Abstract
Capsule Summary
Varicella zoster virus infection in childhood prior to onset of atopic dermatitis (AD) is a protective factor for AD.
Varicella vaccination may be used as an effective, easily available, and economical option to decrease the severity of AD.
INTRODUCTION
Atopic dermatitis (AD) is an intensely itchy, chronic, relapsing inflammatory skin condition affecting approximately 25% of children and 4%–7% of adults. 1 It has been estimated that approximately 85% of AD appears before 5 years of age, but adult-onset AD is also common comprising approximately 25%. 1 Most of the cases of AD resolve in childhood; however, around 33% cases persist till adulthood. 1
AD is considered as the flag-bearer of atopic march, and other atopic conditions such as asthma, allergic rhino-conjunctivitis (AR), and food allergy follow the appearance of AD in predisposed patients. The etiology of AD is multifactorial. In “hygiene hypothesis, it has been suggested that unhygienic contacts and infections in childhood may confer protection against hay fever, AR, and AD. It is based on the observation that the Th1 response induced by infections might counterbalance the Th2 response induced by allergen. 2 However, there has been conflicting evidences regarding the role of infection in protecting against AD. It has been observed that certain childhood viral infections lower immunoglobulin-E (IgE) levels, decrease IgE sensitization, and asthma, but some viral infections like measles increase the risk of AD. 2 Silverberg et al. found that wild-type varicella zoster virus (WTVZV) infection in childhood confers protection against AD.2,3 This may suggest that only specific infection confer protection against AD and others do not.
The number of CD1a+ Langerhans cells is reduced, and their immune functions are altered in the vicinity of herpes zoster skin lesions. 4 Furthermore, it was found that varicella zoster virus (VZV) infection results in downregulation of major histocompatibility complex I and II expression and reduction or inhibition of intercellular adhesion molecule 1 (ICAM-1) expression on keratinocytes, thereby disabling the keratinocytes to act as accessory antigen presenting cells, and inhibiting its role in ICAM-1/leucocyte functional antigen-1 (LFA-1) mediated T-cell responses. 4
In the current study, we use a comparative cross-sectional study to determine the association between VZV infection/varicella virus vaccination (VV) in childhood and the severity of AD.
MATERIAL AND METHODS
Study Design and Setting
The study was a comparative cross-sectional study conducted at the dermatology outpatient department and multidisciplinary research unit of S.C.B Medical College and Hospital, Cuttack, between December 2023 and February 2025. Ethical clearance was taken from the institutional ethical committee [IEC/SCBMCH/1533/23.11.2023].
Study Population
Children in the age group of 1–14 years were considered as study population for this study.
Sample Size Considerations
The maximum sample size for this study was found to be 206. The prevalence was taken as 7.21% as per study done by De et al. 5 The type I error was taken as 0.05. Estimated effect size was taken as 2, and desired level of absolute precision was taken as 0.05.
Inclusion Criteria
Children aged 1–14 years diagnosed as AD according to Hanifin and Rajka criteria.
Exclusion Criteria
Children giving history of AD prior to VZV infection or VV, having history of autoimmune disorder or immunodeficiency, hyper IgE syndrome, and hypereosinophilic or lymphoproliferative disorder were excluded from the study.
Data Collection Technique
Children aged 1–14 years diagnosed with AD according to Hanifin and Rajka were enrolled in the study. Demographic and clinical characteristics like age, gender, season of birth, age of onset of AD, number of episodes, frequency of visits to the hospital, season of first episode, and presenting symptoms were collected using a structured questionnaire. The severity of AD was calculated using SCORing Atopic Dermatitis (SCORAD) severity scale. The history of varicella infection was taken by showing the clinical images to patients and explaining about disease in their local language. The history of varicella VV was taken from the immunization card of the patient. History of other protective (rural setting, breastfeeding, ultra-violet light exposure, consumption of unpasteurised milk during first 2 years of life, exposure to pets, consumption of fresh fruits and fish, and low birth weight) and risk factors (urban setting, delayed weaning, obesity, pollution/tobacco smoke, post natal antibiotic, having attention deficit hyperactive disorder, and high birth weight) were taken.6-8 Serum varicella immunoglobulin G (IgG) titer was done for the AD patients who gave consent for the same. Varicella zoster virus IgG titer was done using a quantitative enzyme-linked immunosorbent assay method, and an optical density ratio of less than or equal to 0.8 was considered negative, and more than 0.8 was considered positive.
