Relationship between repetitive work and the prevalence of carpal tunnel syndrome in part-time and full-time female supermarket cashiers: a quasi-experimental study
- First Online:
- Cite this article as:
- Bonfiglioli, R., Mattioli, S., Fiorentini, C. et al. Int Arch Occup Environ Health (2007) 80: 248. doi:10.1007/s00420-006-0129-0
- 357 Views
To investigate the prevalence of Carpal Tunnel Syndrome (CTS) in full-time and part-time supermarket cashiers exposed to a different weekly duration of biomechanical load.
All the 269 cashiers and 127 office workers were asked to participate. The protocol included ergonomic risk assessment, collection of personal and clinical data and bilateral electrodiagnostic study of the median nerve. CTS symptoms were defined as past and/or current nocturnal and/or diurnal numbness, tingling, burning or pain involving at least one of the first three fingers. Results were evaluated according to two case definitions based on current symptoms and on the combination of current symptoms and slowing of sensory conduction velocity from wrist to palm, respectively. Difference in proportions of CTS symptoms and cases was evaluated by the Pearson’s chi-square (χ2) test, univariate and multivariate logistic regression analyses were performed to determine the impact of weekly exposure.
The final female study population included 71 full-time cashiers, 155 part-time cashiers and 98 office workers. Ergonomic risk level was rated 5 for hand activity level and 4 for peak of force according to ACGIH. The intersection of the two values fell on the threshold limit value line, confirming the possible exposure to biomechanical risk factors for CTS. The prevalence of current CTS symptoms was higher among full-time (31.0%) than in part-time cashiers (19.3%) or controls (16.3%) (p = 0.055). A similar pattern was found for CTS past symptoms and cases. Univariate analysis showed that full-time cashiers had a 2.3 fold increased risk for CTS specific current symptoms than control subjects. A similar trend also emerged for CTS cases but was not significant (Odds ratios 1.23). Multivariate logistic regression analysis confirmed the increased risk for CTS current symptoms in full-time cashiers.
Intensive manual work associated with inadequate recovery time might have generated an impairment of the median nerve at the wrist level proportionally increasing with duration of hand use. Our study can provide useful information both for ergonomic risk assessment and work organization.
KeywordsCarpal tunnel syndromeSupermarket cashiersUpper limb repetitive movements
The relationship between occupational exposure to biomechanical risk factors (repetitiveness, force, posture) and upper limb musculoskeletal disorders is supported by strong scientific evidence (Hagberg et al. 1995; Bernard 1997; National Research Council and the Institute of Medicine 2001). However, although several risk assessment methods have been reported in literature, only few studies have analyzed the dose–response relationship at different exposure levels. This is mainly due to the lack of standard criteria for the classification of musculoskeletal disorders, generally defined by different combinations of symptoms, physical findings and results of diagnostic procedures. Only the uniformity of case definitions would in fact allow a wide comparison of results across epidemiological studies and a better understanding of the dose–response relationship (Fallentin et al. 2001; Fallentin 2003).
Carpal Tunnel Syndrome (CTS) is an entrapment neuropathy of the median nerve at the wrist level. It is common in the general population and has been associated with jobs characterized by repetitiveness, force and awkward postures (Bernard 1997; Nordstrom et al. 1997; Tanaka et al. 1997; Giersiepen et al. 2000). In a consensus report, Rempel et al. (1998) concluded that a combination of electrodiagnostic findings and symptom characteristics provides the most accurate CTS diagnosis. They also recommended a symptoms characterization based on quality and location.
Supermarket cashiers are known to be at risk for CTS, being their job characterized by repetitive movements and forceful exertions of the upper limb. Since the end of the eighties, many epidemiological studies have investigated risk levels for cumulative trauma disorders and CTS in supermarket cashiers (showing 19–62% prevalence of symptoms of CTS), but few have examined the relationship between prevalence of CTS and intensity and/or duration of biomechanical exposure (Margolis and Kraus 1987; Osorio et al. 1994).
