Subjective symptoms and psychological performance of chlorine-alkali workers.

Subjectivesymptoms and psychologicalperformance of chlorine-alkaliworkers. 1989;15:69--74. Subjective symptoms and psychological performances on a computer-administered test battery were studied among a group of 60 chlorine-alkali workers and their matched referents. The exposure time of the group exposed to inorganic mercury vapor had been about 14 years, and the estimated long-term average exposure had been about 25 ug/m' of air. The exposed workers had an actual mercury concentration of 51.9 (SD 25.0, range 15-150) nmolll in blood and 84.1 (SD 56.6, range 15-260) nmolll in urine. Neither perceptual motor nor memory nor learning abilities of the mercury-exposed workers showed any disturbances in a comparison with the referents. However, the exposed group reported statistically significantly more memory disturbances than the referents. Strain caused by three-shift work was a possible cofactor for other increased subjective symptoms, namely, for sleep disorders, fatigue, and confusion.

The central nervous system is the critical target organ during long-term exposure to organic mercury vapor (1). Initial symptoms in intoxications due to elemental mercury have ranged from atactic intention tremor to a bizarre constellation of mental changes (2). Nonspecific subjective symptoms such as fatigue, weakness, loss of memory, insomnia, and increased irritability have been characterized as micromercurialism in subjects exposed for long periods of time to low concentrations of mercury vapor (3,4). In psychological tests disturbances of cognitive and psychomotor functions have been associated with low levels of mercury exposure (5)(6)(7)(8)(9). However, conflicting data have also been reported for workers at similar levels of exposure (l0, 11).
The purpose of the present study was to discover whether exposure conditions of about 25 ug of mer-cury1m 3 of air, as estimated by repeated analyses of blood mercury concentrations, had resulted in any measurable increase in subjective symptoms or disturbances on a psychological test battery.

Materials and methods
Subjects and study design Subjective symptoms and performances on psychological tests were recorded by means of a cross-sectional design. The design was longitudinal for the exposure assessment, which was estimated from earlier blood mercury analyses. After exclusions the exposed group consisted of 60 men who had been exposed to elemental mercury vapor in a chlorine-alkali plant for at least five years. The exclusion criteria were neurological, psychiatric and metabolic diseases, abuse of alcohol, skull injuries, and hypertension checked from the records of the periodic health examinations performed twice a year, as well as from possible case records. Seven of the workers who fulfilled the exposure criterion were excluded from the study group. Four of them had chronic hypertension, and the others suffered from dipsomania (identified by the occupational health staff of the company). In addition, since 1972, seven other workers had been moved to unexposed tasks, two of them because of chronic respiratory disease and the others because of repeatedly increased concentrations of mercury in their blood. The turnover of personnel without known mercury-related problems could not be controlled. No cases of mercury intoxication had been diagnosed among the chlorine-alkali workers since the 1960s.
The mean age of the exposed group was 38.0 (SD 6.6, range 26-56) years. Their average exposure time was 13.7 (SD 5.5, range 5-28) years. The exposed workers had had general and vocational education for a mean of 10.0 (SD 2.6, range 3-15) years. One subgroup (N = 37) had three-shift work with periods of four workdays and 2 d of rest. The other exposed workers (N = 23) had regular daywork with periods of five workdays and 2 d of rest. The mean age was 38.7 (SD 7.2, range 26-56) years for the three-shift workers and 36.7 (SD 5.4, range 29-49) years for the day workers.
One referent, matched for age (± 1 year), sex, vocational status, and length of education, was chosen for each exposed worker. The referents (N = 60) were employed in mechanical woodprocessing plants. The exclusion criteria for the reference group were earlier oc- cupational exposure to known neurotoxic chemicals or a history of the same chro nic disea ses and injuries listed as exclusion criteria for the exposed gro up. Reject ion was based on medical records and on the jud gment of the occupational health sta ff of th e emplo ying company. The mean age of the referents was 38.2 (SD 6.6, range 28-55) years . Their general and vocational education had lasted for a mean of 9.4 (SD 1.8, ran ge [6][7][8][9][10][11][12][13][14] years. All the referents ha d regular da ywork.

