Table 7. Noteworthy observations and sensitivity analysis.
*Cook’s distance (range) (others) refers to the range in Cook’s distance for the other analyzed studies.
| Anatomical variant | Noteworthy studies | Cook's distance | Cook’s distance (others) (range)* | Original estimate (%) | Sensitivity analysis: estimate (%) | Original I² (%) | Sensitivity analysis: I² (%) | Observations/Comments |
| Absent SSN | Kumar et al. (2014) [25] | 2.0 | 0 – 0.3 | 10.7 | 10.75 | 100 | 100 | No substantial impact on heterogeneity or central tendency despite high Cook’s distance. |
| U-shaped | Iqbal et al. (2009) [21] | 0.6 | 0 – 0.4 | 31.7 | 36.4 | 96.03 | 95.68 | Major changes in effect size, indicating study impact. |
| V-shaped | Agrawal et al. (2015) [12] | 1.1 | 0 – 0.8 | 28.981 | 22.64 | 97.58 | 96.34 | 1.24% change in heterogeneity; significant effect size adjustment. |
| SS foramen | Kumar et al. (2014) [25] | 0.7 | 0 – 0.6 | 3.62 | 3.58 | 72.23 | 68.18 | Minor changes in both effect size and heterogeneity, suggesting low influence of this study. |
| SS foramen and notch | Natsis et al. (2007) [1] | 6 | 0 – 1 | 0.942 | N/A | 2.93 | N/A | Unable to perform sensitivity analysis due to high Cook's distance and a limited number of studies. Interpretation of this variant's effect size and heterogeneity should be done with caution. |
| Incomplete ossification | Inoue et al. (2014) [20] | 0.8 | 0 – 0.2 | 6.49 | 5.56 | 90.09 | 81.87 | Moderate changes in both effect size and I², indicating study influence. |
| Small V-shaped | Albino et al. (2013) [15] | 0.9 | 0 – 0.2 | 9.04 | 7.43 | 96.91 | 95.58 | Significant changes in effect size and slight I² reduction, indicating study influence. |
| J-shaped | N/A | N/A | N/A | N/A | N/A | N/A | N/A | Insufficient data to perform a sensitivity analysis. |
| Double foramen | N/A | N/A | N/A | N/A | N/A | N/A | N/A | Insufficient data to perform a sensitivity analysis. |
| “Polguj method” classification | ||||||||
| Type IA | Vyas et al. (2012) [32] | 0.6 | 0 – 0.1 | 12.1 | 14.3 | 85.03 | 0 | Significant change in heterogeneity to 0% and a minor increase in effect size, indicating a notable influence of this study on overall results. |
| Type IB | Polguj et al. (2013) [28] | 2.2 | 0.5 – 1.5 | 5.79 | 8.65 | 82.53 | 85.66 | A moderate increase in both effect size and heterogeneity upon sensitivity analysis, suggesting that this study has a noteworthy influence on the overall meta-analysis. |
| Type IC | N/A | N/A | N/A | N/A | N/A | N/A | N/A | Insufficient data to perform a sensitivity analysis. |
| Type II | Polguj et al. (2013) [28] | 0.35 | 0 – 0.15 | 2.1 | 2.35 | 0 | 0 | Minimal change in both effect size and heterogeneity upon sensitivity analysis, indicating that this study has a negligible influence on the overall meta-analysis. |
| Type IIIA | Vyas et al. (2012) [32] | 6 | 0.2 – 1.2 | 5.3 | 9.57 | 82.24 | 87.31 | Marked variations in effect size and heterogeneity were observed after sensitivity analysis, signifying that this study's strong impact on the overall results of the meta-analysis. |
| Type IIIB | Polguj et al. (2013) [28] | 4.5 | 0 – 2.5 | 3.72 | 7.2 | 87.49 | 89.1 | The significant change in effect size highlights the study's substantial influence. |
| Type IIIC | N/A | N/A | N/A | N/A | N/A | N/A | N/A | Insufficient data to perform a sensitivity analysis. |
| Type IV | Polguj et al. (2013) [28] | 0.5 | 0 – 0.3 | 4.35 | 3.93 | 0 | 0 | A relatively stable but minor impact on the overall meta-analysis. |
| Type V | Vyas et al. (2012) [32] | 0.7 | 0 – 0.3 | 15.8 | 12.0 | 92.97 | 10.37 | The heterogeneity dropped dramatically, implying that this study was a significant source of variability in the original meta-analysis. |