Introduction

More than thirty years of research assessing multiple domains and subdomains of the nervous system have confirmed the relationship between neuropsychological disorders and exposure to organic solvents (OS). These disorders were first studied in order to establish the diagnostic criteria for chronic solvent-induced encephalopathy (CSE).^1,2Diminished attention, memory deficits, motor disabilities and affection disorders were identified as the most common characteristics of CSE.²

High-income countries have reported a reduction in the number of CSE cases due to the elimination of OS-related industrial processes.^3,4A review found only weak and inconsistent associations between long-term human exposure to hydrocarbon solvents at or below occupational exposure limits and adverse neurobehavioral effects.⁵

Many industrial processes depending on OS have been transferred to low and middle-income countries with less stringent occupational health standards where there is little concern about OS-related diseases and health data is scarce. Since the early 1990s, only a handful of studies regarding OS exposure and neurotoxic effects in low and middle-income countries have been published, i.e. about a petrochemical complex in Brazil,⁶and more recently rubber workers in Iran⁷and printing workers in Hong Kong.⁸

In Mexico, trade agreements and economic globalization have contributed to increase the imports and production of OS from less than 1 million tons in 1987 to more than 66 million tons in 2013, and consequently the number of workers exposed to these chemical products has increased exponentially.⁹There is hardly anything known about solvent-induced neurotoxic effects in Mexico, except for a study that showed ototoxicity in workers from a paint factory.¹⁰

Therefore, the aim of this study was to assess the neuropsychological performance of workers from a paint factory in Mexico City, who were exposed to a mixture of organic solvents, mainly toluene and xylene.

Materials and methods

This research project was registered and approved by the Ethics Committee of the High-Specialty Medical Unit from theCentro Médico Nacional Siglo XXIof the Mexican Institute of Social Security (Instituto Mexicano del Seguro Social, IMSS, in Spanish). The paint company agreed to grant workers a Monday as a day off, so they could attend the Occupational Health Research Unit for the examination. Transport was provided by the researchers and all participants signed an informed consent.

In this cross-sectional study we applied a neurobehavioral test battery to all workers from the production departments of a paint factory between June 21st 2004 and September 26^th2005. The participation rate in the study was 100%, but we excluded the only two female workers that attended the study, because of the low number of female participants, and three workers that presented alcohol breath on their test day. The final study population comprised 208 male workers. We used a validated questionnaire by means of a personal interview to collect sociodemographic data and occupational history, lifestyle behaviors (smoking, alcohol intake and recreational drug use), and neurological health (history of head trauma, neurological illness, psychiatric disorders and use of neuroleptic drugs).

Considering the effects of the different OS compounds which can have different clinical manifestations on both the central and peripheral nervous systems, we used a wide-ranging test battery, including tests from the World Health Organization’s Neurobehavioral Core Test Battery (WHO-NCTB),¹¹complemented with additional tests previously used in Latin America¹²and four questionnaires exploring neuropsychiatric symptoms. Before the start of data collection, we trained three psychologists to apply neurobehavioral tests according to a standardized protocol and two nurses to obtain the occupational history and apply the symptom questionnaires.

Workers were tested on the same day (Monday), after 48 hours without exposure to OS. Workers were asked not to drink alcoholic and caffeinated beverages 24 hours prior to the study and to rest properly. Moreover, examiners were blinded to participants’ exposure status.

The tests assessed immediate auditory memory (digit span)-correctly repeated digits (score up to 28), associative memory (digit symbol)-symbols correctly drawn in 90 seconds (score up to 93), sustained attention (digit vigilance)-numbers correctly identified in one time, verbal fluency-number of words mentioned in 60 seconds, and visual-motor perception (trail making A)-time needed to complete the test. For immediate visual memory, Benton’s test was applied; besides, the number of figures correctly selected (up to 10), and motor coordination hand-skill functions were also measured.¹²The WHO-NCTB measured the number of rotated pegboards. Moreover, Pursuit aiming II evaluated the motor speed, precision and stability functions.¹²Motor speed and coordination of finger tapping and hand strength of grip measured in kilograms were assessed. For affective evaluation, the Q-16 questionnaire, Brief Symptom Inventory (BSI),¹²Seppo Aro Questionnaire and Yoshitake Questionnaire were applied. Moreover, for statistical analysis the units of original measurement were used, either the time or number of responses, as corresponding to each test.

