Table 3. Relationship between G6PD deficiency and gender, age, ethnic group, hematocrit, oxygen saturation, heart rate, tea-colored urine and scleral icterus.

N = 657	Univariate Models:	Multivariate Model:
	Unadjusted Odds Ratio	Adjusted Odd Ratio
Predictor	Without RE*: (p-value)	With RE, all predictors (p-value)	Without RE, all predictors (p-value)
Gender	3.54 (<0.0001)	4.26 (<0.0001)	3.59 (<0.0001)
Age, ordinal	1.04 (0.3202)	1.06 (0.3065)	1.05 (0.2790)
Ethnic group
Igbo	0.369 (0.0370)	0.309 (0.0551)	0.382 (0.0500)
Igede	0.860 (0.7414)	0.888 (0.8283)	0.862 (0.7528)
Tiv	1.72 (0.6404)	1.05 (0.9747)	1.03 (0.9789)
Hematocrit	0.980 (0.4084)	0.979 (0.4835)	0.982 (0.4868)
Oxygen sat.	0.987 (0.2141)	0.983 (0.1882)	0.986 (0.1913)
Heart rate	0.998 (0.7070)	1.00 (0.5390)	1.00 (0.5965)
Tea-colored urine	0.727 (0.6093)	0.511 (0.3683)	0.540 (0.3521)
Scleral icterus	1.97 (0.0395)	2.26 (0.0563)	2.12 (0.0351)

^*RE is the random effect of family.

Note: The backward elimination results are not shown in this table.

The odds of being G6PD deficient were 3.6 times higher in males than in females (p<0.0001). The odds of being G6PD deficient were 0.38 times as high in Igbo children compared to Yoruba children (p = 0.0500). The odds for Igede and Tiv children were not significantly different from Yoruba children (p = 0.7528 and 0.9789 respectively). The odds of being G6PD deficient were 2.1 times higher in children with scleral icterus than those without (p = 0.0351). The magnitude of the scleral icterus effect appeared to be lower if these children also reported tea-colored urine, but the number of children reporting both conditions (N = 10) is too small to be certain.