Examinations
At 1300 th e subj ects cam e to the hospital where th e exam inati ons took place . Except for two of the referent s, th e dayshift workers had had five a nd the thre eshift workers six workhou rs before their arri val. Only one sub ject was examined per day. Special attention was given to standardizat ion of the timin g and to adherence to a strict examinati on procedure. At the beginning of th e examinations every subject was int erviewed according to a fixed schedule covering med ical and occupational history, present health status, and life style. Furthermore, a qu estionnaire that the subjects had filled out earlier on subjective symptoms was checked and completed if necessary. Th e interview was followed by a clinical examination and the adm inistration of computer-administered psychological tests. All th e interviews and clinical examination s were performed by one of the authors (LP). On the basis of the int erviews and ob servations made in the clinical examinat ions, the health sta tus of the subjects was judged to be good . The subjects did not report any cur rent use of medicines affecting th e centra l nervou s system, and they had not consumed mor e alcohol th an

Test
Functi onal dom ain Test parameters two bottles of beer (about 0.7 I) during the pre viou s evening.
Durin g the examinat ion veno us blood samples were taken, and the following measurements associated with increased consumpt ion of alcohol were performed (12): gamma glutamyl transferase in serum, aspartate aminotransferase in seru m, and the mean cellula r volum e of erythrocytes. The activty of the two enzymes was analyzed according to the recomm endations of th e Co mmittee on Enzymes of the Scand ina vian Society. The erythrocyte value was determ ined with th e use of an automat ic cell co unter .
Examination ofsubjective symptoms. Subjective symptom s were -s urveyed with the use of the follo wing two methods: (i) a mod ification of the questionnair e used by Husman (13) and (ii) a computer-administe red mood questionnaire included in the psychologi cal test procedure of the Neurobehavioral Eva luation System ada pted by Bak er & Letz (14).
On th e first qu estion na ire the answers co ncerni ng subjective symptoms (N = 31) had th e follo wing thr ee choices: (i) sympt om occurs hardly ever (rating I), (ii) symptom occurs now and then or to some degree (rating 2), and (iii) sympto m occur s often or to a disturbing degree (rating 3). The following symptom variables were form ed and used in the calculations: distu rbance of memor y (score 2-6), difficulties to con centr at e (score 2-6), fatigue (score 2-6), sleep disorders (score 2-6), and tremor of the hands (score 2-6).
The inventory of a ffects included in the computeradministered test batt ery is a modi fication of the profile Mood States (15). One of us (HH) translated and ad apted the items into Finni sh . In this inventor y 25 items (rating 1-5) were used to get a five-dimensional mood pro file called Mood Sta tes. Th e states focused o n were tension, dep ression , anger, fatigue, and confusion (ie, feelings of embarrassment and helplessness). The means of the item scores were used as test parameters in th e study.  (14) . Th e tests and test pa ramet ers used in the st udy are shown in ta ble I. Every session took place in the same quiet and comforta ble room , and th e same personal computer equi pment was used. The subjects had been informed about the testing method beforehand , and the point had been firmly stress ed that the aim of the examination was not to measure general intelligence levels. The mot ivation of the subjects was evaluated by means of a selfadministered sca le, o n which the degree of effo rt in t he tests was rated by every subjec t on a scale fro m very poor (ratin g I) to very hard (ra ting 5). T he mot ivatio n of th e subjects was also judged on th e basis of observations of their concent ration durin g the session . Number The figures of the indicators of actual exposure are shown in table 2. Three of the exposed subjects had a blood concentration of organic mercury that was higher than the inorganic fraction. No referent had more inorganic than organic mercury in their blood.
The exposed day workers had significantly higher concentrations of mercury in their blood and urine than the exposed three-shift workers (figure I). The difference was also seen for the indicators of long-term exposure. The TWA concentration of mercury in the blood of the day workers was 63.3 (SO 13.5, range 37.5-82.4) nmol/I, and that of the three-shift workers was 43.3 (SO 16.5, range 24.7-90.0) nmol/I.
The expo sed and reference workers reported similar average drinking habits (times per week or quantity of intake per time). The cutoff points for the division into subgroups were both the medians of the dose indicators and the form of workshift. The enzyme and erthrocyte measurements did not show any differences between the exposed and reference groups, except for  (18). Both the inorganic and organic fractions of mercury were determined with this method, which has been described earlier along with the methods used for quality control (19).
The dose indicators of the actual exposure of the study groups were the inorganic mercury level of the blood, the organic mercury level of the blood, the total mercury level of the blood , and the urinary mercury level. Morning urine samples collected at home on the morning of the examination da y were used for the urinary mercury analyses . A venou s blood sample (25 ml) was taken during the examination. The polyethylene blood tubes (containing 0.15 ml of heparin, Medica " 5000 U'/ml, Finland) and the urine flasks had been analyzed and found to be free of mercury contamination. The modification of Magos & Cernik (17) was also applied in the analyses of the actual mercury concentration of the blood and urine. The urinary mercury level was corrected for the urinary creatinine concentration according to Jaffe's standard method (20).

