ORIGINAL ARTICLE Year : 2019  Volume : 37  Issue : 3  Page : 134138 The relationship between pityriasis rosea, seasonal factors, and other herpetic infections: A time series analysis Sibel Berksoy Hayta^{1}, Rukiye Güner^{1}, Selim Çam^{2}, Melih Akyol^{1}, ^{1} Department of Dermatology, Cumhuriyet University School of Medicine, Sivas, Turkey ^{2} Hospital Statistics Unit, Cumhuriyet University School of Medicine, Sivas, Turkey Correspondence Address: Background: Data on seasonal variation and viral etiology in Pityriasis rosea (PR) have been conflicting. The aim of this study was to investigate the association of PR and other herpetic infections, taking seasonal changes into account. Methods: The data were collected retrospectively from electronic health registry systems in Sivas in the Central Anatolia region of Turkey between 2008 and 2016. According to their clinical types, other herpetic infections were investigated. Environmental factors such as humidity, temperature, and rainfall for the relevant period were added into the model. Time series methods (augmented Dickey–Fuller unit root test and regression analysis) were used in the analysis. Results: A total of 1207 PR patients were included in the study. The number of PR patients was calculated to be 0.462 times that of the same period in the previous year. The incidence of PR increased significantly when the rate of infections caused by varicellazoster virus decreased and the rate of infections caused by herpesvirus Type 1 and humidity increased (P < 0.05). Conclusion: Environmental factors such as humidity are important in the emergence of the PR. Furthermore, the incidence of PR may be inversely affected by varicellazoster infections contrary to the relationship between PR and herpesvirus Type 1 infections.
Introduction Pityriasis rosea (PR) is a common and selflimiting skin disorder characterized by large patches known as herald patch and mostly settled along cleavage lines. PR is most common in children and young adults between the age of 10 and 35 years. The etiology of PR is unclear, but it is suggested that viral infections are one of the causes.[1],[2] Seasonal variations may predict the frequencies of PR; however, there are conflicting results.[3],[4],[5],[6] The viruses, including Herpesviridae, are currently important in the pathogenesis of PR. DNA of human herpesvirus (HHV)6 and HHV7 has been isolated from tissues, including lesional and nonlesional skin, peripheral blood mononuclear cells, and serum and saliva samples of patients with PR.[7],[8],[9] Although acyclovir is effective in the treatment of PR, the efficacy of acyclovir on HHV6 and HHV7 deserves further investigation.[1],[10],[11] Herpesvirus infections are very common worldwide. Approximately 130 herpesvirus types have been identified. Nine of the herpesvirus types are human pathogens, i.e., herpes simplex virus type (HSV)1, HSV2, human cytomegalovirus, varicellazoster virus (VZV), Epstein–Barr virus and HHV6A, HHV6B, HHV7, and HHV8.[12] It is know that environmental and immunological factors may be responsible for the seasonality of many viral infections.[13] Epidemiological investigations, as well as clinical studies on infectious agents, will be important to reveal the cause of PR. Infectious origin of the disease may lead to clustering. Other environmental and seasonal factors such as humidity and temperature should be taken into account. In addition, it is known that the interaction between different viral infections has an important role in the occurrence of some viral infections.[14] Previous studies only focused on seasonal clusters. However, there is no epidemiologic data on PRs relation to other herpetic infections. The aim of this study was to investigate the association of PR with other herpetic infections, taking seasonal changes into account. Methods Study design The study was approved by the Local Ethics Committee. This retrospectively data were obtained from electronic registry systems from secondary state hospitals, a private hospital and a university hospital in the province of Sivas in the Central Anatolia region of Turkey between 2008 and 2016. Herpetic infections were clustered into three groups according to virus types (VZV, HSV1, and HSV2) [Table 1]. First, PR was defined as a dependent variable. Then, in order to determine the twincausality between the groups, each other study group was also further analyzed as a dependent variable in separate stages of statistical analysis. In addition, data related to the periodic humidity, temperature, and rainfall were created. Monthly average values of climate data based on dataset were taken from Sivas province center of General Directorate of Meteorology.