Occupational and non-occupational risk factors correlating with the severity of clinical manifestations of carpal tunnel syndrome and related work disability among workers who work with a computer

The study included 190 workers aged 20–65 years (100 men, 90 women) who were previously diagnosed with CTS and who worked with computers at least 1 h per day. In February 2023, these participants took their regular occupational medicine check-up at the Primary Health Care Institution Besa Meditor, Oslomej, Kičevo, North Macedonia. They came from different business sectors (from local administration to industry clerks in the thermal power and solar power plants, and automobile industry). The check-up did not include CTS diagnosis, as it had been established earlier by clinical specialists according to clinical guidelines (40, 41).

All participants signed an informed consent to participate, and the study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University of Niš Faculty of Medicine, Niš, Serbia (Decision No. 12-1285/2-1 of 2 February 2023).

To make our research as comprehensive as possible, we included a wide variety of participants information collected through a specifically designed questionnaire and medical documentation: sociodemographic data, medical history of CTS and other diseases, Boston Carpal Tunnel Questionnaire (BCTQ) score, electrodiagnostic (EDX) test results, work disability data, occupational factors, diet, physical activity, anthropometric data, and laboratory results.

All participants completed the BCTQ (42) version translated into Macedonian, which consists of the Symptom Severity Scale (SSS) and the Functional Status Scale (FSS). The first consists of 11 items through which participants rate the severity of pain, paraesthesia, numbness, weakness, nocturnal symptoms, and difficulty grasping objects. The second consists of eight items through which participants rate hand function difficulties in writing, buttoning, holding a book while reading, gripping a telephone handle, opening jars, performing household chores, carrying grocery bags, bathing/taking shower, and dressing (42, 43). Each item is scored one (no symptoms/difficulties) to five (worst symptoms/cannot perform the activity at all) and the total given as mean score for each scale. Higher scores indicate severe symptoms (SSS) or impaired function (FSS) (43).

Electrophysiological testing was run on the Medelec-Synergy EMG instrument (Oxford Instruments, Surrey, UK) by a single specialist of physical medicine. Standardised measurements included motor and fibre conduction velocities and F-wave latencies of the median and ulnar nerves as described elsewhere (44).

Briefly, motor nerve conduction velocity was measured using the orthodromic technique by placing percutaneous detection electrodes on the abductor pollicis brevis (for n. medianus) and on the abductor digiti minimi (for n. ulnaris) and stimulating the motor points of the nerves with supramaximal intensity on the wrist, 8 cm proximal to the detection electrodes.

Sensory nerve conduction velocity was measured using the antidromic technique by stimulating the wrist nerves with electrodes and taking readings with ring electrodes placed on the index finger (for n. medianus) or little finger (for n. ulnaris), 14 cm distal to the stimulating electrode.

To control for confounding factors of age, temperature, illness (i.e., diabetes), gender, and hand size, sensory distal onset latency was measured on the ring finger by stimulating both the median and ulnar nerve at the level of the wrist, 14 cm proximal to the detection ring electrode. In case both hands were affected by CTS, we show the measurements on the more affected hand.

Data evidencing temporary work disability related to CTS and its duration (in days) over the last year were obtained from participants’ GP medical records, while data on recommendations for a job change and applications for permanent work disability related to CTS were obtained from participants and related medical documentation during their occupational-medicine check-ups.

Our assessment was divided between two large groups of factors: those associated and those not associated with working on a computer. Details are available in the Results section. Most data were self-reported and collected through a specifically designed questionnaire but microclimatic factors and vibrations at the workplace were measured by specialists of occupational medicine and safety at work as described elsewhere (45, 46). General job satisfaction was self-rated by the participants as “poorly satisfied”, “moderately satisfied”, or “very satisfied”.

Participants’ body height and weight were measured by trained medical nurses at 8 am on empty stomach and bladder, wearing only underwear. Body mass index (BMI) was calculated, and the participants classified in three groups: normal weight (BMI: 18.5–24.9 kg/m²), overweight (BMI: 25–29.9 kg/m²), and obese (BMI: >30 kg/m²) (47).

