In subjects with normal lung mechanics, inspiratory muscle strength can be reliably and easily assessed by the sniff nasal inspiratory pressure (SNIP), which is the pressure measured in an occluded nostril during a maximal sniff performed through the contralateral nostril. The aim of this study was to assess the validity of the SNIP in patients with chronic obstructive pulmonary disease (COPD), where pressure transmission from alveoli to upper airways is likely to be dampened. Twenty eight patients with COPD were studied (mean forced expiratory volume in one second (FEV1) = 36% of predicted). The SNIP and the sniff oesophageal pressure (sniff Poes) were measured simultaneously during maximal sniffs, and were compared to the maximal inspiratory pressure obtained against an occlusion (MIP). All measurements were performed from functional residual capacity in the sitting position. The ratio SNIP/sniff Poes was 0.80, and did not correlate with the degree of airflow limitation. The ratio MIP/sniff Poes was 0.87, and the ratio SNIP/MIP was 0.97. Inspiratory muscle weakness, as defined by a low sniff Poes, was present in 17 of the 28 patients. A false diagnosis of weakness was made in eight patients when MIP was considered alone, in four when SNIP was considered alone, and in only three patients when MIP and SNIP were combined. We conclude that both the sniff nasal inspiratory pressure and the maximal inspiratory pressure moderately underestimate sniff oesophageal pressure in chronic obstructive pulmonary disease. Although suboptimal in this condition, the sniff nasal inspiratory pressure appears useful to complement the maximal inspiratory pressure for assessing inspiratory muscle strength in patients with chronic obstructive pulmonary disease.