TABLE 9.
Sensitivity analysis of meta-analyses of fish and ω-3 fatty acid intake and cancer risk by study design (cohort and case-control)1
| Observational studies | Cohort | Case-control | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Author & year, type of cancer | Type of ω-3 fatty acid intake2 | n | Summary estimate (95% CI)3 | Level of evidence | n | Summary estimate (95% CI)3 | Level of evidence | n | Summary estimate (95% CI)3 | Level of evidence |
| Wu S et al., 2011 (18) | ||||||||||
| Gastric cancer | High fish consumption | 17 | 0.87 (0.71, 1.07) | Not significant | 2 | 1.10 (0.75, 1.61) | Not significant | 15 | 0.85 (0.68, 1.06) | Not significant |
| Chen G-C et al., 2015 (20) | ||||||||||
| Colorectal cancer | Total n-3 PUFA intake (high vs. low) | 10 | 0.99 (0.92–1.06) | Not significant | 2 | 1.02 (0.92, 1.12) | Not significant | 11 | 0.97 (0.87, 1.08) | Not significant |
| Colorectal cancer | Marine n-3 PUFA intake (high vs. low) | 11 | 1.00 (0.93–1.07) | Not significant | 2 | 1.04 (0.92, 1.17) | Not significant | 9 | 0.98 (0.90, 1.07) | Not significant |
| Zheng J-S et al., 2013 (10) | ||||||||||
| Breast cancer | Highest marine n-3 PUFA intake | 17 | 0.86 (0.78–0.94) | Weak | 11 | 0.86 (0.77, 0.96) | Weak | 6 | 0.83 (0.67, 1.03) | Not significant |
| Breast cancer | Total n-3 PUFA | 10 | 0.96 (0.86–1.06) | Not significant | 4 | 0.99 (0.91, 1.08) | Not significant | 6 | 0.95 (0.61, 1.19) | Not significant |
| Breast cancer | Highest dietary fish intake | 11 | 1.03 (0.93–1.14) | Not significant | 9 | 1.05 (0.94, 1.18) | Not significant | 2 | 0.83 (0.57, 1.20) | Not significant |
| Breast cancer | Per 15 g/d increment of fish intake | 11 | 1.00 (0.97–1.03) | Not significant | 9 | 1.00 (0.97, 1.03) | Not significant | 2 | 0.92 (0.72, 1.19) | Not significant |
| Breast cancer | Marine n-3 fatty (EPA) | 10 | 0.93 (0.85–1.02) | Not significant | 4 | 0.89 (0.74, 1.07) | Not significant | 6 | 0.78 (0.60, 1.02) | Not significant |
| Breast cancer | Marine n-3 fatty (DHA) | 10 | 0.88 (0.75–1.03) | Not significant | 4 | 0.92 (0.74, 1.15) | Not significant | 6 | 0.83 (0.63, 1.08) | Not significant |
| Breast cancer | Marine n-3 fatty (DPA) | 4 | 0.99 (0.98, 1.01) | Not significant | 1 | 0.94 (0.63, 1.38) | Not significant | 3 | 0.89 (0.56, 1.42) | Not significant |
| Breast cancer | ALA (tissue biomarker and diet) | 12 | 0.97 (0.90, 1.04) | Not significant | 6 | 0.97 (0.90, 1.06) | Not significant | 6 | 0.87 (0.67, 1.12) | Not significant |
| Hoang T et al., 2019 (38) | ||||||||||
| Ovarian cancer | ALA intake (high vs. low) | 3 | 0.99 (0.77, 1.26) | Not significant | 1 | 1.00 (0.72, 1.39) | Not significant | 2 | 0.97 (0.66, 1.43) | Not significant |
| Lian W et al., 2017 (39) | ||||||||||
| Brain tumor | Fish intake (high vs. low) | 9 | 0.83 (0.70, 0.99) | Weak | 1 | 1.05 (0.82, 1.34) | Not significant | 8 | 0.79 (0.66, 0.95) | Weak |
| Brain tumor | Per 100 g/wk increase fish intakes | 9 | 0.95 (0.91, 0.98) | Weak | 1 | 0.96 (0.92, 1.01) | Not significant | 8 | 0.94 (0.89, 0.99) | Weak |
| Noel SE et al., 2014 (40) | ||||||||||
| Skin cancer, squamous cellcarcinoma | n-3 PUFA intake (high vs. low) | 3 | 0.86 (0.59, 1.23) | Not significant | 2 | 0.98 (0.71, 1.36) | Not significant | 1 | 0.70 (0.49, 1.00) | Not significant |
1
n represents the number of studies included in the meta-analysis. ALA, α-linolenic acid; DPA, docosapentaenoic acid; NA, not assessible.
2
Definitions of comparison of each category follow that described in the original studies.
3
All summary estimates and 95% CIs were obtained by reanalysis. They were based on a random-effects model.