Metabolic Health Has Greater Impact on Diabetes than Simple Overweight/Obesity in Mexican Americans

Table 2

Diabetes by overweight/obese type and metabolic health status.

Diabetes	Metabolically healthy		Metabolically unhealthy		P value
	Normal weight	Overweight/obese	Normal weight	Overweight/obese
	()	()	()	()
Primary analysis
Frequency
Yes [, (%)]	57 (12.00)	349 (21.89)	26 (36.11)	446 (40.33)	<0.0001^1
No [, (%)]	401 (84.42)	1208 (75.78)	42 (58.33)	641 (57.96)
Weighted OR (95% CI)
Unadjusted model	Reference	2.30 (1.47, 3.60)	5.20 (2.41, 11.19)	6.23 (3.94, 9.85)	<0.0001^2
Multivariable adjusted model 1^3	Reference	2.25 (1.34, 3.79)	3.78 (1.57, 9.09)	5.39 (3.16, 9.20)	<0.0001^2
Multivariable adjusted model 2^4	Reference	2.14 (1.07, 4.28)	3.18 (1.02, 9.92)	5.01 (2.43, 10.34)	<0.0001^2
Sensitivity analysis^5
Frequency
Yes [, (%)]	41 (9.53)	145 (12.49)	42 (35.90)	650 (42.24)	<0.0001^1
No [, (%)]	375 (87.21)	982 (84.58)	68 (58.12)	867 (56.34)
Weighted OR (95% CI)
Unadjusted model	Reference	1.48 (0.87, 2.54)	4.25 (1.74, 10.37)	7.81 (4.56, 13.37)	<0.0001^2
Multivariable adjusted model 1^3	Reference	1.47 (0.83, 2.61)	2.93 (1.04, 8.23)	6.25 (3.54, 11.02)	<0.0001^2
Multivariable adjusted model 2^4	Reference	1.36 (0.59, 3.13)	3.23 (1.08, 10.01)	6.57 (2.81, 15.36)	<0.0001^2

 ¹ approximation of Rao-Scott design adjusted chi-square test P value.
²P values from Wald chi-square test for the effect of overweight/obese phenotype.
³Adjusted for age at enrollment. Other covariates were not significant and not included in the final model. The models were adjusted for the probability of sampling using weights taking into consideration clustering effects arising from the same census block and household.
⁴Adjusted for age at enrollment and family history of diabetes. Restricted to the participants who had data for family history of diabetes (,234, 68%).
⁵The definition of metabolically health included glucose component.