OBJECTIVES--To investigate whether in vivo capillary microscopy of the lower lip mucosa can be used to assess microvascular disease in systemic sclerosis. METHODS--Thirteen patients with systemic sclerosis and 11 healthy control subjects were studied by conventional nailfold capillary microscopy and labial capillaroscopy. The following parameters were analysed: loop length; loop width (maximum distance between the arteriolar and venular limbs); loop density (number of capillaries/mm2); venular plexus visibility; megacapillaries; and the architectural arrangement of the capillary network. RESULTS--A typical 'scleroderma pattern' at the nailfold was observed in 12 of 13 (92%) patients with systemic sclerosis. Labial capillaroscopy showed a different morphological pattern of microangiopathy. A diffuse architectural derangement of the capillary network was the most striking abnormality in 12 (92%) patients. Labial capillaries in the patients with systemic sclerosis were shorter (mean (SD) loop length 133 (32.2) microns) than in healthy controls (211 (48.4) microns) and showed an increased loop width (41.7 (13.1) v 27.6 (5.5) microns in controls. The loop density was 10.5 (4.6) capillaries/mm2 in patients with systemic sclerosis and 9 (1.7) capillaries/mm2 in controls. Labial capillaroscopy in patients with systemic sclerosis did not provide definite evidence of enlarged capillaries or avascular areas, or both, even where such abnormalities were clearly evident at the nailfold. CONCLUSIONS--This study shows that labial capillary microscopy is a simple, non-invasive technique which allows a careful morphological assessment of the mucosal microcirculation. Labial capillaroscopy in patients with systemic sclerosis showed significant microvascular changes with respect to the controls. The results of labial and nailfold capillaroscopy are not superimposable, even if some common findings, such as architectural derangement, are present.