Seasonal variation of Mycoplasma infections in Iceland and Israel: different associations with meteorological variables.

We read with great interest the article by LIEBERMAN et al. [1] published in this journal last December regarding community-acquired pneumonia in the Beer-Sheva area of Israel. We were especially interested in their results regarding Mycoplasma infections. They found the peak incidence of pneumonia caused by Mycoplasma to be during spring, but not in the winter months (p=0.054). Furthermore, they concluded that meteorological variables, such as low temperature and high humidity, were not predominant causes of community-acquired pneumonia. Their findings regarding Mycoplasma infections are in some aspects in contrast to our experience, which is detailed below. In Iceland, all measurements of Mycoplasma antibodies are performed at the Dept of Medical Virology, National University Hospital in Reykjavík. We have reviewed all test results over a period of 9 yrs from 1988 to 1996. During this period, a total of 11,791 serum samples were tested for Mycoplasma antibodies and 716 individuals turned out to be seropositive by a conventional complement fixation test. The Mycoplasma antigen used in the test was supplied by Behringwerke AG (OTHI 04/05; Mycoplasma pneumoniae antigen, Marburg, Germany). In our study, seropositivity was defined as a titre of ≥1:64 in a single sample or a fourfold increase of Mycoplasma titre in paired samples as reported in some other studies [2–5]. Statistical significance was evaluated with the Chi-squared test and the level of significance set at p<0.05. Mean age of the seropositive individuals was 22 yrs; 52% females and 48% males. Figure 1 shows the seasonal distribution of Mycoplasma infections in Iceland and Israel [1]. In Iceland, a significant seasonal difference in Mycoplasma infections was observed during this period (p=0.001). Thus, it is clear from figure 1 that Mycoplasma infections in Iceland were considerably less common in the spring and early summer than during the winter months. In contrast, in Israel, pneumonias caused by Mycoplasma were significantly more common in the spring compared to the winter months (p=0.054). It should be noted that the findings of LIEBERMAN et al. [1] are based on an observation period of only 1 yr whilst our results are based on 9 consecutive yrs. It is, therefore, possible, as LIEBERMAN et al. [1] pointed out, that a single outbreak in the spring of 1992 might explain the high incidence of Mycoplasma pneumonias during this period. In this context it is also relevant that in our material there was a significant difference in the incidence of Mycoplasma infections between years with outbreaks every 3–4 yrs, as reported by others [5, 6]. It should be noted that the selection of our patients and the group of patients reported by LIEBERMAN et al. [1] is different. Since the selection of our patients is solely based on serology it is not certain that all of our Mycoplasma patients had pneumonia. However, it must be considered highly unlikely that pneumonias caused by Mycoplasma have a different seasonal predilection than other Mycoplasma infections. Figure 1 also shows the relationship between Mycoplasma infections and temperature during the observation periods. According to our judgement, LIEBERMAN et al. [1] concluded correctly from their data that meteorological factors, such as temperature and humidity, could CORRESPONDENCE