Cutaneous tuberculosis (CTB) may be over-diagnosed due to imprecise diagnostic criteria or overlooked where mycobacterial investigations are negative. We evaluated the distinction between multibacillary and paucibacillary forms of CTB, as well as drug resistance and cure rates according to the results of mycobacterial investigations.

We included retrospectively all patients diagnosed with CTB from 1995 to 2018 in two hospitals in Paris. Clinical forms were classified according to dermatological descriptions, into multibacillary (e.g. gumma, scrofuloderma, orificial TB) and paucibacillary forms (lupus vulgaris, verrucous tuberculosis, papulonecrotic tuberculids, nodular panniculitis). A distinction was made between microbiologically confirmed CTB and presumed CTB forms, which were treated presumptively. Cure was defined as the complete resolution of CTB in patients who completed anti-tuberculosis treatment.

Among the 124 patients with CTB, the most common forms were nodular panniculitis (30.6%), scrofuloderma (22.6%), gumma (18.6%), and lupus vulgaris (12.1%). Tuberculosis was confirmed in 78 patients (62.9%), among whom 13 (16.7%) exhibited resistance to anti-tuberculous drugs, and 46 were presumptively treated. Mycobacterial investigations were significantly more frequently positive for multibacillary (88.2%) than for paucibacillary CTB (39.3%) (p < 10⁻⁶). Patients with mycobacterial evidence of CTB exhibited significantly better cure rates than patients without (96.7% vs. 66.7%, p < 10⁻⁴), particularly among those with nodular panniculitis (100% vs. 63.0%, p < 10⁻³).

The distinction between paucibacillary and multibacillary CTB is relevant. Resistant strains may be isolated. Antituberculosis drugs should be prescribed with caution in cases of panniculitis in the absence of evidence of mycobacterial infection.