{Table 1} Statistical analysis Statistical analysis was performed using the SPSS for Windows version 16.5 (SPSS Inc., Chicago, IL, USA) software package. Descriptive statistics was presented as mean ± standard deviation and median (interquartile range) for continuous variables according to the results of normality tests, as appropriate. Normality testing was done by the onesample Kolmogorov–Smirnov test. Time series methods (augmented Dickey–Fuller [ADF] unit root test and regression analysis) were used in the analysis of observation data. In the study, addressing the total number of patients per month between 2008 and 2016, there are data for 96 cycles (monthly). The use of climate features may lead to unit differences (cm3, °C, etc.) between data. To avoid a possible problem, the ADF test was used instead of the Dickey–Fuller test. Thus, it is aimed to achieve more accurate results with a method that allows more complex dynamics and eliminates the autocorrelation of the remains. In addition, the ADF test may also show the trend effect of the variables and the trend effect of such variables may be eliminated. Time series analysis of variables with stability was made with Gnu Regression, Econometrics and Timeseries Library (GRETL), version 15 opensource code package program. Autoregressive integrated moving average (ARIMA) models are the most general class of models for predicting a time series, which can be made to be “stationary” by differencing, perhaps in conjunction with nonlinear transformations such as logging or deflating. X13 ARIMA analysis in the package program was used to predict models with the available variables for the patients with PR. After the most suitable model has been determined by X13 ARIMA analysis, statistical analyses have been performed. Nonsignificant variables were taken out to create a new model because the coefficients of the variables to be used in estimating will change the appearance of nonsignificant variables. The modeling steps were repeated to avoid errors in the calculation. The X13 plugin in the GRETL program automatically calculates the appropriate autoregressive, delay, and seasonality values in the ARIMA process. To validate the X13 results, all autoregulatory, delay, and seasonality values were calculated manually and the results were compared. In both trials (X13 and singular trials), the appropriate model is ARIMA (0, 0, 1). To verify the X13 ARIMA analysis, the basic ARIMA (p, d, q) model has been tested at each parameter 0–5. Thus, according to the test statistics, the most suitable model was selected among the 36 models. Results The study included a total of 22,367 patients with 1207 PR patients (781 females [the mean age, 30.3 ± 12.5] and 426 males [the mean age, 29.3 ± 12.7]), 18,227 Group 1, 2816 Group 2, and 117 Group 3. The weather conditions during the study were conducted as follow: humidity (%), 63.38 ± 11.82; temperature (°C), 10.14 ± 8.53; and rainfall (mm), 40.08 ± 27.64. The result of the unit root test for ensuring the stability of the variables is shown in [Table 2]. To be able to do the time series operation, first of all, the stationarity of the variables included in the model was examined. To calculate the stationarity in the variables, the existence of the fixed effect and the trend effect must be eliminated. In the tests used in the GRETL package program, the stability test was performed with ADF test. The results showed that there was no trend effect but a variable effect. Therefore, the ARIMA model was created.{Table 2} [Figure 1] shows the monthly numbers of PR patients are opposite or parallel movements to humidity, temperature, and Group 1 and Group 2 patients in years. The effect of temperature is not related to the increase or decrease in PR patients (P = 0.40). The study variables were examined to determine if they influenced the number of PR patients [Table 3]. Accordingly, the change in Group 3 did not cause any change in the number of PR patients (P > 0.05). The humidity was the only seasonal variant with a significant effect on the number of monthly PR patients (P < 0.05). The decrease in Group 1 patients and the increase in Group 2 patients significantly increase the number of PR patients (P < 0.05).{Figure 1}{Table 3} [Table 3] shows the modeling significance of all variables. Even though some variables did not affect the number of PR patients, they influenced all coefficients of the analysis. Therefore, the following variables, i.e., Group 3, temperature and rainfall were excluded. A new model was created with significant variables to eliminate the bias and to reach the final statistical model and coefficients. [Table 4] shows the final coefficients of factors affecting PR disease. [Table 5] shows the final coefficients and interactions among PR and other study groups.