate recall 9.9 3.4 10.6 2.2 11.1 2.6 11.8 2.0 Digits forward 6.5 1.3 6.4 1.2 6.6 1.1 6.3 1.1 Digits backward 4.7 1.4 4.7 1.2 4.7 0.8 4.9 1.2

Cognitive function Block score (WAIS) 32,2 7.3 35.4 9.3 35.6 7.7 35.7 9.7

Perceptual motor speed Slotted pegboard (sec) 67.5 15.5 68.4 9.3 72.3 16.0 70.5 15.5 Choice reaction time (msec) 571 89 545 59 655 134 620 105 Trails A (sec) 28.2 7.6 30.2 13.0 34.2 12.8 28.9 9.0 Trails B (sec) 64.3 25.1 65.2 29.3 69.6 39.5 61.8 23.3

Peripheral sensory Fingers right 1.8 2.0 1.7 1.8 1.7 1.7 1.5 1.8 Fingers left 1.4 1.5 1.6 1.6 1.3 1.6 1.7 2.4

Balance (sway) distance Eyes open (cm/sec) 1.12 0.52 1.22 .42 Eyes closed (cm/sec) 1.12 0.52 1.19 .38

POMS score (affect 39 14 13 6

* P < 0.05.

sure (hours per day), solvent exposure (slides covered per day), and years of histotech exposure. POMS score (composite) had no significant age and educational level coefficients, nor did any of its six components. As an independent variable, POMS score had no significant coefficients for neurobehavioral test scores. There were significant negative age coefficients for verbal (story 1) and visual recall and block design score and significant positive age coefficients for pegboard time, choice reaction time, trail making A and B times, and errors on right finger writing. The variance explained (r 2) was from 0.075 for block design down to 0.028 for story 1 (Table 4). For the 144-subject test group of 1989, educational level had a significant positive coefficient for Culture Fair A and a significant negative coefficient for two-choice visual reaction time, in addition to their coefficients for age of reverse sign (Table 5). Grip strength, body fat content, sway speed with eyes open and with eyes closed, and POMS were also modeled for age and educational level. Grip strength, body fat content, Culture Fair, sway speed with eyes open, blink (R-1 response on the left), and CRT had significant age coefficients but these coefficients for sway speed with eyes closed and blink R-1 on the right were insignificant. Culture Fair mean score was most dependent on age and educational level (r z = 0.25 and 0.34, respectively).


TABLE 3 DESCRIPTIVE AND NEUROBEHAVIORAL TESTS (GROUP MEANS) AND STANDARD DEVIATIONS FOR

299 RETESTED SUBJECTS

Year

1982 1983 1985 1986


Age (years) Education (years) Memory (Wechsler Memory Scale)

Story 1 immediate recall 13.9 3.5 Story 2 immediate recall 8.4 3.1 Visual immediate recall 9.0 3.3 Digits forward 6.1 1.2 Digits backward 4.6 1.3

Cognitive function Block score (WAIS) 31.6 7.5

Perceptual motor speed Slotted pegboard (sec) 71.7 18.5 Choice reaction time (msec) 562 86 Trails A (sec) 28.0 7.8 Trails B (sec) 69.4 28.4

Peripheral sensory Fingers right (errors) 2.4 2.4 Fingers left (errors) 1.5 1.5

Balance (sway) distance Eyes open (cm/sec) Eyes closed (cra/sec)

POMS score (affect)

(n = 47) (n = 98) (n = 91) (n = 63) 44.0 10.2 46.1 10.1 46.1 10.1 47.9 10,2 14.1 1.7 14.1 1.8 14.1 1.7 14,6 1,8

14.4 2.6 14.4 3.4 15.4 3,7 10.2 2.6 11.2 3.0 10,6 4,4 12.2 3.0 11.0 2.4 11.3 2.1 6.5 1.1 6.6 1.1 6.4 1,2 4.8 1.13 4.9 1.0 4.6 1.2

34.2 8.6 33.0 8.4 34.0 9.7

69.9 11.8 76.5 18.6 75.1 18.6 558 63 623 152 606 101

34.6 15.7 34.5 12.6 29.9 9.2 69.4 34.8 69.2 39.1 69.2 28.7

2.3 2.1 1.4 1.4 2.3 2.5 1.7 1.6 1.1 1.7 2.3 3.2

1.04 0.42 1.04 0.36 1.17 0.45 1.17 0.38

34 8.0 11 5.4

DISCUSSION

1. In repeated samples of naive technicians working in histology a neurobehavioral battery gave virtually identical results from year to year.

