BACKGROUND: The effects of environmental factors and ambient concentrations of grass pollen on allergic asthma are yet to be established. OBJECTIVE: We sought to estimate the independent effects of grass pollen concentrations in the air over Melbourne on asthma hospital admissions for the 1992-1993 pollen season. METHODS: Daily grass pollen concentrations were monitored over a 24-h period at three stations in Melbourne. The outcome variable was defined as all-age asthma hospital admissions with ICD9-493 codes. The ambient air pollutants were average daily measures of ozone, nitrogen dioxide and sulphur dioxide, and the airborne particle index representing fine particulate pollution. Semi-parametric Poisson regression models were used to estimate these effects, adjusted for air temperature, humidity, wind speed, rainfall, day-of-the-week effects and seasonal variation. RESULTS: Grass pollen was a strong independent non-linear predictor of asthma hospital admissions in a multi-pollutant model (P=0.01). Our data suggest that grass pollen had an increasing effect on asthma hospital admissions up to a threshold of 30 grains/m3, and that the effect remains stable thereafter. CONCLUSION: Our findings suggest that grass pollen levels influence asthma hospital admissions. High grass pollen days, currently defined as more than 50 grains/m3, are days when most sensitive individuals will experience allergic symptoms. However, some asthmatic patients may be at a significant risk even when airborne grass pollen levels are below this level. Patients with pollen allergies and asthma would be advised to take additional preventive medication at lower ambient concentrations.