Table 2. Risk estimates for diabetes mellitus-associated thyroid cancer overall and within subgroups.
| N of studies | N of thyroid cancer cases | Summary RR (95% CI) a , c | p-heterogeneity | ||
| All studies | 13 | 4,051 | 1.34 (1.11–1.63) | <0.001 | |
| Sensitivity analysis b | 9 | 3,566 | 1.18 (1.08–1.28) | 0.84 | |
| Study design b | Cohort studies | 7 | 3,143 | 1.18 (1.09–1.29) | 0.76 |
| Case-control studies | 2 | 423 | 0.91 (0.51–1.64) | 0.97 | |
| Geographical area b | High incidence regions | 7 | 3,446 | 1.18 (1.09–1.29) | 0.76 |
| Low incidence regions | 2 | 120 | 0.98 (0.66–1.47) | 0.98 | |
| Study quality b | Score ≥6 | 7 | 3224 | 1.18 (1.08–1.28) | 0.74 |
| Score <6 | 2 | 322 | 1.18 (0.76–1.81) | 0.46 |
a
All summary ORs/RRs (95% CIs) were calculated by the random-effect model.
b
We excluded three studies using the risk estimates with SIRs ([14] and [32]) and the different definition of diabetes ([33]was included with IFG and IGT and [34] used quintile of glucose level).
c
No publication bias by Egger and Begg test (p>0.05).