2. For the 19 technicians tested yearly, verbal recall (memory) mean score increased from the first to second testing and then stabilized. All other tests showed no differences from year to year or across the entire interval.

3. The 299 technicians retested one or more times showed slight improvements in means for verbal and visual recall, block design score, and trails A time, but these were not statistically significant.

4. The slight improvements in recall, which were opposite the age decrements, may reflect test familiarity or effects of learning.

5. Retesting at yearly intervals produced no measurable effects on neurobehavioral performance after the first interval. Such stability would assist in assessing the effects of cumulative toxic exposures across long intervals (years) or the effects of possible therapeutic interventions. Retesting across shorter intervals, especially within 30 days, may have greater effects on performance and should be studied.

142 K1LBURN A N D W A R S H A W

T A B L E 4 COEFFICIENTS OF REGRESSION OF NEUROBEHAVIORAL TESTS WITH AGE, P VALUES OF

COEFFICIENTS, AND PROPORTION OF VARIANCE EXPLAINED BY AGE (r 2) IN 350 ALL-NAIVE HISTOLOGY TECHNICIANS

Cons tan t Coefficient P r 2

Verbal recall Story 1 15.29 - 0.0674 0.003 0.028 Story 2 9.26 - 0.0272 0.220 0.006

Visual recall 13.50 0.0654 0.000 0.056 Digits forward 7.29 - 0.0119 0.121 0.005 Digits backward 5.26 - 0.0110 0.180 0.003 Block design score 38.11 • - 0.1636 0.002 0.033 Pegboard 54.92 0.3292 0.000 0.075 PB errors 0.29 0.0019 0.531 0.001 Choice RT 480 3.086 0.000 0.062 Trails A 24.58 0.1711 0.012 0.019 Trails B 33.31 0.7964 0.000 0.091 Errors Trails B 1.31 0.0173 0.325 0.000 Fingers right - 0.29 0.0568 0.000 0.044 Fingers left 0.41 0.0332 0.207 0.002

6. These histology technicians showed no evidence of progressive impairment from formaldehyde and solvent exposures across 4 years.

This new evidence did not confirm neurobehavioral deficits related to hours per day of formaldehyde exposure, suggested by regression analysis in an earlier cross-sectional study (Kilburn et al., 1987). This absence of long-term effects agrees with a published study of insignificant short-term effects of formaldehyde (Andersen and Molhane, 1983). Before concluding that formaldehyde with these solvents are innocuous, it should be recalled that dose estimates for these chronic exposures to formaldehyde and solvents were not possible. Airborne levels of these fumes depend upon temperature, air exchanges, and worker location with regard to release sites so that reliable estimates would require personal monitoring over the years of employment. Such data have been collected rarely and then only for brief intervals (Kilburn et al., 1985b).

Analysis of chronic neurobehavioral effects of ongoing low level exposures to chemicals should take into account two competing influences. The first is that small (but statistically significant) improvements in recall and in constructional and cognitive function may occur especially across the first retest interval due to test familiarity and the second is that many of these functions gradually deterio- rate with aging.

