Table 3.

Association between treatment approach and having a grade 3-4 adverse event, adjusted for potential confounders.

	Total	No grade 3–4 AE	Grade 3–4 AE
	N	N (%)	N (%)	HRA	95%CI	aHRA	95%CI
Total	158	124 (78.5%)	34 (21.5%)
TB Clinic
Niamey	83	71 (85.5%)	12 (14.5%)	1		1
Maradi	33	24 (72.7%)	9 (27.3%)	1.79	[0.75,4.28]	1.49	[0.62,3.58]
Tahoua	31	20 (64.5%)	11 (35.5%)	2.51*	[1.10,5.72]	2.29*	[1.01,5.21]
Zinder	11	9 (81.8%)	2 (18.2%)	1.18	[0.26,5.31]	0.91	[0.20,4.12]
Gender						NS
Male	122	96 (78.7%)	26 (21.3%)	1
Female	36	28 (77.8%)	8 (22.2%)	1.06	[0.48,2.34]
Age						NS
For every year increase in age	NA	NA	NA	1.02	[0.99,1.04]
HIV statusB						NS
Negative	147	119 (81.0%)	28 (19.0%)	1
Positive	9	5 (55.6%)	4 (44.4%)	2.35	[0.82,6.78]
Previous TB treatment						NS
No	38	29 (76.3%)	9 (23.7%)	1
Yes	120	95 (79.2%)	25 (20.8%)	0.91	[0.42,1.98]
BMI						NS
For every increase in BMI unit	NA	NA	NA	1.02	[0.92,1.13]
Approach
aSTR	78	70 (89.7%)	8 (10.3%)	1		1
oSTR	80	54 (67.5%)	26 (32.5%)	3.19**	[1.45,7.06]	3.04**	[1.36,6.80]

^A Frailty bivariable and multivariable Cox regression models, accounting for the effect of randomisation month (clusters were defined by month of RR-TB diagnosis).

^B 2 patients had HIV test results showing undetermined, category not shown. AE = adverse event; BMI = body mass index; aSTR = Niger treatment strategy with a second-line injectable drug (SLID)-containing short treatment regimen (STR), with linezolid replacing the SLID in case of any grade of ototoxicity on monthly audiometry; oSTR = all oral short treatment regimen with linezolid throughout the intensive phase; HR = hazard ratio; aHR = adjusted hazard ratio; NA = not applicable; NS = not significant;

^*p < 0.05;

^**p < 0.01.