Italian supermarkets usually employ both part-time and full-time cashiers, thus providing the opportunity, without modifying the personnel schedule, to investigate workers exposed to the same levels of repetitiveness and hand force but to a different weekly duration of exposure. In order to analyze the dose–response relationship, we compared the prevalence of CTS in the two groups, using administrative personnel as control group.
All the 269 cashiers and 127 office workers of four big supermarket stores were informed about the study and participants signed an informed consent. The study protocol included risk assessment, clinical evaluation and electrodiagnostic study of the median nerve. Clinical and electrodiagnostic examinations were performed by different operators, blinded to each other’s results.
The ergonomic risk assessment was carried out by trained observers using scales of hand activity level (HAL) and normalized peak of force (PF) proposed by ACGIH. The method is intended for “mono-task” jobs performed for 4 or more hours per day, and allows to identify combinations of PF and HAL associated with significantly increased prevalence of musculoskeletal disorders [i.e. the combinations falling within the area over the threshold limit value (TLV) line] (ACGIH Worldwide 2002).
After observing several operators, three cashiers with different anthropometrical profiles were chosen and videotaped. Job analysis was performed at different times of the day, corresponding to low (08.30–09.30 am), medium (13.00–15.30 am) and high (10.00–12.30 pm and 04.30–07.30 pm) customer flow.
The clinical evaluation of each worker was performed by a properly trained clinician by means of a structured questionnaire about personal data, work history and work schedule, lifestyle (physical activity, food choices, cigarette smoking, alcohol and medicines consumption, hobbies) and clinical history. The questionnaire included a section about CTS, containing questions about past and current symptoms, previously diagnosed and medically/surgically treated CTS. Current symptoms were defined as presence of nocturnal and/or diurnal numbness, tingling, burning, or pain involving at least one of the first three fingers within the last month (classic/probable or possible symptoms for CTS) (Rempel et al. 1998). Each worker reporting the presence of symptoms was asked to locate their exact distribution by blackening the affected area on a hand diagram (Katz and Stirrat 1990; Katz et al. 1990).
Median motor and sensory nerve conduction studies (NCS) were performed bilaterally, according to the segmental palmar technique described by Kimura (1978). Techniques which evaluate median sensory nerve conduction over a short distance across the carpal tunnel have demonstrated a higher sensitivity in the diagnosis of CTS (Jablecki et al. 2002a, b). All NCS were performed in a warm room (22–25°C); since NCS are affected by body temperature, both hands were warmed by immersion in warm water until they reached a palmar temperature of at least 32°C (Denys 1995). Values of sensory conduction velocity from wrist to palm (SCV-WP) falling below the lower 99% confidence limit (43.8 m/s) of the electrodiagnostic reference values described by Kimura (1979) were considered “abnormal.”
Results of clinical and instrumental examination were evaluated according to two different case definitions. The first was based on typical current symptoms (classic/probable or possible), whereas the second case definition of CTS was based on the combination of typical current symptoms and electrodiagnostic findings (reduction of SCV-WP), according to the Consensus Criteria for the Classification of CTS published by Rempel et al. (1998).
Data were analyzed using STATA 8 software (Stata Corporation, College Station, TX, USA). The difference in proportions of CTS symptoms and cases was evaluated by the Pearson’s χ2-test. In order to quantify the risk of CTS in relation to both case definitions, logistic regression analysis was performed. Odds ratios (OR) of the full-time and part-time group were used to determine the impact of weekly exposure, adjusting for possible or probable confounders. Interaction between levels of weekly exposure and high Body Mass Index (BMI) was tested by means of the likelihood-ratio test. A p-value of less than 0.05 was regarded as significant.
The overall response rate was 88.8% for cashiers (239 subjects) and 87.4% for office workers (111 subjects). Twelve male workers, seven cashiers and five office workers, were excluded from data analysis, as they represented 2.9 and 4.5% of the participants, respectively. Seven subjects who had undergone surgery for CTS and seven office workers who had been previously employed as cashiers for more than 5 years were also excluded. The final female study population included 71 full-time cashiers, 155 part-time cashiers and 98 office workers (mean ages 38.7 ± 7.1, 37.6 ± 6.8, and 38.3 ± 6.5 year, respectively).
