Table 3.

Cox regression analysis^a of time to multiple clinical malaria episodes^b between days 90 and 240

	AMA1 vaccine group		Control group
	No primary outcome episodes before day 90 (n = 66)		No primary outcome episodes before day 90 (n = 22)
Predictor variable	Hazard ratioc	P	Hazard ratio	P
log10 (IgG1 day 90)	1.17 [0.47, 2.93]	0.74	1.14 [0.65, 2.00]	0.66
log10 (IgG2 day 90)	0.92 [0.5, 1.70]	0.79	1.27 [0.43, 3.72]	0.66
log10 (IgG3 day 90)	1.89 [0.99, 3.61]	0.053	0.93 [0.35, 2.43]	0.87
log10 (IgG4 day 90)	1.06 [0.69, 1.63]	0.79	8.45 [0.61, 116.86]	0.11
log10 (cytophilic ratiod day 90)	0.83 [0.40, 1.70]	0.61	1.05 [0.44, 2.53]	0.91

^aGap time models with common regression coefficients using Firth’s penalized likelihood

^bClinical malaria episode defined using the primary outcome definition, i.e., an episode of malaria with axillary temperature of ≥ 37.5 °C and asexual P. falciparum density of ≥ 2500 parasites/mm³ of blood

^cHazard ratio and 95% confidence intervals for risk of multiple clinical malaria episodes after day 90

^dCytophilic ratio is the sum of IgG1 and IgG2 titers divided by the sum of the IgG3 and IgG4 titers