Statistical Analysis
The data were entered into a computer-based spreadsheet and cleaned. The cleaned data were analyzed using Statistical Package for Social Sciences (SPSS) software, IBM manufacturer, Chicago, USA, version 25.0. The statistical analysis comprised calculating mean, median, and proportions. Categorical variables were presented in the form of frequency and percentage (%), and the quantitative data with normal distribution were presented as mean along with standard deviation and the data with non-normal distribution (or skewed distribution) as median with 25th and 75th percentiles (interquartile range). For normally distributed data, t-test was used for comparison of mean, and for skewed distribution, Mann–Whitney U test was used for comparison of median. Chi-square test (Fisher’s exact test wherever applicable) was used for comparing the proportion. Spearman’s rho correlation coefficient was used for the calculation of correlation between SCORAD and IgG titer. The level of significance was taken as P < 0.05. Logistic regression was used to determine whether VZV infection was associated with a decreased number of hospital visits for AD. Multinomial logistic regression was used to determine whether VZV infection was associated with decreased severity of AD.
RESULTS
Population Characteristics
Two hundred and 34 patients having AD were recruited for the study according to inclusion and exclusion criteria. History of VZV infection prior to onset of AD was present in 68 patients (VZV+) and absent in 166 patients (VZV−). None of the patients gave history of VV. Serum IgG titers were done in 78 study participants. The mean and median age of the patients were found to be 7.23 ± 4.05 and 7 (4–11), respectively. Among 234 cases, there were 110 (47%) males and 124 (53%) females. The demographic and clinical characteristics of all the study participants is described in Table 1.
Demographic and Clinical Characteristics of Study Participants
AD, atopic dermatitis; SCORAD, SCORing Atopic Dermatitis.
Comparision of Various Parameters According to History of Varicella Zoster Infection
The mean and median age of VZV+ study participants was 9.0 ± 4.02 years and 9 (6–12) respectively, and among VZV− participants was 6.51 ± 3.86 years and 6 (3–10), respectively. The mean age of VZV+ study participants was significantly higher as compared to VZV− study participants (P value: 0.002) [Table 2, Figure 1a]. There was no statistically significant association between history of VZV infection and gender (P value: 0.994) [Table 2]. VZV+ study participants had mild severity of AD as compared VZV− participants. There was statistically significant association between history of VZV infection and the severity of AD (P value: <0.001) [Table 2]. VZV infection was associated with decreased odds of moderate AD (multinomial logistic regression: odds ratio [OR] 0.03; 95% confidence interval [CI] 0.01–0.12, P < 0.01) or severe AD (multinomial logistic regression: OR 0.02; 95% CI 0.01–0.19, P < 0.01) [Table 3]. There was no statistically significant difference between other protective (rural setting, breastfeeding, ultra-violet light exposure, consumption of unpasteurised milk during first 2 years of life, exposure to pets, consumption of fresh fruits and fish, and low birth weight) and risk factors (urban setting, delayed weaning, obesity, pollution/tobacco smoke, post natal antibiotic, having attention deficit hyperactive disorder, and high birth weight) between VZV+ and VZV− study participants [Table 4].
Comparison of Various Parameters According to History of VZV Infection
t test and
Mann–Whitney U test.
AD, atopic dermatitis; SCORAD, SCORing Atopic Dermatitis; VZV, varicella zoster virus.
Multinomial Logistic Regression Showing VZV Infection Was Associated with Decreased Severity of AD
AD, atopic dermatitis; VZV, varicella zoster virus.
Comparision of Other Risk and Protective Factors According to the History of VZV Infection
UV, ultra-violet; VZV, varicella zoster virus.
Fisher’s exact test.
The mean and median age of the first episode of AD in VZV+ study participants was 6.78 ± 3.53 years and 7 (4–9.3), respectively, and among VZV− was 3.29 ± 2.38 years and 2.5 (1–5), respectively. The median age of first episode of AD of VZV+ study participants was significantly higher as compared to VZV− participants (P value: <0.001) [Table 2, Figure 1b]. The mean and median number of hospital visits each year for AD in VZV+ study participants was 1.18 ± 0.46 and 1 (1–1), respectively, and among VZV− was 2.14 ± 1.97 and 2 (1–3), respectively. The median number of hospital visit each year for AD of VZV+ study participants was significantly lower as compared with VZV− participants (P value: <0.001) [Table 2, Figure 1c]. The average number of hospital visits each year was taken as dependent variable and VZV infection and age as independent covariate. VZV infection was associated with decreased odds of more than one hospital visit each year for AD (logistic regression: OR 0.12; 95% CI 0.04–0.37, P < 0.01). The mean and median total number of AD episodes in VZV+ study participants was 2.65 ± 1.41 and 2 (2–3), respectively, and among VZV− was 5.54 ± 4.39 and 5 (2–7), respectively. The median total number of AD episodes in VZV+ study participants was significantly lower as compared to VZV− participants (P value: <0.001) [Table 2, Figure 1d]. The mean and median SCORAD of VZV+ study participants were 19.64 ± 7.81 and 18.3 (15.3–22.2), respectively, and among VZV− were 39.20 ± 13.57 and 39.9 (27–50.5), respectively. The median SCORAD of VZV+ study participants was significantly lower as compared to VZV− participants (P value: <0.001) [Table 2, Figure 2a]. The mean and median IgG titer of VZV+ study participants were 1.02 ± 0.64 and 1 (0.6–1.3), respectively, and among VZV− were 0.25 ± 0.26 and 0.2 (0.1–0.3), respectively. The median IgG titer of VZV+ study participants was significantly higher as compared to VZV− participants (P value: <0.001) [Table 2, Figure 2b].