Arterial blood pressure was measured on the left upper arm after sitting for at least 10 min with a calibrated mechanic sphygmomanometer and stethoscope (Becton Dickinson, Franklin Lakes, NJ, United States). Hypertension was determined in participants with systolic pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg (48).

Data on the frequency of consumption of certain food groups and dietary supplements over the past year and level of physical activity over the past week were collected by a trained medical doctor using a food propensity questionnaire (FPQ) in accordance with the European Food Safety Authority (EFSA) EU Menu methodology (49, 50) and the International Physical Activity Questionnaire – Short Form (IPAQ-SF) (51). Data on the frequency of consumption of certain food groups and dietary supplements over the past year were categorised into seven categories: 1) never; 2) less than once a month; 3) one to three times a month; 4) once a week; 5) two to three times a week; 6) four to five times a week; and 7) six to seven times a week. Metabolic equivalents (METs) (expressed in minutes per week) were calculated as described in detail by Craig et al. (51).

Fasting blood samples were collected between 7 and 8 am. Erythrocyte sedimentation rates (ESR) were analysed with the Westergren's method at 18 °C and the results show one-hour values. Fibrinogen was measured with the Clauss fibrinogen assay (52). Standard biochemical assays were run on an ARCHITECT c8000 Abbott clinical chemistry analyser (Abbott Laboratories, Abbott Park, IL, USA) using commercially available kits (Abbott Laboratories) and included serum glucose, total cholesterol, high-density lipoproteins (HDL)-cholesterol and low-density lipoproteins (LDL)-cholesterol, triglycerides, urea, creatinine, aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), uric acid, and C-reactive protein (CRP). Intra- and inter-assay coefficients of variation (CVs) for all measurements were <6 %.

The required minimum sample size for Spearman correlation analyses with meaningful correlation coefficients and moderate effect (r_s≥0.300) was 85 participants, and for the correlation analyses with correlation coefficients r_s≥0.600 it was 20 participants, when the level of significance was set to 5 % (α=0.05) and the study power was set to 80 % (β=0.20) (53). However, since we had multiple testing, we had to increase the number of the subjects included (54).

The normality of continuous data distribution was tested with the Kolmogorov-Smirnov test. Since no continuous data had normal distribution, nominal/categorical data are presented either as a number and percent, while all other data as median and interquartile range.

Correlations between different occupational and non-occupational factors with the severity of subjective symptoms (SSS and FSS scales), objective impairments (EDX results), and temporary and permanent work disability indicators were assessed using the Spearman's r_s correlation coefficient (to assess correlations between two non-parametrically distributed numeric or ordinal variables), Glass rank-biserial r_rb correlation coefficient (to assess correlations between one non-parametrically distributed numeric or ordinal variable and one dichotomous categorical variable) (55, 56, 57), and the phi φ or Cramér's V φ_c correlation coefficients (to assess correlations between two categorical variables).

Two categories (groups) were compared using the Mann-Whitney U and chi-squared test.

For all analyses, statistical significance was assumed at a two-tailed p<0.05, but also the Bonferroni correction for multiple testing was applied (58). All analyses were run on the SPSS 22.0 software (SPSS Inc., Chicago, IL, United States).

Detailed demographic, anthropometric, overall health, and lifestyle characteristics of the study participants are given in Table 1. Interestingly, most are overweight or obese, have quite an unhealthy lifestyle (including smoking, high alcohol consumption, physical inactivity), and suffer from hypertension and endocrine glands disorders (diabetes or thyroid disease).

Table 2 shows participants’ medical history related to CTS, including subjective (BCSQ) and objective (EDX) impairments and work disability.

