Inhaled corticosteroid use was also associated with a 53 per cent increased risk of death within 30 days of hospitalisation.
The findings cast doubt on NICE guidance for patients to be given inhaled steroids if they have had one exacerbation in the past year despite treatment and have an FEV1 below 50 per cent.
The study included 23,942 patients with COPD who were hospitalised for pneumonia and 95,768 controls also with COPD.
Current inhaled steroid use increased the risk of hospitalisation for pneumonia by 70 per cent, even when severity of COPD was accounted for.
Those receiving a high steroid dose - fluticasone 1,000µg daily - were 125 per cent more likely to be admitted to hospital with pneumonia than those not taking an inhaled corticosteroid.
However, the association between inhaled corticosteroids and pneumonia seems to dissipate within six months of discontinuing the medication.
For those last prescribed an inhaled steroid nine to 12 months before the index date, the increased risk of hospitalisation for pneumonia was 20 per cent.
But steroid therapy did not result in a longer hospital stay or increased all-cause mortality in those who developed severe pneumonia, suggesting response to pneumonia treatment is not affected.
Professor Neil Barnes, from the London Chest Hospital, said: 'Inhaled corticosteroids reduce the number of exacerbations of COPD, but if you break through and have an exacerbation, it's more likely to be pneumonia.'
However, he said the findings should not affect current practice because, while not treating COPD with inhaled corticosteroids may slightly reduce cases of pneumonia, patients would have more chest infections from increased exacerbations.
Am J Respir Crit Care Med 2003; 176: 162-6
Comment below and tell us what you think