Hand hygiene and hand eczema: A systematic review and meta‐analysis

Abstract Hand eczema is a common inflammatory condition of the skin that has been linked to hand hygiene. This systematic review and meta‐analysis aims to determine the risks of hand eczema associated with hand hygiene, including frequency of handwashing, wet‐work and use of alcohol hand rub. A comprehensive search of MEDLINE, EMBASE and Cochrane Library was performed for cohort, case–control or cross‐sectional studies that analysed the association between hand hygiene and risk of hand eczema. Results of individual studies were presented in respective forest plots and pooled summary relative risks were estimated using a random‐effects model. Forty‐five studies were included in analysis. Handwashing at least 8–10 times daily significantly increased risk of hand eczema (relative risk [RR] 1.51; 95% confidence interval [CI]: 1.35–1.68; p < 0.001). The risk was related to handwashing frequency, with higher pooled RR of 1.66 (95% CI: 1.51–1.83; p < 0.001) with increased handwashing at least 15–20 times daily. However, use of alcohol‐based hand sanitizer was not significantly associated with risk of hand eczema. Given the widespread implementation of hand hygiene practices during the COVID‐19 pandemic, there is a pertinent need to understand skin care habits specific to the hands to avoid a greater incidence of hand eczema.


| Selection of articles
Studies published in English language from all countries and evaluating all populations were considered. Cohort, case-control or cross-sectional studies that analysed the association between hand hygiene and risk of hand eczema were included. Studies had to report adequate information such as relative risks (RR), odds ratio (OR) and confidence interval (CI) in order for further meta-analysis to be performed. For studies which did not report such ratios, studies should have crude data such as total cases of hand eczema among those exposed and unexposed. The primary outcome measured in this study is the relative risk of hand eczema in the different exposure groups. Definition of hand eczema in the various studies includes physician diagnosis as well as characteristic signs and symptoms of hand eczema.
From the title and abstract, two reviewers independently selected studies for full-text review based on the inclusion criteria. The fulltext articles were then evaluated independently by two reviewers to determine eligibility for inclusion, and any disagreements were resolved by consensus. Study quality was assessed by two independent reviewers using the Newcastle-Ottawa scale (NOS) 14 for cohort and case-control studies, while cross-sectional studies were assessed using an adapted version of NOS. 15 Studies were scored in three areas: selection of study population, comparability between groups and assessment of outcome. A maximum score of 9 or 10 could be achieved for cohort studies and cross-sectional studies respectively, and NOS score of ≥7 was considered low risk of bias or high quality.
Certainty of evidence for each outcome was rated by two independent reviewers using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) assessment tool.
Evidence from observational studies started at the low quality level, and was subsequently assessed across various domains including risk of bias, imprecision, inconsistency, indirectness and publication bias for downgrading or upgrading. 16,17 Any discrepancy in rating was resolved by consensus.

| Data extraction
Data were extracted from the selected studies by two reviewers independently, using a standardized data extraction form. Information extracted included: study year, country and population, study design, total number of participants, outcome and assessment of outcome, risk ratios and 95% CI for groups compared. For studies which did not report risk ratios, crude data were extracted, including total cases of hand eczema, number of cases exposed and unexposed. Studies that provided insufficient information to calculate relative risks of hand eczema or their standard errors among groups of interests were excluded. Where possible, efforts were made to contact the authors for more information.