Characteristics of the study group
The individual TWA of the blood mercury concentrations was based on 22 (SO 5.8, range 8-32) measurement s. The mean TWA blood mercury concentration of the chlorine-alkali group was 51.3 (SO 15.6, range 24.7-90.0) nmol/1. The mean TWA of the inorganic mercury in the blood was 29.6 (SO 3.3, range 9.8-64.1) nmol/l, which was ba sed on 19 (SO 4.0, range 8-27) measurements. For 11 exposed subjects the TWA concentration of the organic mercury in the blood exceeded the TWA concentration of the inorganic mercury in the blood .    Table 4. Comparis o n of scores for SUbjective symptom s between t he exposed workers and their re ferents grouped ac c o rding to the form of workshi ft of the exposed w o rkers. T he pai red t-tsst was used for comparison . one subgro up. The more highly exposed subgroup , whose urinary mercury level was 8.2 umol /rnol of creatinine or more had a higher serum gamma glutamyl transfer ase level [mean 31.3 (SO 26.5) U/I ) and a higher mean cellular volume of erthrocytes [mean 92.8 (SO 3.6) tl) than their referents [gamma glutamyl transfera se: mean 21. 8 (SD 9.9) U /I; mean cellular volume of erthrocytes: mean 90.9 (SO 4.0) fl], but no difference in the activity of the serum aspartate aminotransfera se. The mean number of teeth with amalgam fillings was 5.0 (SO 5.3, range 0-17) for the expo sed workers and 4.5 (SO 5.5, ran ge 0-17) for the reference workers.

Diff erence in score me ans
All the items of the psychological test battery were performed by every subject. Seven of the exposed work ers and 14 of the referents evaluated their degree of effort in the test perfor mances as mode rate (rating 3 on the scale 1-5). Two exposed workers but none of the refe rents evaluated their effort as rather poor (rating 2). All the other subjects had tried hard or very hard (ratings 4 and 5). However, on the basis of the ob servations made during the session s, no performan ces were excluded from the calculations.

Subjective symptoms
The expo sed workers had systematically more subjective symptoms than their referents (table 3). Th e multivariate anal ysis of variance of the symptom scor es was sta tistically significant (F = 3.91, P <0.001) . The differences between the exposed and referen ce groups were highl y significant for memory disturbances and sleep disorders. The chlorine-alkali workers also reported significa ntly more ange r, fat igue, and confusion .
To control the confounding effect of three-shift work, we divided the exposed workers into subgroups of three-s hift workers and day workers. First, the symptoms of the exposed sub gro ups were compared with the symptoms of their referents (tabl e 4). Both subgroups reported significantly more disturbances of memory than their referents. Tremor of the hands had been more common in the more highly expo sed da y worker group. The three-shift workers had more complaint s of sleep disorders, fatigu e, and co nfusion than their referents. Second , comparison s between the exposed subgroups showed that the more highly exposed day workers had reported significantly more memory symptoms than the lowly exposed three-shift workers (t = -3.20, P < 0.0 I) . A multi variate anal ysis of varian ce confirmed the differences between the gro ups . Third , a two-wa y analysis of variance was used to detect th e dependence of the symptoms of the chlorinealkali workers on exposure, indicated by the blood and ur inar y mercury levels, and on the form of workshift. According to thi s analysis the memory disorders were signifi cantl y associated with the form of work shift  No other significa nt asso ciations were found bet ween th e sym pto ms, level of expos ure, and th e fo rm o f workshift.

Performan ces on the psychological tests
The results of the performances on the psy ch ological tests are shown in table 5. A sign ificant difference between the expo sed and reference gro ups was show n by th e multivari at e analysis o f va ria nce (F = 2.72, P < 0.05 ). Unexpectedl y, the performa nce of th e exposed workers o n th e hand-eye coordinatio n test was significa ntly bett er than that o f th e referen ce workers , a nd the exposed workers per formed as well as their referents o n th e othe r tests . Th e exposed workers were a lso gro uped accordi ng to th e medi ans o f th e dose indicators (TWA blood mer cury concentration 49.1 nmolll, inorgan ic mercury in blood 30 nm olll , tota l concentration of mercury in blo od 45.0 nmolll , and ur inary mer cur y co ncentratio n 8.2 umol/rnol creatinine) into high a nd low exp osure subgro ups, as well as acco rdi ng to th e fo rm o f wor kshift. Even so th e perfo rman ces o f th e exposed subgroups were simi lar to those of th eir referents. Furthermore, th e test performances of th e expos ed three-sh ift wo rkers a nd th e day wo rk ers were compared. Ho we ver , no significa nt d iffer en ces were fo und bet ween the gro ups. The twoway an al ysis of va riance even sho wed no sig nifi ca nt interactio ns betw een the test perfo rma nces, the level of exposure, and the form o f workshift.

