28.02.2013 Views

Full Report - Food, Nutrition, and the Prevention of Cancer

Full Report - Food, Nutrition, and the Prevention of Cancer

Full Report - Food, Nutrition, and the Prevention of Cancer

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

CHAPTER 4 • FOODS AND DRINKS<br />

Figure 4.2.12 Raw vegetables <strong>and</strong> stomach cancer;<br />

cohort <strong>and</strong> case-control studies<br />

Cohort<br />

Botterweck 1998 0.51 (0.05–4.73)<br />

Galanis 1998 0.81 (0.53–1.23)<br />

Khan 2004 Men 0.63 (0.16–2.51)<br />

Khan 2004 Women 2.03 (0.14–29.46)<br />

Summary estimate 0.80 (0.54–1.18)<br />

Case control<br />

Jedrychowski 1981 0.10 (0.02–0.41)<br />

Jedrychowski 1986 1.56 (0.12–20.74)<br />

Buiatti 1989 0.56 (0.45–0.71)<br />

Caggon 1989 0.13 (0.01–2.88)<br />

Kato 1990 Men 0.59 (0.38–0.91)<br />

Kato 1990 Women 0.86 (0.48–1.48)<br />

Hoshiyama <strong>and</strong> Sasaba 1992 0.44 (0.29–0.68)<br />

Ramon 1993 0.35 (0.16–0.78)<br />

Cornee 1995 0.27 (0.10–0.72)<br />

Huang 1999 0.77 (0.62–0.94)<br />

De Stefani 2001 0.46 (0.16–1.34)<br />

Sriamporn 2002 0.40 (0.01–15.73)<br />

Lee 2003 0.11 (0.04–0.27)<br />

Lissowska 2004 0.82 (0.52–1.28)<br />

Summary estimate 0.50 (0.38–0.65)<br />

0.5 1 2<br />

Relative risk, per 100 g/day<br />

Relative risk (95% CI)<br />

intake group when compared to <strong>the</strong> lowest). 241<br />

Most case-control studies showed decreased risk with<br />

increased intake, which was statistically significant in 10. 58<br />

109 152 156 164 171 232 246-248 No studies showed statistically significant<br />

increased risk. Meta-analysis was possible on 6 casecontrol<br />

studies, giving a summary effect estimate <strong>of</strong> 0.40<br />

(95% CI 0.19–0.82) per 100 g/day, with high hetero-<br />

109 152 171 232 244 250<br />

geneity.<br />

White or pale vegetables<br />

This incorporates a wide range <strong>of</strong> vegetables. For example,<br />

in Japan white vegetables such as daikon (radish) are commonly<br />

consumed. Descriptions used for this exposure were<br />

white vegetables, pale green or light green vegetables, <strong>and</strong><br />

raw chicory.<br />

Both cohort studies showed non-significant decreased risk<br />

with increased intake. 150 214 Meta-analysis was possible on<br />

both studies, giving a summary effect estimate <strong>of</strong> 0.49 (95%<br />

CI 0.24–1.01) per 100 g/day, with no heterogeneity (figure<br />

4.2.11).<br />

All six case-control studies showed decreased risk with<br />

increased intake, 157 165 169 226 228 243 which was statistically significant<br />

in three. 165 169 243 Meta-analysis was possible on three<br />

studies, giving a summary effect estimate <strong>of</strong> 0.57 (95% CI<br />

0.32–1.02) per 100 g/day, with high heterogeneity, which<br />

was caused by varying size, not direction <strong>of</strong> <strong>the</strong> effect (figure<br />

4.2.11).<br />

Figure 4.2.13 Raw vegetables <strong>and</strong> stomach cancer;<br />

case-control studies: dose response<br />

Jedrychowski 1981<br />

Jedrychowski 1986<br />

Buiatti 1989<br />

Coggon 1989<br />

Kato 1990 Men<br />

Kato 1990 Women<br />

Hoshiyama <strong>and</strong> Sasaba 1992<br />

Ramon 1993<br />

Cornee 1995<br />

Huang 1999<br />

De Stefani 2001<br />

Sriamporn 2002<br />

Lee 2003<br />

0 50 100<br />

150<br />

Raw vegetables (g/day)<br />

Raw vegetables<br />

Of seven independent estimates from six cohort studies that<br />

reported on raw vegetables, four reported non-significant<br />

reduced risk, 146 214 253 254 two reported non-significant<br />

increased risk, 214 252 <strong>and</strong> <strong>the</strong> o<strong>the</strong>r reported a significant<br />

increased risk. 148 Two <strong>of</strong> <strong>the</strong> increased risk estimates, including<br />

<strong>the</strong> one that reached statistical significance, were stratified<br />

for women only. Meta-analysis was possible on four<br />

estimates from three studies (not including <strong>the</strong> one that was<br />

statistically significant), giving a summary effect estimate <strong>of</strong><br />

0.80 (95% CI 0.54–1.18) per 100 g/day, with no heterogeneity<br />

(figure 4.2.12).<br />

Of <strong>the</strong> 25 case-control studies that reported on raw veg-<br />

109 129 161 162 167 172 174 183<br />

etables, 21 reported decreased risk,<br />

184 191 219 225 226 243 247 248 250 255 256 258 260 261 264 which was sta-<br />

129 161 172 174 225 226 243 247 248 256 260 261<br />

tistically significant in 13.<br />

264 None <strong>of</strong> <strong>the</strong> remaining 4 studies that reported increased<br />

risk reached statistical significance. 257 259 262 263 Meta-analysis<br />

was possible on 14 independent estimates from 13 casecontrol<br />

studies, giving a summary effect estimate <strong>of</strong> 0.50<br />

(95% CI 0.38–0.65) per 100 g/day, with moderate heterogeneity<br />

(figure 4.2.12).<br />

A dose-response relationship was apparent from casecontrol<br />

but not cohort data (figure 4.2.13).<br />

Of <strong>the</strong> three ecological studies, two reported statistically significant<br />

reduced risk208 238 <strong>and</strong> one reported a non-significant<br />

increased risk with increased raw vegetable consumption. 202<br />

89

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!