| Meta-analysis
Selected studies were classified based on whether they analysed frequency of handwashing or use of alcohol hand rub. Studies related to handwashing were then divided into two groups based on the threshold of daily handwashing frequency that was compared: at least 8-10 times versus <8-10 times, and at least 15-20 times versus <15-20 times.
Studies that investigated wet work (defined as contact with liquids >2 h/day, use of occlusive gloves >2 h/day or handwashing >20 times/ day) were analysed separately. Studies related to alcohol hand rub were divided into three groups: use of alcohol hand rub versus no use of alcohol hand rub, alcohol hand rub >10 times daily versus ≤10 times daily and alcohol hand rub >20 times daily versus ≤20 times daily.
In both controlled and uncontrolled studies, most of the included studies reported odds ratios or risk ratios. They were included for metaanalysis when available; otherwise, ratios were estimated from the crude data. Pooled estimate of relative risk (RR) from selected studies was derived from these ratios, as they approximated one another mathematically under the rare disease assumption. 18 Results of individual studies were presented as a forest plot and the pooled summary relative risks was estimated using random-effects model of DerSimonian and Laird to account for variance between and within the studies. Heterogeneity between studies was assessed using χ 2 test and the I 2 statistic; values of 25, 50 and 75% were considered to be low, moderate and high heterogeneity, respectively. A funnel plot was constructed and visually inspected for asymmetry to qualitatively assess publication bias.
All analyses were performed using STATA Version 13.0 (StataCorp).

| Search results
The comprehensive search of MEDLINE, Embase and Cochrane Library yielded a total of 1020 studies, of which 256 duplicates were removed from further evaluation. Based on the titles and abstracts of the remaining 764 articles, 605 studies were excluded. Among the 159 full-text articles assessed, 44 studies fulfilled the inclusion criteria and the rest were excluded for reasons reported in the PRISMA flow diagram ( Figure 1). An additional article was found from the citation list of another article. Finally, 45 studies were included in the systematic review, and 42 studies included in at least one meta-analysis.

| Description of included studies
Six studies utilized a cohort study design, 5,12,[19][20][21][22] three studies were case-control studies, [23][24][25] while the other studies (n = 36) employed a cross-sectional study design. There were 17 studies performed in Asia, and 28 were from non-Asia countries. The majority (n = 28, 62.2%) of studies were performed in healthcare workers, 12 (26.7%) were in non-healthcare workers and five included the general population regardless of occupation. 23,[26][27][28][29] Outcomes were assessed by self-reported questionnaire in most studies (n = 35), of which 12 were based on the NOSQ-2002 questionnaire, 30 and 10 studies relied on clinical examination by dermatologist or trained professional. Most of the studies were of high quality (n = 35), and 10 were considered lower quality (NOS < 7). The general characteristics of each study are summarized in Table 1. Further details are provided in Tables S2-S5. 3.3 | Risks of hand eczema with at least 8-10 times of daily handwashing The meta-analysis included 29 studies that examined the risks of hand eczema with at least 8-10 times of daily handwashing versus fewer than 8-10 times (Table S2). The pooled RR of hand eczema among those who washed their hands at least 8-10 times daily was 1.51 (95% CI: 1.35-1.68; p < 0.001), as compared to those who washed their hands fewer times ( Figure 2). There was moderate heterogeneity between studies (I 2 = 71.0%, p < 0.001), hence a random-effects model was used. The funnel plot appeared symmetrical and did not show obvious publication bias ( Figure S1).

Identification of studies via databases and registers
Identification of studies via other methods       Figure S4).
Geographical region did not significantly affect the association of handwashing 15-20 or more times daily with hand eczema.
The two studies with a low NOS quality score 36

| Risks of hand eczema with wet work
Seven studies examined the risks of hand eczema with wet work versus no wet work (Table S4). As represented in Figure 4, the pooled RR of hand eczema with wet work was 1.37 (95% CI: 1.24-1.51, p < 0.001). There was low heterogeneity among the studies (I 2 = 19.2%, p = 0.283) and there was no significant publication bias seen in the funnel plot ( Figure S5).

