Table 2.
Reported results and statistical adjustments in the prospective studies included in the meta-analysis
| Author, years | Consumption, highest vs. lowest | RR (95% CI), highest vs. lowest | Covariate adjustment | Smoking adjustment |
|---|---|---|---|---|
| Jacobsen, 1986 [19] | ≥7 vs. ≤ 2 cups/d |
0.99 (0.53–1.86) 0.98 (0.47–2.03)(M) |
Age, sex, residence, and smoking (for M only) | Never, former, current (1–9, 10–19, ≥20 cig/d) (for M only) |
| Mills, 1991 [26] | ≥2 cups/d vs. never |
1.99 (0.91–4.34) 2.03 (0.70–5.87)(NS) 1.14 (0.46–2.80)(FS/CS) |
Age, sex, and smoking | Never, former, current |
| Chyou, 1993 [27] | ≥5 vs. ≤ 1 times/wk | 2.07 (0.84–5.12) | Age and smoking | Pack-years (0, > 0–30, > 30) |
| Stensvold, 1994 [15] | ≥7 vs. ≤ 2 cups/d |
1.50 (0.45–5.02)(M) 2.40 (0.28–20.5)(W) |
Age, residence, and smoking | Cig/d (continuous) |
| Michaud, 1999 [28] | ≥4 cups/d vs. < 1 cup/mo | 0.79 (0.48–1.30) | Age, region, energy intake, fruit and vegetable intake, and smoking | Smoking status (smoker, nonsmoker) and pack-years (6 categories) |
| Nagano, 2000 [29] | ≥5 vs. 0 times/wk | 0.90 (0.52–1.56) | Age, sex, radiation dose, education, BMI, calendar time, and smoking | Never, former, current (≤20, > 20 cig/d) |
| Zeegers, 2001 [16] | ≥7 (M)/≥5(W) vs. < 2 cups/d |
1.36 (0.82–2.04)(M) 0.32 (0.15–0.68)(W) |
Age, tea consumption, and smoking | Cig/d (continuous), years of smoking (continuous) |
| Tripathi, 2002 [30] | ≥4 cups/d vs. < 1 cup/mo | 1.59 (0.95–2.68) | Age | None |
| Kurahashi, 2019 [17] | ≥3(M)/≥1(W) cup/d vs. almost never |
1.37 (0.75–2.51)(M) 0.55 (0.23–1.33)(W) 2.48 (0.88–7.05)(NS)(M) 2.09 (0.96–4.54)(FS)(M) 2.24 (1.21–4.16)(NS/FS)(M) 1.13 (0.65–1.97)(CS)(M) |
Age, area, alcohol, green tea consumption, and smoking | Never, former, current (< 25, ≥25 pack-years) |
| Ros, 2011 [21] | ≥875(M)/500(W) ml/d vs. < 429 (M)/250(W) ml/d | 1.11 (0.85–1.43) | Age, sex, center, energy intake, and smoking. | Smoking status (never, former and current), duration (continuous), and lifetime intensity (continuous) |
| Hashibe, 2015 [31] | ≥2 vs. < 1 cup/d | 1.08 (0.85–1.39) | Age, sex, race, education, and smoking | Smoking status (never, former, current), frequency (1–10, 11–20, 21–30, > 30 cig/d), duration (1–10, 11–20, > 20 yr), years since quitting (> 0–2, 3–5, 6–10, 11–20, > 20 yr). |
| Loftfield, 2017 [6] | ≥4 cup/d vs. none |
1.18 (1.05–1.33) 1.25 (1.09–1.43)(M) 0.97 (0.74–1.25)(W) 0.87 (0.65–1.17)(NS) 1.23 (1.04–1.33)(FS) 1.32 (0.95–1.81)(CS) |
Age, sex, race/ethnicity, BMI, education, reported health status, fruit intake, vegetable intake, supplement use, physical activity, diabetes, family history of cancer, and smoking | Pipes or cigars (ever, never), smoking frequency (1–10, 11–20, 21–30, 31–40, 41–60, ≥60 cig/d), years since quitting (≥1–4, 5–9, ≥10 yr). |
| Sugiyama, 2017 [7] | ≥3 cup/d vs. none |
0.56 (0.32–0.99) 0.57 (0.31–1.07)(M) 0.44 (0.10–1.97)(W) 0.62 (0.14–2.72)(NS) 0.61 (0.32–1.17)(FS/CS) |
Age, sex, BMI, hypertension, diabetes, MI, stroke, job status, education, alcohol, green tea consumption, walking, and smoking | Never, former, current (< 20, ≥20 cig/d) |
| Gapstur, 2017 [5] | ≥6 cup/d vs. never |
0.89 (0.73–1.09)(NS/FS) 0.80 (0.57–1.12)(NS) 0.97 (0.74–1.27)(FS) |
Age, sex, race, marital status, education, alcohol consumption, BMI, physical activity, family history of cancer, red and processed meat intake, vegetable intake, tea consumption, and smoking. | Years since quitting (< 10, 10- < 20, ≥20 yr) and cig/d (< 20, 20–29, ≥30). |
| Lukic, 2018 [8] | ≥4 vs. < 1 cup/d |
1.34 (0.94–1.90) 1.23 (0.78–1.95)(M) 1.46 (0.84–2.51)(W) 1.87 (1.01–3.45)(NS) 1.18 (0.77–1.81)(FS/CS) |
Age, sex, and smoking | Never, former, current |
| Hashemian, 2019 [9] | ≥4 vs. < 1 cup/d | 1.10 (0.81–1.49)(CS) | Age, education, alcohol, diabetes, physical activity, fruit intake, vegetable intake, tea consumption, and smoking | Smoking years (continuous), cig/d (continuous) |
Abbreviations: BMI body mass index, cig cigarettes, CS current smoker, FS former smoker, M men, MI myocardial infarction, mo month, NS never smoker, W women, wk, week
a 95% CIs were calculated using raw data
b Data were rescaled by using the lowest consumption group as the reference