Table 3.

HRs (95% CIs) for Baseline Thyroid Dysfunction and Risk of MI or Stroke, the ARIC Study (1990–2014)

	HR (95% CI)
	Model 1 (Crude)	Model 2 (Confounders)	Model 3 (Confounders + Mediators)
MI (n = 1102 events)
Log-transformed T3	0.90 (0.72–1.12)	0.78a (0.62–0.98)	0.88 (0.70–1.11)
N	11,356	11,235	11,206
Thyroid group
N	11,356	11,235	11,206
Moderate or severe chemical hypothyroidism	1.36 (0.96–1.92)	1.43a (1.01–2.03)	1.29 (0.91–1.84)
Mild chemical hypothyroidism	1.02 (0.77–1.34)	1.09 (0.83–1.45)	1.09 (0.83–1.45)
Euthyroidism	1 (Ref.)	1 (Ref.)	1 (Ref.)
Mild chemical hyperthyroidism	1.15 (0.84–1.57)	1.18 (0.85–1.62)	1.14 (0.83–1.58)
Moderate or severe chemical hyperthyroidism	0.89 (0.55–1.43)	1.07 (0.66–1.73)	1.06 (0.64–1.73)
Stroke (n = 838 events)
Log-transformed T3	0.77a (0.60–0.99)	0.81 (0.62–1.05)	0.80 (0.61–1.06)
N	11,356	11,235	11,062
Thyroid group
N	11,356	11,235	11,062
Moderate or severe chemical hypothyroidism	0.96 (0.60–1.53)	1.03 (0.64–1.64)	1.06 (0.66–1.70)
Mild chemical hypothyroidism	0.96 (0.69–1.32)	1.04 (0.75–1.44)	1.08 (0.77–1.50)
Euthyroidism	1 (Ref.)	1 (Ref.)	1 (Ref.)
Mild chemical hyperthyroidism	1.29 (0.92–1.82)	1.23 (0.87–1.74)	1.14 (0.80–1.63)
Moderate or severe chemical hyperthyroidism	0.97 (0.57–1.64)	1.10 (0.64–1.87)	1.06 (0.61–1.85)

T3 was log-transformed in base 2; therefore, estimates are shown for a twofold increase in T3. Model 1 was unadjusted, Model 2 was adjusted for confounders (age, sex, race-center, HDL-C, alcohol use, tobacco use), and Model 3 was adjusted for Model 2 variables plus mediators (systolic blood pressure, diastolic blood pressure, antihypertensive therapy, non-HDL-C, lipid-lowering therapy, diabetes, body mass index, and heart rate). For the model with stroke as the end point, we also included baseline or incident atrial fibrillation as a mediator.

^aP < 0.05.