Comparison of SCORAD According to IgG Titer Among Study Participants
The mean and median SCORAD of study participants who had IgG titer >0.8 were 19.95 + 8.92 and 19 (15.3–23.15), respectively, and among those in which IgG titer was ≤0.8 were 27.52 + 15.02 and 35 (23.6–50.75), respectively. The median SCORAD of study participants who had IgG titer >0.8 was significantly lower as compared to those who had IgG titer ≤0.8 (P value: <0.001) [Figure 2c].
Correlation Between SCORAD and IgG Titer Among Study Participants (n = 78)
The spearman’s rho corelation coefficient was −0.452 and there was statistically significant negative (inverse) correlation between SCORAD and IgG titer (P value: <0.001) [Figure 2d].
Correlation between SCORAD and Age of VZV Infection among Study Participants Having Positive History of VZV Infection (n = 68)
The spearman’s rho correlation coefficient was −0.113, and there was no statistically significant correlation between SCORAD and age of VZV infection (P value: 0.526).
DISCUSSION
The pathogenesis of AD is complex and is mainly characterised by immune response dysfunction, T helper type 2 (Th2) dominance, over expression of Th2 cytokines such as interleukin (IL)-4 and IL-13, and lower interferon-γ production. 9 It also exhibits impairment in skin barrier function due to defect in filaggrin and lower production of antimicrobial peptide cathelicidin, causing unbalanced skin microbiota dysbiosis leading to Staphylococcus aureus colonisation and increased risk of bacterial and viral infections.10-12 Although infections by bacteria and viruses are considered harmful, improvement in clinical lesions of AD has been noted, particularly at the site of VZV infection and kaposi varicelliform eruption.9,13,14
The current study demonstrates a strong association between VZV infection and prevention of AD. Our study demonstrates that a single episode of VZV infection in childhood prior to the onset of AD is associated with delayed age of onset of AD, decreased number of hospital visits, and decreased severity of AD. These findings corroborate with the results of Silverberg et al. who reported that WTVZV in children up to 8–10 years of age was associated with decreased OR of developing AD and asthma, delayed onset of AD symptomatology, decreased AD severity, and fewer AD-related office visits. 2 Furthermore, Silverberg et al. found that WTVZV results in decreased allergic sensitization, lower serum IgE levels, and decreased numbers of peripheral blood lymphocytes, monocytes, and basophils leading to decreased severity of AD. 3 However, both the studies of Silverberg et al. were retrospective in design and did not corelate severity with varicella antibody concentrations. Moreover, other risk and protective factors that could act as confounding factors in the study were not analyzed. Our study was of comparative cross-sectional design and quantitatively corelated the severity of AD with varicella IgG titer.
