Incident Mycobacterium tuberculosis infection in household contacts of infectious tuberculosis patients in Brazil

Table 4 Multivariable analysis of factors associated with tuberculin skin test (TST) positivity, conversion or negativity in 838 household contacts in Vitória, Brazil

Variable	TST-positive at baseline (n = 523)	TST Converter (n = 62)	OR^a (95% CI)	P +	TST negative (n = 253)	OR^b (95% CI)*	P ++
Contact Factors
Age (years)	22 [12–41]	15 [10–39]	0.99 (0.97–1.00)	0.17	18 [8–34]	1.01 (0.99–1.03)	0.24
Female	292 (56)	41 (66)	1.58 (0.96–2.60)	0.07	140 (55)	1.70 (0.95–3.06)	0.07
BCG scar present	425 (81)	54 (87)	1.58 (0.96–2.60)	0.32	207 (82)	2.96 (1.01–8.67)	0.05
Other TB contact	127 (25)	6 (10)	0.36 (0.16—0.82)	0.01	33 (13)	-	-
IGRA negative	78 (19)	18 (31)	1.73 (0.98–3.07)	0.06	196 (84)	-	-
Index Factors
Age (years)	34 [23–44]	32 [22–39]	-	-	37 [27–42]	0.98 (0.94–1.01)	0.15
Cough duration (weeks)	8 [4–16]	8 [4–12]	0.96 (0.91–1.02)	0.16	8 [4–16]	0.97 (0.93–1.00)	0.08
VACS	8 [5–9]	9 [8–10]	-	-	5 [3–8.5]	1.20 (1.03–1.40)	0.02
Extent of disease on CxR¹	257 (50)	20 (32)	0.87 (0.40–1.94)	0.74	945 (38)	-	-
Presence of cavities	412 (80)	40 (65)	0.68 (0.22–2.05)	0.49	163 (65)	0.96 (0.40–2.29)	0.93
Environmental Factors
Own residence	425 (81)	60 (97)	4.87 (1.05–22.5)	0.04	238 (94)	-	-
Sleep in the same room	353 (67)	46 (74)	-	-	137 (54)	2.22 (0.91–5.36)	0.08
> 6 h / day of contact	376 (73)	46 (74)	-	-	156 (62)	1.37 (0.64–2.95)	0.42

Definition of abbreviations: BCG Bacille Calmette-Guérin vaccine, CI Confidence interval, CxR Chest x-ray, IGRA Interferon gamma release assay (Quantiferon Gold-In-Tube), OR Odds ratio, TB Tuberculosis, VACS Visual Analog Cough Scale
Values are median [Interquartile range] or number (percentage), unless otherwise specified
^aOdds ratio and 95% CI estimated for TST converters when compared to TST-positive at baseline. For numeric variables, odds ratio represent odds of being a TST converter per 1 unit of increase of the variable
^bOdds ratio and 95% CI estimated for TST converters when compared to TST-negative at 8–12 weekss. For numeric variables, odds ratio represent odds of being a TST converter per 1 unit of increase of the variable
+ Estimated using multivariate GEE to compare risks of being TST converters vs. TST-positive at baseline. Model is adjusted for other variables listed in the corresponding table column
++ Estimated using conditional multivariate GEE to compare TST converters vs. TST-negative at 8–12 weeks. Model is adjusted for other variables listed in the corresponding table column
OR estimated comparing far advanced disease versus normal/minimal disease on chest radiograph

ISSN: 1471-2334