Table 1.
Principal characteristics of selected meta-analyses and the adjustment performed according to the "trim and fill method"
| Meta-analysis | Number of trials | Estimated number of missing studies | Model of Effect used | Original combined effect estimates | Adjusted effect estimates | Year of last trial included |
| Hofmeyr 18 | 12 | 0 | Fixed | 0.35 (0.22–0.56) | Unchanged | 1999 |
| Smaill 19 | 83 | 18 | Random | 0.39 (0.35–0.43) | 0.42 (0.37–0.47) | 2001 |
| Hopkins 14 | 14 | 7 | Fixed | 0.92 (0.70–1.23) | 0.57 (0.45–0.73)* | 1993 |
| Hopkins 20 | 11 | 3 | Fixed | 1.08 (0.74–1.58) | 0.96 (0.68–1.36) | 1996 |
| Smaill 21 | 13 | 0 | Fixed | 0.24 (0.19–0.32) | Unchanged | 1987 |
| Kenyon 22 | 13 | 1 | Fixed | 0.90 (0.74–1.10) | 0.91 (0.74–1.11) | 2001 |
| McDonald 23 | 13 | 2 | Fixed | 0.87 (0.74–1.02) | 0.92 (0.78–1.07) | 2003 |
| Abalos 24 | 22 | 1 | Fixed | 0.75 (0.47–1.19) | 0.73 (0.46–1.15) | 1995 |
| Knight 25 | 37 | 14 | Fixed | 0.87 (0.80–0.94) | 0.91 (0.84–0.99) | 1999 |
| King 26 | 11 | 1 | Fixed | 1.49 (0.67–3.34) | 1.35 (0.62–2.95) | 2002 |
| Atallah 27 | 11 | 5 | Random | 0.35 (0.20–0.60) | 0.50 (0.32–0.78) | 2001 |
| Martin-Hirsch 13 | 11 | 3 | Fixed | 0.28 (0.25–0.31) | 0.27 (0.25–0.30) | 1994 |
| Thacker 28 | 10 | 5 | Fixed | 1.39 (1.21–1.59) | 1.17 (1.04–1.33) | 1993 |
| Hodnett 15 | 15 | 5 | Fixed | 0.91 (0.83–0.99) | 0.94 (0.86–1.03)* | 2002 |
| Cheng 16 | 10 | 1 | Fixed | 0.63 (0.44–0.92) | 0.70 (0.49–1.00)* | 2002 |
| Crowley 29 | 19 | 3 | Fixed | 0.20 (0.06–0.70) | 0.25 (0.08–0.73) | 1992 |
| Oates-Whitehead 30 | 14 | 0 | Fixed | 1.05 (0.83–1.34) | Unchanged | 2002 |
| Hodnett 31 | 13 | 2 | Fixed | 0.99 (0.90–1.09) | 0.98 (0.89–1.08) | 2001 |
| Carroli 32 | 10 | 2 | Fixed | 0.81 (0.70–0.93) | 0.83 (0.72–0.95) | 1998 |
| Johanson 33 | 12 | 2 | Fixed | 0.59 (0.51–0.68) | 0.60 (0.52–0.69) | 1996 |
| Hofmeyr 34 | 25 | 7 | Random | 0.65 (0.58–0.73) | 0.72 (0.64–0.81) | 2003 |
* Conclusion is possibly affected by publication bias