We tested between three and four workers per day, from 8:00 am to 2:00 pm, with breaks between tests. The average test duration for individuals was 120 minutes. A comfortable setting was provided, as well as the necessary equipment and sufficient material for adequate development of the tests. All workers received the same information and motivation.¹²

During 2006 and 2007 we measured workers’ exposure to 14 of the most used solvents in paint manufacture in the different departments of the factory. A subset of workers (n=134) carried Gillian gravimetric low-flux pumps at 1 to 500 milliliters/minute (±5%) intervals, during an 8 hour work day, with 4 or 5 continued collected samples for each worker. Adsorbent activated carbon tubes were sent to the IMSS Occupational Health Laboratory in Veracruz, Mexico, for gas chromatography analysis according to international standards. The results were reported in milligrams per cubic meter (mg/m³). The average concentrations for specific jobs and specific work areas were assigned to the remaining 74 workers who did not participate in personal measurements according to job category, work shift and years worked in the corresponding job category.

For every work shift we had one measurement of exposure which was used to assign workers’ exposure levels. Then, a cumulative toluene exposure dose was estimated according to working years at the company (time-weighted average [TWA] × years working at the company). A low exposure group (group 1) was created with the first three quartiles and a high exposure group (group 2) with the one above the 75th percentile.

Since the company manufactured paints with lead oxides and some workers used leaded-glazed pottery, and as lead in blood has been associated with neuropsychological disorders and psychiatric symptoms,¹³we measured lead with an anodic voltameter (LeadCare) in capillary blood (micrograms per deciliter [μg/dl]) in a sample of workers.¹³

Statistical analysis was done with Stata v.10, which included descriptive analysis, mean differences, analysis of variance, and between proportions. Personal, sociodemographic and neuropsychiatric characteristics were considered as confounding variables (age, smoking, alcohol intake, drug use, neurologic diseases, level of education and lead exposure). Multiple linear regression models were constructed, considering 40 years of age as a cut-off point to contrast the two groups, given that cognitive decline was observed with age. A value ofp<0.05 was considered as significant.

Results

The study included 208 workers from different production departments who completed the tests. Significant differences between categories were observed for mean values of age, education level, and time working at the company; for drug use, alcohol intake, head trauma and use of leaded pottery there were no differences (table I). Considering this, we evaluated the performance in each neurobehavioral test and neuropsychiatric symptom questionnaire by each work area and we did not find significant differences. We also analyzed job performance in subgroups with less or more than 10 years of employment, considering the cumulative dose and contrasted group one and two. The groups were similar in age, seniority at work and scholar achievements.

Toluene and xylene were present in all working areas, however the mean concentration in 80% of the working places was higher for toluene. Considering this and knowing that toluene (as well as xylene) is associated with neurotoxic effects,¹⁴we calculated the exposure related only to this substance. The 14 OS TWA concentrations were lower than the allowed values, but the highest toluene and benzene concentration surpassed this in 9 and 3 departments, respectively.

OS levels were lower than those permitted by the Mexican Official Norms (mean time weighted concentration-TWA) and United States Norms (TLV-TWA), which were very similar (table II). The average toluene weighted concentration in group one and group two was 5.65 and 21.14 mg/m³, respectively. The cumulative index for toluene in the low exposure group was significantly lower than in the higher one (p<0.05). The means of 57% of the other 13 substances were higher in group 2, including xylene, as well as the means of 31% (table II).

Mean blood lead levels were not significantly different among users of leaded-glazed pottery (8.23 ± 6.2 μg/dl) and non-users (6.6 ± 3.7 μg/dl), and neither between the high (8.7 ± 3.8 μg/dl) and the low OS exposure groups (7.3 ± 5.9 μg/dl).

Regarding neuropsychological tests’ results, we observed that 72% of the workers with high exposure to OS mixture had low performance, but the differences were only significant at 55%. The lowest performance was present in the following cognitive domains: associative memory (Digits and symbols), visual-motor perception (Trail making A), sustained attention (Digit vigilance), as well as in the cognitive motor-domain and motor coordination (Purdue pegboard, Grooved pegboard, Santa Ana and Pursuit aiming II). No differences in motor domain and neuropsychiatric symptoms tests were observed (tables III and IV).

Multiple linear regression models adjusted for age, education level and leaded-glazed pottery use showed that the higher exposure group had a lower performance in most of the neuropsychological tests; motor perception β=5.5 (p=0.04), motor speed β=-15.7 (p=0.02) and precision β=-16.4 (p=0.007) were statistically significant. Residuals analysis was performed for each model to corroborate that they fulfilled generalized linear models assumptions (table V). Other confounders were tested such as drug use, smoking, alcohol intake, traumatic brain injury, but they did not explain lower performance on any test-these confounders were categorized as present/absent and since they did not modified the model significantly, they were not included.