Discussion
In th is study the psychomotor , memory, and learning a bilit ies of the expos ed wo rkers were simila r to those o f their mat ched referents o n the psychological tests. The mercury co nce ntratio n in the blood and ur ine indica ted both a low long-term a nd a lo w actual expos ure level of th e chlo rine-a lkali workers . Adverse effects on hand steadiness (5, 10) and on sho rt-term memory (6,21) have earlier been shown at similar levels of mercury exposure. T he selection and motivation of the subjects may have influenced the results of our study howeve r. The wo rke rs with lo w personal hygien e at wor k, who po ssibl y were un inclined to follo w wo rk safety instructions and po ssibl y also had problems with cogniti ve abiliti es as well , had been excluded fro m mercu ry-expos ed tasks. The motivati on of the referen ce gro up was lo wer than that of the ex pos ed gro up, even though all th e re fere nts coo perated adeq ua tely in the exa m inatio ns. The exposed workers complained of more numerous subjective sym ptoms than did their referents. Simila r symptoms ha ve earlier been de scribed in co nnection with long-term low exposur e to mercu ry vapor (3). In the st udy of Roel s et al (10) the expos ure ind icat ed by the blo od mercur y con centrations was eq ual to tha t in our study , a nd the wo rkers ha d also had subjective sympto ms simila r to those of our expose d group, ie, memory disturbances, fati gue , and irritabilit y. In our st udy th e results o f th e memory te sts of th e exp osed wo rk ers did not co n fir m the ir subject ive memory sympt o ms, bu t it sh o uld be rem embered that tests of learning and memory can never co ver all the memo ry funct ion s needed in ever yday life. In the interview the exposed workers de scribed their symptoms as absentmindedness and lapses of memory in small ordinary pursuits. However, on e expla natio n for the se reports might be that disturb an ces of mem ory had been routinel y as ked ab out in the periodic health examinations, and therefo re th e ch lo rine-alkali workers might ha ve been guided to emphasize self-observa tion and perhaps to exaggerate such sym p to ms . Indeed , the memory symp to ms showed a stro nger associa tion with the form o f workshift (namely, daywork) th an with the level of mercury exposure. St rain in the three-shift work was a p ossible cofactor fo r the other subj ective disturbances, namely, sleep disorders, fatigue, and confusion .
The da y workers, whose expos ure levels had been higher th an those o f th e three-shift wo rkers, su ffe red more frequently from un steadiness o f their hands th an thei r refe rents did. Tremor was well kn own as an ea rly mercury effect amon g the chlo rine-alkali wo rke rs becau se it had been measured with the aid of drawings mad e by a n un supported hand during the periodic health examinat ion s performed by the occupation al health physi cia ns. In that way the exposed wor kers were more familiar with the motor tasks, which , to gethe r with the lower motivation of the referents on the performance test s, co uld perhaps explain the significantly poorer per form an ce of the referen ce su bjects in the hand-eye coordi natio n test. The methods used for measuring hand steadiness in our study did not reveal individ ual parameters o f tremor pattern , eg , velocit y or th e effect of loading on the tremor spectra that were distu rbed initially (5,7,10) . On the other hand , fo rearm tremor has been show n to be reversible (12), and the exposed workers had had unstead ines s in their hands only now and then .
The exposed subgroup with the higher urinary mercury level had a h igh er activity of seru m gamma gluta my l transferase and a la rger mea n volu me of erythrocytes than their referents, a nd th is high er va lue might have be en ca used by th e co ns um ptio n of a lcoh ol. Ea rlier , a high excretio n of urina ry gamma gluta my l tr an sferase, which is richly found in the kidne y, has also been co nnected with a high level o f expo sure to inorganic mercury (I , 23). However , in o ur st udy, th e subgroup with a higher expos ure level to mer cury, as indicated by the urinary mer cury level, had results similar to those of their matched referents on the per formance tests.
In thi s study no objecti ve distu rbances in perceptual motor, memory, or learning a bilities were co nnected with lo w lon g-t erm exposur e to elem ent al me rcury vapor, which was es tima ted (24) to have been ab out on the health-based level for the average long-term mer cury vapor concentration in air , ie, 25 ug/m", recommended by the WHO Study Grou p (4). Although the exposed workers had subjective symptom s which have earli er been described in connection with micromercurialism, one can conclude that no significant adverse effects were produced by the long-term exposure to mercury vapor at an average mercury concentration of 25 ug/rn ' of air.