All the participants met the criteria for the EDX diagnosis of CTS in active workers or general population (44), while data from BCSQ indicated a moderate handicap (59, 60). All the participants also reported temporary work disability related to CTS during the preceding year, with the median duration of about one month (ranging from 2 to 85 days), and 62 applied for permanent (long-term) work disability pension because of CTS (on personal request, request of the employer, or request of medical commission due to prolonged work disability), but only 10 were approved preterm pension after this study was completed. Still, about 40 % received a job change recommendation because of CTS after medical checkup by a specialist of occupational medicine. The 62 participants who had applied for permanent work disability were older (median age: 55 years), with longer duration of disease (median: 13.1 years), worked longer with computer during their career (median: 15.4 years), spent more hours on work with computer (median: 6.2 h a day), and were on longer sick leaves in the past year (median; 55.5 days) compared to those who had never applied for disability pension (Mann-Whitney U test, p<0.001 for all).

Table 3 shows specific data about occupational factors, both related and unrelated to computer work. Taken as a whole, the participants spent 4.4 years (ranging from 1 month to 20 years) working on a computer for 4 h a day at work (ranging from 1 h to 6.8 h) and 1 h a day in their free time (ranging from 5 min to 3 h). Although the overall median time spent on a computer is moderate, the wide variety of ranges allowed us to analyse correlations between the duration of computer work and severity of CTS impairments.

An interesting finding is that the total duration of work with computer (in months, hours, or both) negatively correlates with the use of ergonomically designed devices, hand exercises, and training for safe and healthy work with computers reported by 40 % of the participants (p<0.001 for all, data not shown), indicating that new workers and those spending less time on computers had better training and practiced ergonomic measures better.

Heavy physical exertion with arms/hands at work was not performed by many participants, but other repetitive jobs with hands in addition to computer-work were reported by about half of them. About half the participants also reported occupational exposure to vibrations lasting between 10 min and 40 min, but only five participants were exposed to vibrations above the regulatory exposure limit of 5 m/s² in average for eight-hour work, while about 30.0 % were also exposed to vibrations outside of work. Curiously enough, the total duration of work with computer highly correlates with the duration of exposure to vibrations, heavy physical exertion, other repetitive movements, and negative microclimatic conditions (p<0.001 for all, data not shown).

For the sake of easy viewing, Table 4 shows food and dietary supplement consumption in three categories of frequency, although originally there were seven (see the Methods section). The vast majority of the participants reported quite unhealthy eating habits, such as frequent consumption of refined grain products, red and processed meat, tallow, sweets and dried fruits, alcoholic beverages (spirits in particular), and carbonated sugar-added beverages. In contrast, the reported consumption of dairy products, eggs, poultry meat, fish and sea food, integral grains, fresh vegetables and fruits, nuts, vegetable oils (including sunflower oil and olive oil), and pork lard was surprisingly low. Between 55 % and 65 % of the participants reported not to consume any dietary supplements, and those who consumed them show no particular preference.

Our correlation analysis shows that all dietary factors (except for dairy products, legumes, and wine) and consumption of supplements highly inter-correlate and tend to cluster in two opposite eating patterns: a) the predominant “unhealthy” one with high consumption of refined grains, red and processed meat, tallow, sweets, dried fruits, spirits, and carbonated beverages (“dietary cluster 1”) and very low consumption of integral grains, eggs, poultry, fish/sea products, fresh vegetables and fruits, fruit-juices, nuts, vegetable oils, pork lard, and beer (“dietary cluster 2”) and supplements and b) the reverse “more healthy” pattern, which was much rarer. The food groups from one cluster highly positively correlate within the same cluster and highly negatively correlate with the food groups from the other cluster (p<0.001 for all, data not shown).

The laboratory data seem to reflect the anthropometric measurements and reports of dietary and physical (in)activity habits, as most participants showed values indicative of metabolic syndrome and inflammation: low HDL- and high LDL-cholesterol, triglycerides, glucose, ALT, AST, GGT, uric acid, urea, creatinine, fibrinogen, CRP, and sedimentation (Table 5). In fact, we found significant correlations between all laboratory indicators of metabolic risk and inflammation (except AST) and anthropometric and dietary data, smoking, alcohol use, physical inactivity, and hypertension (p<0.001 for all, data not shown).