{Table 4}{Table 5} Our results show that PR disease can be calculated seasonally. The model created as a result of the analysis was determined as ARIMA (0, 0, 1). The number of PR patients was calculated to be 0.462 times that of the same period in the previous year. Discussion Although the etiology of the disease is not completely known, epidemiological and clinical studies support that infectious agents can cause PR.[1],[2],[13] PR is seen throughout the year but studies have shown that it can occur more frequently in certain seasons, in contract to this, other studies have shown that no seasonal effect on PR.[4],[5],[6],[15],[16],[17] Clusters can be observed in terms of variables such as seasonal temperature, rainfall, and humidity. For example, although a slightly higher incidence of PR occurs in the months with less rainfall, there are conflicting results in the literature with regard to rainfall.[3],[15] ,[18],[19],[20],[21] Many viral pathogens are seasonal due to dynamic environmental and immunological mechanisms.[13] However, it is possible that PR may be due to reactivation of a latent virus rather than primary viral infection.[22] Chuh et al.[15],[16] believe that temporal clustering of PR is compatible with viral reactivation. HHV6 and HHV7 are capable of reactivation and PR is an acute, selflimiting exanthem associated with endogenous reactivation of HHV6 and HHV7.[23] VZV infections may also cause viral reactivation diseases and have temporal clustering. VZV causes two distinct diseases, i.e., varicella and herpes zoster. Although the incidence of varicella peaks during the cold months, the incidence of herpes zoster peaks most frequently during the summer months. The variations in temperature are important for the incidence of varicella and herpes zoster infections.[24] Furthermore, relative humidity is inversely correlated with varicella incidence.[25] Thus, it should be taken into account that meteorological factors are important in the epidemiology of these viral diseases.[26] Studies have shown that certain viruses are more active when the humidity increases.[27],[28] We found that PR incidences increased when the humidity was higher; however, the temperature had not an influence on PR prevalence. This could be explained by the fact that HHVs may be more active in increasing humidity. Humidity is an important seasonal factor in the development of PR. The results of our study showed that PR was inversely related to VZV infections. We know that a virusinfected cell becomes resistant toward a second infection by a closely related or a different virus species.[14] At this point, the reason for this reverse relationship between PR and VZV infections may be related to the interactions of both viral infections. It is known that Type I IFNs are important for host defense against viruses, but they can cause immunopathology in some acute viral infections. Indeed, high concentrations of Type I IFNs block Bcell responses or lead to the production of immunosuppressive molecules.[29],[30] On the other hand, there is a positive relationship between PR and Group 2 (HSV1 group). Interactions between viral infections may be the same or may be inverse. Coinfections may be observed with some virus species.[31] In addition, HSV1 infections are highly contagious, which are common and endemic throughout the world. The incubation and the duration periods of HSV1 infections are rather short. Therefore, it may not be possible time intervals in which interaction may occur between both PR and HSV1 infections. Although varicella infections also increased significantly during periods of increased HSV1 infections, this increase does not seem to affect the negative relationship between PR and HSV1 infections according to our results. In conclusion, to the best of our knowledge, this study presents the first epidemiological data showing a relationship between PR, environmental factors, and other herpetic infections and also supports a viral etiology in PR. The results of our study indicate that environmental factors such as humidity are important in the emergence of the disease. The incidence of PR may be inversely affected by varicella infections contrary to HSV1 infections. Although there is a significant epidemiological relationship between PR and other herpetic infections, further molecular studies are needed to explain these arguments. Acknowledgment We would like to thank to Prof. Nazif Elaldi (Cumhuriyet University School of Medicine, Department of Infectious Disease, Turkey) for the comments in discussion. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest. References