| Risks of hand eczema with use of alcohol hand rub
Fourteen studies examined the risks of hand eczema with the use of alcohol hand rub; however, two studies 22,43 were not included in final meta-analysis as the frequency of alcohol hand rub reported was different from the rest of the other studies (Table S5a,b). The remaining 12 studies were analysed as three groups based on frequency of alcohol disinfectant use: use of alcohol disinfectant versus no use of alcohol disinfectant, more than 10 times daily versus ≤10 times daily and more than 20 times daily versus ≤20 times daily. There was high heterogeneity among the studies included in the meta-analysis (I 2 = 93.6%, p < 0.001).
The funnel plot appeared asymmetrical, suggesting a publication bias ( Figure S6).
There was no statistically significant relationship between risks of hand eczema and use of alcohol hand rub (p = 0.548), alcohol hand rub more than 10 times daily (p = 0.196) or alcohol hand rub more than 20 times daily (p = 0.452), as shown in Figure 5.
In the studies that were only included in qualitative analysis, Lan et al. 43 found that there was no statistically significant risk of hand eczema with use of alcohol hand rub more than nine times within 4 h versus ≤9 times (p = 0.2886). A prospective cohort study by Yüksel et al. 22 described that increased use of alcohol-based hand rubs on wet skin by healthcare workers during the COVID-19 pandemic was associated with increased 1-year prevalence of hand eczema at follow up (RR: 1.78; 95% CI: 1.11-2.87).

| GRADE assessment: certainty in evidence
The GRADE certainty ratings for the following outcomes: risks of hand eczema from at least 8 to 10 times handwashing, 15 to 20 times handwashing and wet work were low. The GRADE certainty rating for risk of hand eczema from alcohol hand rub was very low (Table 2).
Therefore, the overall GRADE quality rating for risks of hand eczema

| DISCUSSION
This study demonstrated a significant increase in risk of hand eczema associated with frequency of handwashing and wet work, but not with use of alcohol hand rub.
Handwashing at least 8-10 times a day significantly increased the risk of hand eczema (RR: 1.51) as compared to washing hands fewer times; the risk was even higher when handwashing frequency was increased to at least 15-20 times a day (RR: 1.66). The associations between hand hygiene practices and risk of hand eczema were consistent regardless of geographical region or occupation. It is also noted that there could possibly be a dose-response relationship given that the pooled risk ratios of hand eczema were higher with more frequent handwashing. However, our meta-analysis results showed that 8-10 times of daily handwashing is enough to cause a significantly higher risk of hand eczema than someone who washes hands less frequently.
On the other hand, no significant association has been established between use of alcohol hand rub and risk of hand eczema. However, healthcare workers often perceive alcohol disinfection to be more damaging to the skin than handwashing, 49 Besides handwashing and alcohol hand rubs, wearing of occlusive gloves is also a risk factor for hand eczema. 38,47 The gloves lead to a state of hyper-hydration causing maceration of the skin, enhancing the penetration of soaps and alcohol sanitizers. 68 This meta-analysis included studies on wet work which encompassed the use of occlusive gloves, but did not examine the independent role of gloves in increasing the risk of hand eczema. Although glove use is less common among the general population, it is a factor that should also be considered in future studies in the context of healthcare workers who routinely use gloves at work. This review has several limitations. First, the data were gathered from observational studies that were prone to the effects of  30 for screening of hand eczema and exposures, others relied on self-reported questionnaires that may not have been validated in detecting hand eczema accurately. Self-reported questionnaires have been found to demonstrate high specificity of over 90%, but sensitivity is less than 70%, and hence tend to underestimate the true prevalence of hand eczema. 30 In conclusion, this meta-analysis highlights the significant risk of hand eczema associated with handwashing, but not the use of alcohol hand rubs. This risk is observed regardless of geographical region or population. The burden of hand eczema is especially significant amidst the current COVID-19 pandemic, when a higher frequency of hand hygiene has been recommended for the general public. Knowledge of this risk is valuable in underscoring the need to encourage hand care to reduce the incidence of hand eczema.

CONFLICTS OF INTEREST
The authors declare that there are no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are provided in supplementary tables. These data were derived from resources available in the public domain and have been referenced.