Table 3.
Meta-analyses on circulating IL-1β levels in type 1 diabetes mellitus.
| Pooled Data | Heterogeneity | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Meta-Analyses | No. of Studies | No. of Patients | Stat. Model | Wt | SMD (95% CI) | p-Value | Phet |
I2 (%) |
Supplementary Materials a |
| Determination by immunoassays | |||||||||
| All b | 20 | 3490 | REM | D-L | 2.45 (1.73 to 3.17) | <0.001 | <0.001 | 98.6 | —— |
| Subgroup analysis by geographical area c | Figure S1, p. 11 | ||||||||
| Africa | 3 | 403 | REM | D-L | 10.41 (2.58 to 18.23) | 0.01 | <0.001 | 99.5 | |
| Asia | 5 | 885 | REM | D-L | 2.61 (0.56 to 4.66) | 0.01 | <0.001 | 99.0 | |
| Europe | 9 | 1875 | REM | D-L | 1.04 (0.49 to 1.59) | <0.001 | <0.001 | 95.0 | |
| North America | 1 | 38 | —— | —— | 0.35 (−0.30 to 0.99) | 0.29 | —— | —— | |
| South America | 2 | 289 | REM | D-L | −0.29 (−2.37 to 1.79) | 0.78 | < 0.001 | 97.4 | |
| Subgroup analysis by age c | Figure S2, p. 12 | ||||||||
| <18 years old | 14 | 2870 | REM | D-L | 2.81 (1.88 to 3.74) | <0.001 | <0.001 | 98.9 | |
| >18 years old | 6 | 620 | REM | D-L | 1.56 (0.48 to 2.65) | 0.002 | <0.001 | 96.5 | |
| Subgroup analysis by HbAc1 levels in patients <18 years old c,d | Figure S3, p. 13 | ||||||||
| <7 | 2 | 79 | REM | D-L | −0.04 (−2.67 to 2.58) | 0.97 | <0.001 | 96.2 | |
| >7 | 8 | 1138 | REM | D-L | 5.43 (3.31 to 7.56) | 0.001 | <0.001 | 99.1 | |
| Subgroup analysis by age matching c | Figure S4, p. 14 | ||||||||
| Matched | 15 | 3172 | REM | D-L | 3.06 (2.19 to 3.94) | <0.001 | <0.001 | 98.8 | |
| Unmatched | 5 | 318 | REM | D-L | 0.90 (−0.18 to 1.97) | 0.10 | <0.001 | 94.4 | |
| Subgroup analysis by sex matching c | Figure S5, p. 15 | ||||||||
| Matched | 11 | 2379 | REM | D-L | 0.55 (0.19 to 0.91) | 0.003 | <0.001 | 92.9 | |
| Unmatched | 3 | 224 | REM | D-L | 0.88 (−1.15 to 2.90) | 0.40 | <0.001 | 97.5 | |
| NA | 6 | 887 | REM | D-L | 8.66 (5.37 to 11.96) | <0.001 | <0.001 | 98.9 | |
| Subgroup analysis by sample source c | Figure S6, p. 16 | ||||||||
| Serum | 15 | 3111 | REM | D-L | 2.73 (1.85 to 3.61) | <0.001 | <0.001 | 98.9 | |
| Plasma | 5 | 379 | REM | D-L | 1.34 (0.28 to 2.41) | 0.01 | <0.001 | 94.3 | |
| Subgroup analysis by type of analysis c | Figure S7, p. 17 | ||||||||
| ELISA | 16 | 2235 | REM | D-L | 3.29 (2.27 to 4.30) | <0.001 | <0.001 | 98.8 | |
| Immunoassay panel | 4 | 1255 | REM | D-L | 0.25 (−0.08 to 0.58) | 0.14 | 0.02 | 70.5 | |
| Subgroup analysis by study design c | Figure S8, p. 18 | ||||||||
| Case-control | 16 | 2447 | REM | D-L | 2.77 (2.00 to 3.55) | <0.001 | <0.001 | 98.1 | |
| Cohort | 1 | 398 | — | — | 0.03 (−0.164 to 0.23) | 0.74 | — | — | |
| Cross-sectional | 3 | 645 | REM | D-L | 1.39 (−1.56 to 4.34) | 0.36 | <0.001 | 99.3 | |
| Univariable meta-regression e | |||||||||
| Sex (% of T1DM males) | 17 | 2928 | Random-effects Meta-regression |
Coef = 0.011 (−0.619 to 0.641) |
0.97 | —— | —— | Figure S9, p. 19 | |
| Risk of bias (NOS score) | 20 | 3490 | Random-effects Meta-regression |
Coef = 0.195 (−3.209 to 3.598) |
0.91 | —— | —— | Figure S10. p. 20 | |
| Determination by qRT-PCR | |||||||||
| All b | 2 | 216 | REM | D-L | −0.66 (−3.02 to 1.71) | 0.59 | <0.001 | 97.1 | —— |
| Determination by Flow Citometry | |||||||||
| All b | 3 | 455 | REM | D-L | 1.40 (−0.19 to 3.00) | 0.08 | <0.001 | 91.8 | —— |
Abbreviations: Stat., statistical; Wt, method of weighting; SMD, standardized mean difference; CI, confidence intervals; REM, random-effects model; D-L, DerSimonian and Laird method; HbAc1, hemoglobin Ac1; T1DM, type 1 diabetes mellitus; NOS, Newcastle–Ottawa Scale; NA, not available. a More information in the Supplementary Materials; b meta-analyses; c subgroup meta-analyses; d the studies recruiting patients >18 years old or with missing data were excluded for this analysis; e effect of study covariates on circulating IL-1β levels among patients with T1DM compared with controls.