The full-time and part-time group followed a 6 and 3 days/week work schedule, respectively, with a daily work shift of 6 h 10 min. All the cashiers performed no other duty than the checkstand operations.
The production standard was calculated on the basis of the sold articles/running workstations ratio per considered time interval. The data collected refer to the entire working week, including higher and lower flow days. The mean number of handled items was 300/h, reaching peaks of 600/h within the Saturday 12–13 pm interval. Supermarkets articles are extremely variable in size and weight and the scanning operations vary accordingly. Objects weighting less than 3 kg are normally lifted to find the barcode, while heavier objects are grasped, trailed and pushed with both hands and the barcode is read by means of a hand laser scanner.
The biomechanical risk level was rated 5 for HAL and 4 for PF. The intersection of these two values fell on the TLV line, confirming the possible exposure to biomechanical risk factors for CTS.
The prevalence of past CTS symptoms was higher among full-time (53.5%, i.e., n = 38) than in part-time cashiers (35.5%, i.e., n = 55) and controls (33.5%, i.e., n = 33) (p = 0.016 at Pearson χ2-test). A similar pattern was found for CTS current symptoms (i.e. present in the last month and confirmed by the blackening on the hand diagram): the reported prevalences were 31.0% (n = 22), 19.3% (n = 30), and 16.3% (n = 16), respectively (p = 0.055 at Pearson χ2-test). The prevalences of reported cases of CTS were 9.9% (n = 7) among full-time cashiers, 7.1% (n = 11) among part-time cashiers and 8.2% (n = 8) among controls (p = 0.776 at Pearson χ2-test).
Univariate analysis showed that full-time cashiers had a 2.3 fold increased risk for CTS specific current symptoms than the control group (95% CI 1.10–4.80, p = 0.026). A similar trend also emerged for CTS cases but was not significant (OR 1.23; 95% CI 0.42–3.56, p = 0.702).
Some variables reported by the literature as possible or probable determinants of CTS were then included in a logistic regression model: BMI ≥30, age, duration of previous at-risk jobs expressed in years, family cases of CTS, presence of children, hand-knitting and/or needlework, education >8 years. Multivariate logistic regression analysis confirmed the increased risk for CTS current symptoms in full-time cashiers. A significant increased risk for CTS current symptoms also emerged for age and hand-knitting/needlework.
Multivariate logistic regression analysis according to job title and other possible determinants of CTS
CTS current symptoms
Previous at-risk jobs
CTS familiar history
Education >8 years
Discussion and conclusions
The relationship between “high” levels of exposure to biomechanical risk factors (repetition, force, and posture) and upper limb musculoskeletal disorders is well-known, whereas few data exist regarding the dose–response relationship for “moderate” levels of exposure. Moreover, the relationship between biomechanical risk factors and upper limb musculoskeletal disorders still presents some “grey areas.” To begin with, dose estimation is usually performed through exposure assessment (number of movements per time interval, weight of objects handled, joint deviation from neutral position, etc.). In addition, there is a lack of scientific evidence about the contribution of different combinations of risk intensity and duration (peak, average, and cumulative) (Bernard 1997; Nordstrom et al. 1997; Tanaka et al. 1997; Giersiepen et al. 2000; Fallentin et al. 2001; Fallentin 2003).
In line with other studies, the biomechanical risk assessment performed according to the method proposed by ACGIH showed that our overall population of supermarket cashiers was at increased risk for CTS.
Full-time supermarket cashiers had a higher prevalence of CTS past symptoms and cases than both part-time cashiers and control subjects. The evaluation of current symptoms followed the same trend. In particular, the prevalence of current symptoms in part-time workers and controls resulted to be similar to that of the general female population (Atroshi et al. 1999).
