Nonlinear short-term effects of airborne Poaceae levels on hay fever symptoms

More sophisticated and rigorous statistical methods such as generalized additive and generalized estimated equations models, which check the shapes of concentration-responses and control for confounders, are however available for studying factors influencing health in air pollution studies.1 To our knowledge, our panel study, performed on 106 volunteers, is one of the first to use these methods to explore the relationship between symptoms of hay fever and pollen levels. The higher level of nasal and ocular symptoms, without any trigger threshold, observed in the whole population at the beginning of pollination may be explained by a priming effect, as described by Connell.6 This means that during natural exposure, decreasing amounts of pollen are required to elicit symptoms as the season progresses, and the severity of allergic response is therefore increased.\n Statistical methods For each nasal, ocular, and respiratory symptom, daily score was transformed in a dichotomous response variable: "no symptoms" for no symptoms or mild symptoms (score 1).