Table 1. Estimates of the secondary attack rate of tuberculosis in 1918 household contacts in Kampala, Uganda, 1995–2004.

Characteristic	Category	No. at Risk	No. Positive Culture Cases	No. without RFLP	No. RFLP Matched Isolates	Estimated No. Matched Isolates†	SAR - Tuberculosis (%)	95% CI
Overall		1918	76	15	46	57.3	3.0	2.2, 3.8
Age (y)	≤5	508	28	3	23	25.8	5.1	3.2, 7.0
	6–15	691	7	3	3	5.3	0.8	0.1, 1.4
	16–25	364	16	3	8	9.8	2.7	1.0, 4.4
	26–45	283	22	5	11	14.2	5.0	2.5, 7.6
	≥46	72	3	1	1	1.5	2.1	0, 5.4
	>5	1410	48	12	23	30.7	2.2	1.4, 2.9
HIV Status	HIV+	201	30	8	13	17.7	8.8	4.9, 12.7
	HIV−	1455	44	7	31	36.9	2.5	1.7, 3.3
BCG Vaccine	Yes	1349	46	6	32	36.8	2.7	1.9, 3.6
	No	499	27	7	13	17.6	3.5	1.9, 5.1

**Co-prevalent cases with the same finger print pattern as the index case. Since 15 cases did not have RLFP results, this number is estimated using the observed proportion (see methods) of RLFP matches. 46/61 observed matches; thus, 46/61*76 culture confirmed cases = 57.3 = 57.

^†The total number of cases with matched RFLP patterns is the number of isolates with observed matches plus expected number of matches from isolates grown in culture but not analyzed with RFLP. Expected number of matches was estimated as the product of the observed proportion of matches and the number of pairs without RFLP results plus observed matches.

*HIV serostatus was not available in 262 (13.7%) of contacts. HIV serostatus was not measured in community control households; the general secondary attack rate for infection was therefore used to estimate risk of disease after household infection.

^‡Vaccination status missing or uncertain in 70 household contacts and 4 community members.