Verbal recall improved across the first interval and then became constant despite a significant negative age coefficient. This improvement could have ob- scured an interval decrement from exposure. That the retest improvement was small is important and reduces the need to distinguish what it is due to, learning (memory), practice, familiarity (test ease), or other factors. The number of repetitions of a test needed to reach stable performance has been addressed particularly carefully in designing testing to detect loss of vigilance over a working day's

N E U R O B E H A V I O R A L TESTS ON HISTOLOGY T E C H N I C I A N S 143

T A B L E 5

COEFFICIENTS OF AGE AND EDUCATIONAL ATTAINMENT LEVEL WITH NEUROBEHAVIORAL TEST

SCORES IN 144 SUBJECTS STUDIED IN 1989 WITH P VALUES AND PROPORTION OF VARIANCE

EXPLAINED BY THE COEFFICIENT (r 2)

P values of coefficient Constant Coefficient P r 2

Grip right 44.06 - 0.1771 0.043 0.023

Grip left 41.78 - 0.1928 0.030 0.028

Fat 30.21 0.1311 0.034 0.029

Lean weight 108.95 - 0.1371 0.522 - 0.005

Culture fair A 46.50 -0 .3732 0.000 0.250

Education level 28.87 1.152 0.000 0.344 a

Sway Distance eyes open (3) 0.93 0.0069 0.072 0.018

Distance eyes closed (3) 0.95 0.0084 0.063 0.020

Speed eyes open (3) 0.606 0.0068 0.009 0.045

Speed eyes closed (3) 1.01 0.0066 0.148 0.009

Choice RT 433.0 1.858 0.005 0.053

Education level 572.2 - 9.107 0.012 0.093 a

Simple RT 233.0 0.889 0.026 0.031

Blink (R-l)

Right 11.38 0.039 0.098 0.016

Left 10.61 0.1256 0.006 - 0.058

POMS score 10.48 0.1072 0.625 - 0.006

Tension 8.72 - 0.0238 0.608 - 0.006

Depression 4.39 0.0424 0.55 - 0.004

Anger 7.08 - 0.0544 0.273 0.002

Vigor 19.99 - 0.0283 0.604 - 0.006

Fatigue 5.86 0.0882 0.078 0.016

Confusion 4.44 0.0264 0.375 - 0.002

a Combined (age and education level).

exposure to styrene or other solvents (Gamberale and Kjellberg, 1983). Simple reaction time exemplifies a test which stabilizes after a few repetitions.

There was practical identity of various test scores in four samples of histology technicians (naive subjects) who attended the national convention. As the subjects were self-selected, the similarity of their performances suggested that the subjects who volunteered year after year represented a reproducible sampling of the histotech population attending the annual convention. No claim is made that they were the same as a random sample. However, even after random selection, subjects invariably divide into participants and nonparticipants which may pro- duce bias in an all-volunteer sample.

In examining factors other than exposure which influence scores in neurobehavioral testing, age as discussed and endowment of ability for cognitive and memory capacity or intelligence are obviously important. Advancing age reduced performance of reaction time, recall, and perceptual motor tasks. Age coefficients were developed by linear regression modeling, after elimination of sex, POMS score, years of employment, years of cigarette smoking, and nonoccupational exposures, including those to solvents, drugs, and alcohol. (Histories of alcohol and drug use were practically absent in this population.) Educational attainment

144 K I L B U R N A N D W A R S H A W

is the only easily obtained estimate of endowment of ability or intelligence. As years of school, it was correlated with higher Culture Fair scores and with faster choice reaction times. However, there were no significant effects on other scores in this battery so modeling effects of age did not need to include years of school. These conclusions must be considered tentative because mean educational attainment was 14 years and the range was narrow. Only two subjects were below 12 years (the threshold for job entry) and 17 was the highest level. This truncated distribution probably reduced the statistical power of the coefficients for educational level. Thus these conclusions may not be transferable to groups normally distributed for educational attainment, those with a mean educational attainment of 12 years, or those with wider ranges. To test for representativeness for the general population, the relationships we studied need further analyses, applying this approach in populations which are normally distributed.

