TABLE 3.

Multivariable logistic regression models for fluorosis^* of permanent maxillary incisors using AUC^† fluoride intakes (0.1 milligram per day), according to exposure age.

EXPOSURE VARIABLE, ACCORDING TO PARTICIPANT’S AGE IN MONTHS	ODDS RATIO (95% CI‡)	P VALUE
3- to 9-Month Model
Formula reconstituted from powder	1.09 (1.02, 1.15)	.008
Other beverages with added water§	1.75 (1.08, 2.82)	.03
Eliminated (P > .05): cow’s milk, water alone
16- to 36-Month Model
Water by itself	1.23 (1.04, 1.45)	.02
Dentifrice	1.11 (1.02, 1.21)	.03
Eliminated (P > .05): other beverages with added water,§ selected foods¶
Combined 3- to 9-Month and 16- to 36-Month Model
Formula reconstituted from powder (3–9 months)	1.10 (1.03, 1.17)	.005
Other beverages with added water (3–9 months)§	1.68 (1.02, 2.78)	.05
Dentifrice (16–36 months)	1.13 (1.03, 1.24)	.02
Eliminated (P > .05): water by itself (16–36 months)
Combined Categorical Model#
Formula reconstituted from powder (3–9 months)	1.62 (1.05, 2.51)	.03
Other beverages with added water (3–9 months)†	1.56 (1.01, 2.42)	.045
Dentifrice (16–36 months)	1.66 (1.07, 2.57)	.03

^*Fluorosis was defined as having definitive fluorosis on two or more maxillary incisors, and no flurorosis as having no definitive fluorosis on any of the maxillary incisors. Variable reduction was achieved using the best subset (using the Score statistic) with all retained variables being statistically significant (P < .05).

^†AUC: Area-under-the-curve estimates.

^‡CI: Confidence interval.

^§Includes nonformula beverages made from frozen concentrate or powder.

^¶Selected foods include infant foods with substantial amounts of added water (such as infant cereal, cooked cereal, soup, rice, pasta, gelatin).

^#Indicator variables for categorical model are upper quartiles of each fluoride intake component. The reference group contains the lower three quartiles (combined). There were no significant two-way interactions.