Results
Of the 539, we excluded 25 patients without data of thyroid indices, 10 patients diagnosed with hyperthyroidism or hypothyroidism and 24 patients lost follow-up. Eventually, we included 480 ischemic stroke patients. During 90 days follow-up, 244(50.8%) patients have poor outcome. The baseline characteristics (demographic, clinical, and laboratory data) according to the functional outcome assessed by the mRS score were presented in Table 1. Compared to the patients with good outcome, those with poor outcome were older, lower hemoglobin, albumin, eGFR and higher White Blood Cell, NIHSS score. The proportions of female (p=0.001) , atrial fibrillation(p<0.001), coronary heart disease/heart failure(P=0.001) and recanalized therapy(P=0.005)were also significantly higher in the poor outcome group compared to the good outcome group. In addition, patients with poor outcome had lower level of TSH, FT3,but higher FT4 levels.
To explore the independent risk factors for 3-month poor outcome, we selected variables P < 0.1 in univariate analysis to perform multivariable logistic regression analysis. It revealed that lower FT3 level was an independent risk factor of poor outcome at 3 months (OR= 0.561; 95%CI 0.375-0.841, P=0.005). In addition, older age (OR = 1.023, 95% CI 1.001-1.047, P=0.044), higher NIHSS score (OR=1.286, 95%CI 1.213-1.363, P<0.001) and no recanalized therapy (OR = 3.527, 95% CI 1.808-6.880, P<0.001) were also independent risk factors of poor outcome. However, TSH and FT4 levels were not statistical after adjusting for other variables (See Table 2).
Patients were divided into quartile groups according to FT3 levels. The distribution of mRS scores at 3 months differed according to FT3 quartile (Q1, Q2, Q3, Q4) (Fig1) since higher mRS scores were observed with lowering FT3 levels. Further analyses were performed by constructing two models to explore the relationship between FT3 and poor outcome. In model 1 adjusted for gender, AF, CHD/HF, Hemoglobin, WBC, ALB, eGFR, TSH, TG, TC, FT4 and TOAST Classification, lower FT3 levels were associated with a worse functional outcome at 3 months (OR=4.63; 95% CI 2.33-9.02, p < 0.001 for tertile1 vs. tertile4; OR=2.42; 95% CI 1.35-4.35, p =0.003 for tertile2 vs. tertile4). In model 2, after further adjusting variables of age, NIHSS score and recanalized therapy on the basis of model 1, compared with those in the 1st quartile, FT3 levels in the 4th quartile were still significantly associated with poor outcome (OR=2.56; 95% CI 1.15-5.69, p =0.021) (See Table 3). Additionally, the ROC curve analysis (Fig2) revealed that serum level of FT3<3.69pmol/L in AIS patients was a powerful predictor of poor outcome (sensitivity 62.70%; specificity 72.03%; AUC 0.713).