Table 2.

Trends in the Prevalence of Extreme Obesity From 1998–2003 and From 2003–2010 Among US Children Aged 2 Through 4 Years Living in Low-Income Families^a

Characteristic	Prevalence, %			OR (95% CI) Per Year Change
				Unadjustede		Adjustedf
	1998b	2003c	2010d	1998–2003	2003–2010	1998–2003	2003–2010
Overall	1.75	2.22	2.07	1.048 (1.046–1.050)	0.988 (0.987–0.990)	1.047 (1.045–1.049)	0.983 (0.981–0.984)
Age, y
2	1.18	1.39	1.14	1.028 (1.023–1.032)	0.971 (0.968–0.973)	1.021 (1.017–1.026)	0.965 (0.962–0.967)
3	1.76	2.28	2.08	1.052 (1.049–1.056)	0.988 (0.986–0.990)	1.048 (1.044–1.051)	0.982 (0.979–0.984)
4	2.40	3.21	3.24	1.062 (1.058–1.065)	0.998 (0.996–1.000)	1.059 (1.056–1.063)	0.993 (0.991–0.995)
Sex
Boys	1.69	2.17	2.01	1.053 (1.050–1.057)	0.986 (0.985–0.988)	1.052 (1.049–1.055)	0.980 (0.979–0.982)
Girls	1.81	2.27	2.14	1.042 (1.039–1.045)	0.991 (0.989–0.992)	1.042 (1.039–1.045)	0.985 (0.983–0.987)
Race/ethnicityg
Non-Hispanic white	1.17	1.64	1.60	1.071 (1.067–1.075)	0.987 (0.985–0.990)	1.072 (1.068–1.076)	0.987 (0.985–0.989)
Non-Hispanic black	1.40	1.73	1.50	1.045 (1.040–1.050)	0.979 (0.976–0.983)	1.047 (1.042–1.052)	0.980 (0.976–0.983)
Hispanic	2.85	3.30	2.92	1.028 (1.025–1.031)	0.982 (0.980–0.984)	1.032 (1.029–1.036)	0.982 (0.980–0.984)
American Indian/Alaska Native	2.24	2.95	2.98	1.046 (1.025–1.068)	1.003 (0.989–1.018)	1.048 (1.027–1.070)	1.003 (0.989–1.018)
Asian/Pacific Islander	1.99	1.97	1.58	0.996 (0.985–1.007)	0.968 (0.963–0.973)	1.002 (0.991–1.013)	0.966 (0.960–0.971)

Abbreviation: OR, odds ratio.

^aExtreme obesity defined as body mass index for age and sex in the 120% or greater of the 95th percentile according to 2000 Centers for Disease Control and Prevention growth charts.

^bBonferroni adjustments were used during the multiple subgroup comparisons with t tests. Results of all subgroup comparisons are all statistically significant at P < .005, except for the prevalence difference between American Indian/Alaska Native and Asian/Pacific Islander children.

^cBonferroni adjustments were used during the multiple subgroup comparisons with t tests. Results of all subgroup comparisons are all statistically significant at P < .005, except the prevalence difference between Hispanic and American Indian/Alaska Native children.

^dBonferroni adjustments were used during the multiple subgroup comparisons with t tests. Results of all subgroup comparisons are all statistically significant at P < .005, except the prevalence differences between Asian/Pacific Islander and non-Hispanic white or black children and between Hispanic and American Indian/Alaska Native children.

^eUnadjusted odds of being extremely obese for 1 year’s increase in time, calculated from piecewise logistic regression.

^fAdjusted odds of being extremely obese for 1 year’s increase in time, calculated from piecewise logistic regression controlling for age, sex, and race/ethnicity.

^gTrends are presented because they are important for identifying health disparities.