Development of the gut microbiota during early life in premature and term infants

Table 2 Models relating microbial and clinical data to the odds of parent-reported wheezing episodes and eczema

		Clinical factor	Odds ratio (95% CI); change in odds ratio per unit of factor	P value	Clinical factor value at:
		Clinical factor		P value	Lower quartile	Median	Upper quartile
Modelling of odds ratio for wheezing episodes by 2 years of age	Univariate Models:	# Courses antibiotics since 6 weeks	1.57 (1.19, 2.21)	0.004	0	1	3
		Bacteroidetes reads (% of community)	0.99 (0.99, 1.00)	0.033	0	7	27
		Firmicutes reads (% of community)	1.01 (1.00, 1.01)	0.002	44	62	72
		Subdoligranulum OTU reads	1.05 (1.02, 1.08)	0.002	0	1	4
	Multivariate Model:	# Courses antibiotics since 6 weeks	1.62 (1.16, 2.42)	0.010	0	1	3
		Firmicutes reads (% of community)	1.01 (1.00, 1.01)	0.025	44	62	72
		Subdoligranulum OTU reads (% of community)	1.03 (1.00, 1.07)	0.041	0	1	4
Modelling of odds ratios for eczema by 2 years of age	Univariate Models:	Gestation at birth (weeks)	1.15 (1.02, 1.33)	0.030	28.1	34.4	40.3
	Univariate Models:	Firmicutes reads (% of community)	1.01 (1.00, 1.01)	0.007	44	62	72
	Multivariate Model:	Gestation at birth (weeks)	1.17 (1.03, 1.37)	0.030	28.1	34.4	40.3
	Multivariate Model:	Firmicutes reads (% of community)	1.01 (1.00, 1.01)	0.007	44	62	72

Univariate models included the investigated variable, infant gestational ages and day of sampling and factors have been reported if significant at the 20% level after MHC (with OTUs, phyla and clinical factors considered separate groups for correction). A multivariate model was used to identify dominant factors, with iterative removal of the least influential factors until only significant factors remained. Odds ratios (bold) and 95% confidence intervals have been calculated for each model; values provided indicate the proportional change in odds per unit of the variable

ISSN: 1757-4749