Table 1 List of changes in text. Corrected texts are underlined, and original texts are placed in square brackets in italic. The text deleted are marked with strikethrough

Effect estimates for overall cancer and heterogeneity

Abstract: Results (Page 1)

An inverse association was observed between higher adherence to a plant-based dietary pattern and risks of… cancer (0.88 [0.84–0.92]) [(0.91 [0.87–0.96])] … with moderate to high heterogeneity across studies (I² ranged: 30.2%[47.8%] –95.4%,).

Results: Plant-Based Diet and Risk of T2D, CVD, cancer, and mortality (Page 5-6)

A greater adherence to plant-based dietary patterns was consistently associated with lower risks of T2D, CVD, cancer, and mortality. The random-effects pooled RR was 0.82 (95% CI: 0.77–0.86), 0.90 (95% CI: 0.85–0.94), 0.88 (95% CI: 0.84–0.92),[0.91 (95% CI: 0.87–0.96)], and 0.84 (95% CI: 0.78–0.92), respectively. A moderate heterogeneity was observed among studies for T2D (60.9%), CVD (49.8%), cancer (30.2%)[47.8%], and a high heterogeneity was observed for studies reporting mortality (95.4%).

Effect estimates for cancer subtypes and heterogeneity

Results: Plant-Based Diet and Risk of T2D, CVD, cancer, and mortality (Page 6)

For various cancer types, plant-based dietary patterns were significantly associated with lower risk of breast cancer (0.91 [95% CI: 0.86–0.95; I²=0%])[(0.92 [95% CI: 0.87–0.97; I²=11.1%])], digestive system cancer (0.82 [95% CI: 0.72–0.94; I²=0%]), pancreatic cancer (0.68 [95% CI: 0.55–0.84; I²=0%]) and prostate cancer (0.87 [95% CI: 0.77–0.99; I²=53.3%]), but not with risks of colorectal cancer (0.90 [95% CI: 0.79–1.02; I²=61.8%])[(0.98 [95% CI: 0.85–1.12; I²=68.4%])], liver cancer (0.51 [95% CI: 0.22–1.21; I²=57.7%])[(0.74 [95% CI: 0.52–1.05; I²=48.1%])], lung cancer (0.82 [95% CI: 0.54–1.26; I²=36.6%]) [(0.88 [95% CI: 0.56–1.38; I²=44.2%])], prostate cancer (0.94 [95% CI: 0.81–1.08; I² =60.7%]) or stomach cancer (1.73 [95% CI: 0.90–3.31; I²=0%]) (Supplemental Figure S2).

Test estimates for Publication bias

Results: Assessment of Publication Bias and Risk of Bias in Individual Studies (Page 9)

Furthermore, Egger regression tests and Begg-Mazumdar regression tests … did not detect significant publication bias for cancer (P=0.12 and P=0.96)[(P=0.07 and P=0.22)] and mortality (P=0.83 and P=0.52).

Effect estimates after performing the trim-and-fill analysis

Results: Assessment of Publication Bias and Risk of Bias in Individual Studies (Page 9)

After performing the trim-and-fill analysis to evaluate the robustness of associations after accounting for potential publication bias, our results remained largely unchanged. The random-effects pooled RR was … 0.91 (95% CI: 0.87–0.95)[0.89 (95% CI: 0.84–0.94)] for cancer …

Significancy change of prostate cancer results

Discussion (Page 9)

For specific disease outcomes, the inverse association for CVD was mainly driven by CHD, and for cancer by breast cancer, and pancreatic cancer, and prostate cancer.

Discussion (Page 12)

While we found that the inverse association with cancer is mainly driven by breast cancer, and pancreatic cancer, and prostate cancer, but not by colorectal, liver, lung, prostate, or stomach cancer.

Discussion (Page 13)

Third, study points on some diseases are limited (e.g., heart failure, lung cancer, liver cancer, prostate cancer, stomach cancer), and therefore it is unknown whether the observed non-significant association is true or due to the lack of statistical power.