Tables 6–10 show the findings of our correlation analysis between occupational and non-occupational factors and the severity of CTS-related clinical manifestations and work disability. Since many variables were included in the study, we applied the Bonferroni correction. Significant correlations (p<0.00003472) obtained this way are underlined in the tables.

The SSS, FSS, and EDX scores highly correlate with temporary/permanent work disability and certain clinical characteristics of CTS, including duration, bilateral presentation, family history, treatment, as well as with female gender, age, smoking, alcohol consumption, physical inactivity, BMI, systolic and diastolic blood pressure, and diseases of endocrine glands (e.g., diabetes, thyroid disease). EDX findings of the ulnar nerve on the ring finger (but not on the little finger and hypothenar) show the same correlations but to a lesser extent (Table 6). In addition, BMI highly correlates with blood pressure, physical inactivity, smoking, alcohol consumption, overall health status, and endocrine glands disorders (p<0.001 for all, data not shown).

Table 7 shows a highly significant correlation between hours/months of work with computer and the severity of clinical impairments and temporary or permanent work disability. In addition, time spent daily at computer outside the workplace also highly correlates with these indicators. In contrast, the number and total duration of additional breaks, use of an ergonomically designed devices during work on the computer, hand exercises during breaks at work and outside work, and having completed the training for safe and healthy work with computers all highly negatively correlate with BCTQ and EDX findings and work disability. Other occupational factors unrelated to computer work that highly correlate with the severity of impairments and work disability include other repetitive jobs with hands, exposure to vibrations and its duration, work at low temperature, high air humidity, and draft, and low job satisfaction. As for the ulnar nerve on the ring finger EDX findings, they to correlate with the above factors but to a lesser extent.

Since many participants were exposed to combined occupational risks significantly correlating with CTS (vibrations, other repetitive jobs with hands not related to work with computer, work under decreased air temperature, increased air relative humidity, and increased air-flow speed), we repeated the correlation analysis in a subgroup of workers unexposed to additional occupational risks i.e., in those who worked under normal microclimatic conditions, without vibrations, heavy physical exertion, or other repetitive jobs with hands (N=83) only to find even more significant correlations between some occupational factors related to computer work and the severity of objective clinical manifestations (EDX results) or temporary work disability indicators but a bit less significant correlation with the severity of subjective clinical manifestations (BCTQ results), since their BCTQ scores are generally lower (the median score is 1 on both BCTQ scales) (data not shown). More specifically, EDX results on the median nerve and temporary work disability (with rs correlations coefficients ranging from 0.839 to 0.975; p<0.001) significantly correlate with the duration of work on a computer and use of additional breaks, which has remained significant even after the Bonferroni correction. However, we have found no correlation with permanent work disability indicators.

Likewise, the subgroup of the remaining workers (N=107), those with additional occupational risk-factors, shows a significant correlation between the duration of computer work with objective CTS manifestations and work disability (r_s ranging from 0.950 to 0.973; p<0.001) or subjective (BCTQ) scores (r_s ranging from 0.840 to 0.880; p<0.001) after the Bonferroni correction.

Furthermore, workers without additional occupational risk factors (N=83) scored much lower on BCTQ and EDX than those with additional risks (N=107) and have a lower rate of temporary and permanent work disability (Mann-Whitney U test and chi-squared test, p<0.001 for all, data not shown), which points to an additive (or even synergistic) effect of occupational risk factors, whether related to computer work or not.

Table 8 clearly shows that foods clustered in “dietary cluster 1” positively correlate with the level of impairments and work disability, whereas foods clustered in “dietary cluster 2” and supplementation correlate negatively. The same goes for the correlation of dietary clusters and supplementation with the BMI: “dietary cluster 1” shows a strong positive while “dietary cluster 2” and supplementation a strong negative correlation (p<0.001 for all, data not shown).

As for laboratory findings (Table 9), all inflammation and metabolic impairment markers but AST highly correlate with measures of CTS severity and work disability. As expected, they also highly correlate with BMI, save for AST (p <0.001 for all, data not shown).