Skip to main content
. 2016 Aug 25;6:31975. doi: 10.1038/srep31975

Table 2. Stratified analysis of the association between TSH and dementia.

Subgroup N Risk estimate
Heterogeneity
Pooled RR (95% CI) P1 I2 P2
TSH: low category vs. middle category 10 1.6 (1.27, 2.00) <0.001 0.00 0.58
 Study type
  Cohort 7 1.5 (1.17, 1.92) 0.001 0.00 0.49
  Case-control 3 2.25 (1.28, 3.96) 0.005 0.00 0.78
 Data type
  TSH below vs. within the normal range 6 1.77 (1.31, 2.39) <0.001 0.00 0.76
  Lower vs. middle tertile of TSH within the normal range 4 1.39 (0.98, 1.97) 0.062 0.23 0.27
TSH: high category vs. middle category 9 0.99 (0.76, 1.29) 0.923 0.17 0.29
 Study type
  Cohort 6 1.14 (0.85, 1.54) 0.385 0.00 0.65
  Case-control 3 0.54 (0.30, 1.00) 0.048 0.00 0.44
 Data type
  TSH above vs. within the normal range 6 0.91 (0.61, 1.34) 0.619 0.00 0.77
  Upper vs. middle tertile of TSH within the normal range 3 1.07 (0.74, 1.54) 0.734 0.70 0.03
Gender: male
 Low category vs. middle category of TSH 3 1.28 (0.70, 2.32) 0.419 0.00 0.62
 High category vs. middle category of TSH 3 0.92 (0.60, 1.43) 0.722 0.00 0.57

Key: N, number of studies. P1 is an evaluation of the statistical significance level of the risk estimate, while P2 is an evaluation of the heterogeneity among included studies. Subgroup analyses were performed by study design (cohort or case-control) and data types of TSH levels. Three studies evaluating the relationship between TSH and dementia in men were analyzed separately.

HHS Vulnerability Disclosure