Despite these limitations, the near identity of the test results in the four naive groups and the examination of coefficients for age and for educational attainment bring us closer to prediction equations for each neurobehavioral test, analogous to those already generated for simple and choice reaction times (Miller et al., 1989). For initial evaluation of an exposed population such comparisons could help to estimate the probability of abnormality in subsequent studies; i.e., abnormality is likely when mean test results are more than two standard deviations from the mean.

This would be a useful and economical approximation for assessing results of case series and pilot studies. Unquestionably there are many advantages to having a matched control group which is unexposed to the suspected toxic agent or environment but in all other respects is "identical" to the study group. However, when budgets are lean matched controls which double the work and cost of exploratory studies often compromise the size of the exposed group and its age spread. Also it is increasingly difficult to find and recruit totally "unexposed" occupational or environmental populations as referent groups. Finally, even if a case-match comparison of neurobehavioral test results is made additional comparisons to predicted values adjusted for age and educational attainment would provide assurance that the controls are uneffected, can be used as referents, and do not reflect effects of unknown exposures. It is analogous to the routine comparison of results of pulmonary function tests to predicted values based on height and age for individual subjects or populations (Morris et al., 1971; Kilburn et al., 1989). Standardized neurobehavioral examinations can be used for surveillance of effects of toxins on groups across time (Gamberale and Kjellberg, 1983). After the first retesting, which may show marginal improvement of recall and some cognitive functions, stability is reached or at least age-related deterioration offsets improvement and these variations are modeled effectively by linear regression equations. Long-term exposures associated with steeper than expected declina- tions of neurobehavioral function with age and/or cancellation of the improvement due to repetition should trigger further investigation to distinguish neurotoxicity from confounding.

Coefficients for age and educational level need to be developed for populations with normal age ranges and extended ranges of educational attainment. Unfortu-


nately, many published normative values are inappropriate because they were based on college freshmen or other selected subjects who differ from occupational groups. Other surrogates for endowment, especially for memory and cognitive function, should be explored in this manner. However, the need for these devel- opments should not restrict neurobehavioral surveillance (prospectively), especially from the onset of exposure in workers exposed to hazardous materials such as firemen, toxic cleanup crews, and hazardous materials workers. Selected neuropsychological testing of individual subjects obtains vital information rapidly and efficiently. Applying coefficients for age and educational level to calculate predicted performance permits comparison of test results to predicted scores as percentage of predicted. Such comparisons would help the clinical scientist and physician make critical decisions about neurotoxicity and screen patients intelli- gently for referral, as suggested by Geschwind (1977).

REFERENCES

Andersen, I., and Molhane, L. (1983). Controlled human studies of formaldehyde. In "Formaldehyde Toxicity" (J. E. Gibson, Ed.), pp. 154-165. Hemisphere, New York.

Baker, E. L., Smith, T. J., and Landrigan, P. J. (1985). The neurotoxicity of industrial solvents: A review of the literature. Am. J. Ind. Med. 8, 207-217.

Cattell, R. B. (1951). Classical and standard score IQ standardization of the IPAT: Culture free intelligence scale 2. J. Consult. Psychol. 15, 154-159.

Cattell, R. B., Feingold, S. N., and Sarason, S. B. (1941). A culture free intelligence test. II. Evalu- ation of cultural influences on performance. J. Educ. Psychol. 32, 81-100.

Elofsson, S. A., Gamberale, F., Hindmarsh, T., et al. (1980). Exposure to organic solvents. Scand. J. Work Environ. Health 6, 239-270.

Federal Panel on Formaldehyde (1982). Report of the federal panel on formaldehyde. Environ. Health Perspect. 43, 139-168.

Freghy, A. R., and Graybill, A. (1968). An ataxia test not requiring rails. Aerosp. Med. 39, 177-282. Gamberale, F., and Kjellberg, A. (1983). Behavioral performance assessment as a biological control of

occupational exposure to neurotoxic substances. Adv. Biosci. 46, 111-121. Geschwind, N. (1977). In "The Mental Status Examination in Neurology" (R. L. Strub and F. W.