The goal of our study was to elucidate if VZV was a protective factor for AD. As AD has a multifactorial etiopathogenesis, we have matched other established confounding risk and protective factors and the result was insignificant (P > 0.05) for all the matched factors. The two sub-groups were also matched for demographic characteristics, in which all the other factors in both the sub-groups were comparable except age. The mean age of study participants who had history of VZV infection was significantly higher as compared to those who had no history of infection (P value: 0.002), which may be due to higher probability of acquiring the infection as the age advances, but this finding also helps in understanding the true nature of the data in both the sub-groups as recall bias could be an important limitation of our study. Therefore, we can attribute the differences in age of onset of AD, severity of AD, and number of hospital visits of AD patients between the two groups to prior infection with VZV. The average age of VZV infection in our study is 3.11 years with the age of infection ranging from 3 months to 8 years. In our study, there is no relation between the age of VZV infection and development of AD. In the study by Silverberg et al., VZV infection until the age of 8–10 years protects against AD. 2 In another study by Pembrey et al. age at Epstein–Barr virus (EBV) infection and age at VZV infection were not associated with risk of atopy in unadjusted analyses. 15 These findings oppose the hypothesis that the protective effects of viral illnesses against atopic disease in children are solely attributable to the immature immune system found in the first year of life.16-18 Similar results were also seen in a study by Illi et al. who showed that children with WTVZV by age 3 years were less likely to develop asthma by age 7 years. 16 It has also been found that EBV infection in the early ages of life is associated with decreased IgE sensitisation.19,20 Conflicting results have also been found in some studies where they found either no association between VZV seropositivity and IgE sensitisation; however, the limitaion of both the studies are that these were the secondary outcome in their study.20,21
Induction of Th1 regulatory events has been found during infection with both VZV and herpes simplex virus (HSV), with VZV causing more systemic immunostimulation and HSV causing local immunostimulation in the lesions. 9 During VZV infection, CD14+CD16++ monocytes are recruited after stimulation with microbial products and produce high amounts of tumor necrosis factor (TNF) and IL-12. 9 IL-12 is an important immunoregulatory cytokine helping in defense against wide range of viral infections, and may promote Th1 response and regulate Th1 stability. 22 This switch to Th1 predominance after VZV infection may cause improvement in Th2 predominant AD after varicella infection. Fujimura et al. showed a switch from Th2 to Th1 regulation after varicella infection in patients with AD. They also showed that IL-12 plays a key role in the switch from Th2 to Th1 predominance. 13 It has also been found in previous studies that WTVZV infection causes production of cytokines like IL-4, IL-10, and IFN-gamma in vitro, and IL-10, IL-12, and IFN-gamma in vivo. 3 IL-10 induces anergy to allergen-specific peripheral T-cells and also decreases IgE production. 3
Varicella vaccination in children with atopic eczema is essential, as it has been found that children with AD are prone to acquire as well as develop severe complications of varicella. 23 However, it has been seen that many children miss their immunisation due to fear of exacerbation of AD. But it has been found that not only the VV is safe and effective in children having AD, but the severity of AD also decreased after VV.23,24 Kreth et al. did a prospective, randomised, multicenter study and found that after VV, the overall severity of AD as measured by SCORAD fell by approximately 10 points over the 2 year follow-up period. Varicella vaccination was well tolerated with none of the children being withdrawn due to adverse effects. Injection site reaction, fever, and exanthem (but not vesicle) were the most common side effects. Seroconversion rates were 94.3% at week 8 and 88.9% at month 12. 25 However, Silverberg et al. found that the decrease in lymphocyte, monocyte, and basophils were more persistent with WTVZV strain as compared with VV, and hypothesized that presence of more virulence factors in WTVZV and severe clinical infection may be the cause for the more persistent change in leucocyte distribution after WTVZV infection. 3 In our study, the effects of VV on the severity of AD could not be ascertained, as none of the patients gave history of VV, as VV is optional for Indian population and not present in national immunization schedule. Conducting a prospective study in this regard could help to ascertain the effect of VV to reduce the severity of AD.
The strength of our study is the cross-sectional design. To the best of our knowledge, this is the first study to investigate the effect of VZV infection on childhood AD in the Asian population. We have eliminated the effect of confounding variables by matching the two groups for all established protective and risk factors for AD, which was not done in any previous studies. To increase the robustness of the study, we have determined the varicella IgG titer in all patients and corelated it with the severity of AD measured by SCORAD.
There are a few limitations to this study. There may be recall bias and selection bias while collecting data. We have tried to rectify this by making the questionnaire objective. SCORAD is not completely objective; therefore, subjective aspects of the scoring system like sleep loss may not be ascertained precisely for very young children. This is a site-limited study. Important laboratory parameters such as IgE and characterization of lymphocyte-monocyte subset were not done due to resource constraints.
This study shows that VZV infection in childhood prior to the onset of AD delays the onset of AD, decreases the number of hospital visits, and decreases the severity of AD. This was further strengthened by the finding that varicella IgG titer was inversely proportional with SCORAD. The exact mechanism of the protective action of VZV is not delineated. This provides scope for more research to find the mechanism of immune changes caused by VZV which could help in better understanding of etiopathogenesis of the disease and find new modes of treatment. Moreover, VV can be used to decrease the severity of AD. The effects of VV on the severity of AD need to be studied further. Therefore, from this study, we conclude that VZV infection in childhood prior to onset of AD is a protective factor for AD, and VV may be used as an effective, easily available, and economical option to decrease the severity of AD.