North Atlantic weather oscillation and human infectious diseases in the Czech Republic, 1951-2003.

Longitudinal correlation between the North Atlantic Oscillation large-scale weather system (NAO) and the annual incidence rate of 14 viral, bacterial and protozoan national notifiable human diseases in the Czech Republic was examined. In simple correlation, cases of salmonellosis, erysipelas, infectious mononucleosis and toxoplasmosis were positively correlated with the winter NAO index, while hepatitis A and shigellosis were negatively correlated, and the other diseases tested (rubella, mumps, chickenpox, tick-borne encephalitis, Lyme borreliosis, leptospirosis, tularemia and scarlet fever) were uncorrelated with NAO. However, 8 of the 14 diseases also revealed a significant time trend, either increasing (infectious mononucleosis, salmonellosis, erysipelas, toxoplasmosis) or decreasing (hepatitis A, scarlet fever, leptospirosis, shigellosis) during the period. When the effect of NAO on incidence of the diseases was then controlled for calendar year using partial correlation analysis and detrended regression, only toxoplasmosis and infectious mononucleosis were found significantly positively correlated with the NAO when the index was lagged 1 or 2 years, and leptospirosis was correlated negatively with a lag of 2 years. Large-scale weather changes as described by NAO therefore do not seem to be a crucial factor in the fluctuation of annual incidence rate of the majority of tested infectious diseases in the Czech Republic, while other factors, especially social and public health circumstances, are obviously more important.