Table 3.

Effect of parental BCG scarring on mortality between birth and 42 days for children with scar information available for both parents.

	Mortality rate [Deaths/1000 Person days] (n)		MRR (CI) (Paternal scar/No paternal scar)	Adjusted MRR (CI) (Paternal scar/No paternal scar)b
	Paternal BCG scara	No paternal BCG scara
Maternal BCG scara	0.32 [1/3.2] (87)	2.8 [4/1.4] (42)	0.12 (0.01–1.03)	0.13 (0.01–1.18)e
No maternal BCG scara	6.2 [3/0.5] (16)	2.6 [2/0.8] (23)	2.38 (0.41–13.74)	2.80 (0.49–16.05)f
MRR (CI) (Maternal scar/No maternal scar)	0.06 (0.01–0.50)	1.09 (0.20–5.88)
Adjusted MRR (CI) (Maternal scar/No maternal scar)b	0.04 (0.00–0.32)c	1.14 (0.21–6.16)d		0.11 (0.01–0.87)g

^aPaternal scar data is missing for 291 infants of whom we have maternal scar data, maternal scar data is missing for 49 of whom we have paternal scar data, and two infants were included after 42 days, leaving 168 infants with information of both parents.

^bSeparate Cox proportional hazards models were performed for maternal and paternal scar status, which were adjusted by stabilized Inverse Probability of Treatment Weighting based on the variables maternal schooling, electricity, indoor toilet and age of the child at inclusion.

^cProportional hazards = 0.10.

^dProportional hazards = 0.58.

^eProportional hazards = 0.79.

^fProportional hazards = 0.12.

^gaMRR estimate for children with both parents having a scar compared to children with either one or none of their parents having a scar, proportional hazards = 0.41. Abbreviations: BCG, Bacillus Calmette-Guérin; CI, confidence interval; MRR, mortality rate ratio.