Black, Eds.). Davis, Philadelphia. Hanninen, H., Eskelinen, L., Husman, K. A. J., and Nurminen, M. (t976). Behavioral effects of

long-term exposure to a mixture of organic solvents. Scand. J. Work Environ. Health 4, 240-255. Kilburn, K. H. (1989). Is the human nervous system most sensitive to environmental toxins? Arch.

Environ. Health 44, 343-344. Kilburn, K. H., Warshaw, R., Boylen, C. T., Johnson, S. J. S., Seidman, B., Sinclair, R., and

Takaro, T. (1985a). Pulmonary and neurological effects of formaldehyde exposure. Arch. Environ. Health 40, 254-260.

Kilburn, K. H., Seidman, B., and Warshaw, R. (1985b). Neurobehavioral and respiratory symptoms of formaldehyde and xylene in histology technicians. Arch. Environ. Health 40, 229-233.

Kilburn, K. H., Warshaw, R. H., and Thornton, J. C. (1987). Formaldehyde impairs memory, equilibrium and dexterity in histology technicians: Effects which persist for days after exposure. Arch. Environ. Health 42, 117-120.

Kilburn, K. H., Warshaw, R. H., and Thornton, J. C. (1989). Pulmonary function in histology technicians compared with women from Michigan: Effects of chronic low dose formaldehyde exposure on a national sample of women. Br. J. Ind. Med. 46, 468--472.

Kilburn, K. H., Warshaw, R. H., and Hanscom, B. (1991a). Are hearing loss and balance dysfunction linked in construction iron workers? Br. J. Ind. Med., submitted for publication.

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


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

McNair, D. M., Lorr, M., and Droppleman, L. F. (1971/1981). "Profile of Mood States." Educational and Industrial Testing Service, San Diego, CA.

Mello, N. K. (1975). Behavioral toxicology: A developing discipline. Fed. Proc. 34, 1832-1834. Miller, J. A., Cohen, G. S., Warshaw, R., Thornton, J. C., and Kilburn, K. H. (1989). Choice (CTR)

and simple reaction times (SRT) compared in laboratory technicians: Factors influencing reaction times and a predictive model. Am. J. Ind. Med. 15, 687-697.

Morris, J. F., Koski, A., and Johnson, L. C. (1971). Spirometric standards for healthy nonsmoking adults. Am. Rev. Respir. Dis. 103, 57-67.

Morrow, L. A., Ryan, C. M., Hodgson, M. J., and Robin, N. (1990). Alterations in cognitive and psychological functioning after organic solvent exposure. J. Occup. Med. 32, 444--450.

Reitan, R. M. (1958). Validity of the trail-making test as an indicator of organic brain damage. Percept. Motor Skills 8, 271-276.

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.). Aca- demic Press, New York.

Swenberg, J. A., Kerns, W. D., Mitchell, R., Gralla, E. J., and Pavkov, K. L. (1980). Induction of squamous cell carcinomas of the rat nasal cavity by inhalation exposure to formaldehyde vapor. Cancer Res. 40, 3398-3401.

U.S. Environmental Protection Agency (USEPA) (1990). "National Priorities List, Supplementary Lists, and Supporting Materials." USEPA HW-10-145. USEPA, Washington, DC.

Valciukas, J. A., and Singer, R. M. (1982). "The Embedded Figures Test in Environmental and Occupational Neurotoxicology." Environ. Res. 28, 183-198.

Warshaw, R. H., Hanscom, B., and Kilburn, K. H. (1992). A reproducible trigger for the measurement of Trigeminal latencies elicited by globuller tap Environ. Res., in press.

Wechsler, D. (1945). A standardized memory scale for clinical use. J. Psychol. 19, 87-95. Wechsler, D. (1971). "Adult Intelligence Scale Manual" [revised]. The Psychological Corp., New

York.

Neurobehavioral effects of formaldehyde and solvents on histology technicians: Repeated testing...

Documents

Transcript of Neurobehavioral effects of formaldehyde and solvents on histology technicians: Repeated testing...