Discussion

Drawn from the results of the present work, it can be suggested that working populations with long-lasting high OS exposure time could have lower performance in the usual processing of information, i.e. the cognitive domain. A decrease in visual-motor perception was identified with Trail making test A, along with cognitive-motor alterations, where hand skills with speed and precision (Pursuit aiming II) showed a worse performance compared with the low exposure group. Age also contributed to diminish these functions, whereas higher education level increased them as expected, and this provides certainty that the observed effects are true. Although the remaining multiple linear regression models did not show significant differences between the groups, the group with the highest accumulated exposure consistently performed worse (table V).

The diversity of processes where OS were used poses a challenge and limits extrapolation of the results. Therefore, we considered comparing our results with studies carried out in factories which produce solvent-based paints or industrial painters by the similarity of mixtures to which workers are exposed: toluene, xylene, n-hexane, benzene and acetone.

Paint industry has been transformed since the end of the 1960’s and has gradually decreased exposure levels to OS mixtures.^4,15Despite this decline, Orbaek and colleagues observed a decrease in memory, difficulties in concentration, fatigue, as well as feeling of internal tension and hostility in an exposed group, pointing out the presence of a neurasthenic syndrome.⁴Similar results were found by Elofsson and colleagues in which painters also had significant differences in performance of speed of perception, manual skills and memory, among other tests.¹⁶Even so, we did not find any differences in neuropsychological symptoms (table IV). In addition, Jung-Der and Jong-Dar identified more acute and chronic central nervous system symptoms in the category of higher OS exposure among workers in paint factories from Taiwan.¹⁷

Thus, decrease in the performance of only three tests (14%) of 21, including emotional disorders in our study, is slightly lower than the proportion of lower performance tests found by Son-Yenet al-three (21%) out of 14 in Taiwan.⁸Also in that study, the exposure determined lower performance in continuous pattern matching and memory performance,⁸whereas in our study these changes were related to the lower cognitive performance across domains (Trail making) and cognitive-motor (Pursuit aiming II) related to greater exposure. Similar to our results, Colvin and colleagues, and Kishi and colleagues studied industrial painters exposed to similar mixtures at moderate concentrations; a decrease in the performance in various neuropsychological tests was observed as well as poor performance on vocabulary, block design, digit span, and Santa Ana motor coordination test, along with two subcategories of the profile of mood status. Moreover, multiple linear regression models adjusted for age, education level and alcohol intake showed an association with OS mixture exposure, along with motor, memory and attention tests which are similar to ours in the cognitive-motor Pursuit aiming test.^18,19

Our results were very similar to those in which cognitive deficiencies were evaluated, particularly regarding lesser cognitive performance and decreased immediate memory.^20,21It is likely that concentrations below the allowed limits produce slight neuropsychological alterations in the cognitive sphere with no neuropsychiatric alterations.^22,23,24About these supposition, it is necessary to conduct further research in populations with low OS exposure. Perhaps most of our observations were due to a greater OS mixture exposure in the recent past at U.S. capital enterprises, as flexibility in the compliance with standard limits allowed in Mexico was common. Besides we have to mention that differences between studies, despite similar OS mixtures exposures, are probably due to different OS concentrations, exposure estimation methods, and misclassification of exposure, as well as healthy worker bias and use of different neuropsychological tests.

One limitation of the study is the lack of OS exposure historical data from the company, which makes it difficult to quantify intensity and cumulative exposure time more precisely for each OS and different mixtures. Although there are biomarkers for organic solvents exposure such as phenyl mercapturic acid (benzene), mercapturic acid (toluene) and methyl hippuric acid (xylene), their unique measurement could mislead to an imprecise exposure assessment. In Mexico, no single or repeated measurements of organic solvents in the work environment or their biomarkers are made during the working life of workers.

Moreover, exposure underestimation is possible, as we had no access to previous exposure data of workers with longer seniority, therefore we assumed that the measured OS exposure behaved similarly all the time. Furthermore, we consider that in our study healthy worker bias could be present because three years before, personnel who had been working at the company for the longest time were fired. We also have to mention that the population was young, with mean age of 32 years, which probably contributed to underestimating the observed effect.