BACKGROUND: Suicide is a major public health problem and depression is among the most important risk factors for suicide. Treatment of depression might prevent suicide. To study this hypothesis further we conducted an ecological study. METHODS: An ecological study using sales data for antidepressants and numbers of suicides in Norway and Norwegian counties 1980-2004 was performed. Data on alcohol consumption and unemployment rates were registered and taken into account. Data were analyzed using Cochrane-Orcutt time series for the country as a whole. The county specific data were analyzed with a random coefficient model with county as subject and intercept and time (slope) as random variables using an unstructured covariance matrix. RESULTS: Sales of non-tricyclic antidepressants (non-TCAs) and suicide were clearly negatively related, even when controlling for alcohol and unemployment (adjusted r(2): 0.57). There was an effect modification between time and level of sales of non-TCAs. Studying the relationship between the sales of non-TCAs and the suicide rate, we found that it was significant and stronger for the low sales figures, but non-existent for the high sales figures. LIMITATIONS: Ecological studies cannot infer causality. CONCLUSIONS: The fall in suicide rates in Norway and its counties was related to the increased sales of non-TCAs. The effect was mostly a result of a sales increase in the lower sales segment, indicating that a change from the more toxic TCAs, or heightened awareness of depression and its treatment, could explain the relationship found between sales of newer antidepressants and a decrease in suicide rate.