Influence of environmental temperature on the occurrence of non-necrotizing cellulitis of the leg

Sir, Non‐necrotizing cellulitis of the leg is a common cutaneous bacterial infection whose risk factors include venous insufficiency, lymphoedema and toe‐web intertrigo. The role of environmental temperature remains controversial.1 To study the relationship between environmental temperature and the frequency of non‐necrotizing cellulitis of the leg, we reviewed all patients hospitalized with non‐necrotizing cellulitis of the leg in a university hospital in a temperate region of France during a 4‐year period, and correlated findings with the local environmental temperature during the same period. Patients included in this study were consecutive patients referred for non‐necrotizing cellulitis of the leg from January 1995 to December 1998 in the Departments of Dermatology, Internal Medicine and Infectious Diseases of the Rouen University Hospital. The diagnosis of cellulitis of the leg was retrospectively identified using the database of the Medical Information System. The mean and maximum temperatures of each the 8 days prior to the date of hospital admission were obtained from the local meteorological unit for each case. Correlations between the mean and maximum daily temperature of each day from day − 8 to the admission day (lag 0–8 days) and the daily number of patients hospitalized for cellulitis of the leg were studied using a nonparametric Poisson regression model to adjust for time trends and days of the week (generalized additive model).2 Eight hundred and ninety‐eight patients with cellulitis of the leg [342 men (38%) and 556 women (62%)] were assessed. The mean age of patients was 66·1 years. The mean ± SD daily number of hospitalizations for leg cellulitis was 0·60 ± 0·8. The mean ± SD for maximum and mean daily temperature was + 14·6 ± 7·3 °C and + 10·6 ± 6·2 °C, respectively (range − 11·1 to + 26·8 °C). During the study period, we found a positive correlation between the daily number of hospitalizations for cellulitis of the leg and the maximum daily temperature recorded during the 2 days prior to admission [Student's t‐test = 1·3 (β‐coefficient =0·0062; SD = 0·0048) and 1·79 (β‐coefficient = 0·0085; SD = 0·0047), respectively (Fig. 1)]. No correlation was found when testing the days prior to day − 2 (day − 8 to day − 2). Relationship between the daily number of patients hospitalized with cellulitis of the leg and the maximum daily temperature (°C) on the day of admission (left) and 2 days before admission (right). A relationship between the occurrence of leg cellulitis and seasonal factors was first suggested by Boycott, who conducted a quarterly analysis and reported a slight increase in infections due to Streptococcus pyogenes during the third quarter.3 Most of these were posttraumatic and/or cutaneous infections. Similar results have been reported by Ronnen et al. in Israel, in a series of 526 adult patients who showed an increase in number of cases of erysipelas during the summer period.1 In contrast, no relationship was found between the mean number of monthly cases of erysipelas and seasonal factors in two other European studies.4, 5 It is likely that these contradictory results were at least partly due to the large periods considered in these analyses, which may have produced discrepancies in the average temperature. This prompted us to perform a study in a larger number of patients using a powerful statistical method: we did not consider the quarterly or monthly number of cases of leg cellulitis, but rather the daily number, which was correlated with the environmental temperature recorded for each of the 8 days before hospital admission, using a nonparametric Poisson regression model. Such a method has previously been used to demonstrate a correlation between air pollution and the number of hospital stays for respiratory illness in the U.S.A.2 The present study showed an association between the number of hospital stays for leg cellulitis and the maximum temperature during the 2 days prior to hospitalization. Considering that the main risk factors for cellulitis—venous insufficiency and lymphoedema—are classically aggravated by hot weather, the present study suggests that hot weather might trigger or favour the occurrence of non‐necrotizing cellulitis of the leg.