According to the method used, part-time and full-time cashiers showed the same combination of force and HAL for the hand-wrist-forearm segment. Nevertheless, the clinical picture resulting from our data showed a higher prevalence of both CTS symptoms and CTS cases only among full-time cashiers. It is well-known that the lack of sufficient rest-time intervals, associated with biomechanical risk factors such as force, repetitiveness and awkward postures, increases the risk to develop upper-limb musculoskeletal disorders (Putz-Anderson 1988). Therefore, the different distribution of effects among our study population of full-time and part-time cashiers might be attributed to the different recovery time amount and distribution within the week.
Univariate analysis showed an increased risk for full-time cashiers with both case definitions. Only for CTS current symptoms, the odds ratio reached the level of significance, probably due to the small number of CTS cases registered in the overall study population.
The inclusion of possible or probable determinants of CTS in the logistic regression model did not alter the effect of job title and showed, for both case definitions, a significantly increased risk according to age, confirming what previously reported in other studies (de Krom et al. 1992; Nordstrom et al. 1997; Tanaka et al. 1997; Nathan et al. 2002).
For the case definition based on current symptoms and NCS abnormalities (CTS cases), a more pronounced risk was observed in relation to individual variables (BMI and CTS family history), if compared to environmental variables relating to manual activities (job title, children, hand-knitting/needlework).
The contribution of BMI and of genetic factors to CTS has already been clearly demonstrated by several studies (de Krom et al. 1992; Nordstrom et al. 1997; Tanaka et al. 1997; Hakim et al. 2002; Nathan et al. 2002).
As regards BMI, a possible effect of obesity on nerve conduction could be hypothesized, although studies on this subject have reached different conclusions. Kouyoumdjian et al. did not find an association between more severe nerve conduction abnormalities and higher BMI in clinical patients, whereas Bland demonstrated that increased BMI is a significant independent risk factor for CTS in patients under the age 63 in whom NCS confirmed CTS (Kouyoumdjian et al. 2002; Bland 2005).
It could be argued that the case definition based on CTS current symptoms only might include non-specific hand/wrist pain or tendonitis symptoms. It must be however said that CTS current symptoms were defined according to the specific criteria established in a consensus report and that their exact localization on the hand diagram further increased the specificity of CTS diagnosis (Katz and Stirrat 1990; Katz et al. 1990; Rempel et al. 1998). In addition, the prevalence of CTS symptoms both among cashiers and control subjects resulted consistent with that reported in other studies (Margolis and Kraus 1987; Osorio et al. 1994; Atroshi et al. 1999).
To conclude, we can assume that repetitive manual work produces a compression of the median nerve at the carpal canal level due to cumulative trauma and localized ischemia and that the following reduction in nerve fibers microcirculation produces an alteration of the membrane excitability and subsequent ectopic impulse generation responsible for symptoms. On the other hand, prolonged exposure, associated with predisposing individual factors, could be responsible for myelin sheets disruption or axonal damage and subsequent median nerve conduction slowing (Rosenbaum and Ochoa 1993).
Furthermore, the hypothesis that intensive manual work associated with inadequate recovery time amount and distribution generates an impairment of the median nerve at the wrist level (which produces symptoms and/or conduction slowing) is supported by the fact that such effect is higher in full-time cashiers, proportionally increasing with duration of hand use. In addition, the importance of manual activity as aetiologic factor is also confirmed by the increased risk associated to other non-occupational risk factors responsible for hand/wrist overload (presence of children and hand-knitting/needlework).
Our data, though collected in the context of a cross-sectional study, allowed us to identify the risk of CTS associated with two levels of exposure characterized by the same force and repetitiveness but different recovery time amount. Increased occupational risk for CTS was in fact found in full-time cashiers, whereas the risk level in part-time cashiers did not result significantly higher than that of the control group.
Only longitudinal investigations on the future development of symptoms and nerve conduction abnormalities related to CTS will contribute to a further understanding of the exposure–response relationship at medium levels of exposure. In a prevention perspective of upper limb work-related musculoskeletal disorders, the results of our study can provide useful information both for ergonomic risk assessment and work organization.
We are grateful to Chiara Scardoni for scientific editing.