Prevention of Gastric Cancer: A Miss

Despite its declining global incidence and mortality, gastric cancer remains one of the most common cancers in the world, with nearly one million new patients expected to be afflicted with the disease this year. Given the 5-year relative survival rate among gastric cancer patients of only 23% ( 1 ), prevention is clearly the preferred strategy to reduce morbidity and mortality from this disease. In addition to secondary prevention approaches, which rely on screening for early detection, primary prevention efforts have focused on chemoprevention, using either nutritional interventions or anti– Helicobacter pylori therapies.

In this issue, Plummer et al. ( 2 ) report results of a nutritional intervention for the chemoprevention of gastric premalignancy. This randomized, double-blind, placebo-controlled trial tested the efficacy of daily supplementation for 3 years with a combination of three different nutrients—vitamin C (750 mg/day), vitamin E (600 mg/day), and beta-carotene (18 mg/day)—versus placebo on the regression or progression of premalignant gastric lesions in nearly 2000 subjects from Venezuela. Despite a tendency toward more regression and less progression of gastric premalignant lesions comparing the treatment and control arms, none of the results approached statistical significance. Disappointingly, this must be considered a negative trial.

While the fundamentals of the trial were sound—the intervention was conducted in a high-risk population, the antioxidants tested were strong a priori candidates based on substantial prior epidemiologic data, compliance was excellent, endpoints were carefully and systematically assessed, and results were appropriately analyzed—there are several design and implementation issues that may have contributed to this null result. The most important issue is the attributes of the population itself. Baseline median plasma values for the three micronutrients supplemented in the trial exceeded median values in the US population by 18%–56% ( 3 ), evidence of a population with good nutriture. In addition, the supplement dosages selected were high—eightfold the Recommended Dietary Allowance (RDA) for vitamin C, 40-fold the RDA for vitamin E, and six- to sevenfold the estimated average US dietary intake of beta-carotene ( 3 )—and they produced marked increases in status for each of these nutrients. At 1 year, plasma vitamin C levels increased 30% over baseline, vitamin E levels doubled, and beta-carotene levels were up 6.7-fold [to levels that were approximately three-fourths those seen in beta-carotene recipients in the Beta-Carotene and Retinol Efficacy Trial (CARET) and approximately one-half the level of beta-carotene recipients in the Alpha-Tocopherol, Beta-Carotene (ATBC) Trial ( 3 )]. Thus, this trial was a test of pharmacologic doses of micronutrients in an already well-nourished population. It is also possible that the intervention duration and follow-up may have been too short and that benefit will emerge only later after longer latency and follow-up. Another potential explanation for the lack of benefit of vitamin supplementation is that the high prevalence of advanced premalignant lesions and H. pylori infection rates were simply too much to overcome with the study design employed here. Although the authors appropriately conducted an intent-to-treat analysis, concerns about beta-carotene toxicity in smokers from the ATBC Trial and CARET prompted investigators to switch all current and previous smokers in the vitamin treatment group (24% of the group) to placebo, thereby diluting any potential intervention effect. Finally, a large number of subjects were censored (30% of those randomized), and although censoring did not differ by treatment group, censored subjects varied from uncensored subjects for several potentially important risk factors (e.g., censored subjects were younger than uncensored subjects).

There have now been, by my count, a total of seven different trials that have reported nutritional interventions in the prevention of gastric premalignancy using intent-to-treat analyses ( 2 , 4 – 9 ). These trials enrolled from 216 to 3365 subjects, tested 13 different nutrients alone or in combination, and supplemented participants for periods varying from 2 to more than 7 years. Three of these seven trials reported one or more statistically significant findings, including one unfavorable and two favorable results: ascorbic acid plus molybdenum increased gastric dysplasia/cancer in the first trial ( 4 ); beta-carotene and ascorbic acid each (separately) reportedly regressed gastric premalignant lesions in the second ( 7 ); and folic acid with B12 produced more reversions of dysplasia than placebo in the third ( 8 ).

While premalignancy is valuable as a surrogate, the real endpoints of interest are incident or fatal cancer. At least nine reports from six different nutritional chemoprevention trials with an endpoint of gastric cancer have also been published ( 4 – 6 , 8 – 13 ). Of these gastric cancer endpoint trials, only two reported statistically significant findings, both favorable. The first (and largest by far) was the Linxian General Population Trial in which the combination of beta-carotene, alpha-tocopherol, and selenium supplemented for 5.25 years to a nutritionally inadequate population with extraordinarily high rates of upper gastrointestinal tract cancers reduced gastric cancer mortality by 21% ( 10 ). In the same trial, gastric noncardia cancer deaths were also reduced 41% among persons supplemented with retinol/zinc. Li et al. ( 13 ) recently reported that a synthetic garlic compound (allitridum) given with a very small amount of selenium for 2 years reduced the incidence of gastric cancer for a period of 5 years of follow-up; this result, however, was based on just 29 total cases and was statistically significant only among males. Benefit was much reduced when the full 10 years of available follow-up was considered.

The other focus of attempts to achieve primary chemoprevention of gastric cancer has been anti– H. pylori therapy. Just four such trials that used gastric premalignancy as an endpoint have been reported, but the anti– H. pylori therapy arm benefited to a statistically significant extent in all of them ( 7 , 9 , 14 , 15 ). A single randomized trial set out to prevent gastric cancer (rather than premalignancy) using anti– H. pylori therapy ( 16 ). Although only 18 new gastric cancers were observed, fewer occurred in the anti– H. pylori arm than in the placebo arm (7 versus 11, respectively). Of potential import was the finding that in a subgroup analysis among participants with the mildest gastric abnormalities at baseline (i.e., chronic active gastritis), no gastric cancers developed in the anti– H. pylori arm compared to six in the placebo arm ( P = .02).

So where does this leave us in the battle to prevent gastric cancer? Despite the declining trends worldwide, prevention of gastric cancer should remain a high priority. There are still too many cases and too many deaths to hope that the slow march of modernity will simply overcome this disease in a reasonable time without further help. How best to help then? Of the two primary prevention approaches evaluated in randomized trials to date, anti– H. pylori therapy appears to be more effective, although the data are still quite limited. However, there remains a role for nutritional interventions as well. The most likely reason for our limited success to date with nutritional interventions is that we have not optimized our interventions to our target populations. The Linxian story is a case in point. In this population with low nutrient status and high risk, supplementation with beta-carotene/alpha-tocopherol/selenium reduced gastric cancer mortality ( 10 ). Observational studies in the trial cohort showed a strong association between high serum levels of selenium and reduced risk of gastric cancer ( 17 ), and follow-up of trial participants for 10 years after the end of supplementation continues to show benefit for the intervention ( 18 ). These findings affirm a role for selenium in the prevention of gastric cancer, at least in one population with low status for this nutrient, and they suggest a logical public health strategy (i.e., nutrient fortification) to prevent gastric cancer in this and possibly other similar populations.

Evidence for the effectiveness of anti– H. pylori therapy in the prevention of gastric premalignancy is accumulating and appears consistent; evidence for the prevention of gastric cancer itself with this treatment approach, although quite limited, is also promising. The available data suggest that an appropriate next step is a proof-of-principle test of anti– H. pylori therapy in a trial of sufficient size and duration to be truly informative regarding the prevention of gastric cancer. There are reasons, however, why anti– H. pylori microbial therapy will not be the sole or ultimate solution: drug costs are substantial, compliance is problematic (particularly with the longer, more effective regimens), reinfection is common, widespread use of antibiotics leads to microbial resistance, H. pylori eradication may increase gastroesophageal junction cancers, and a H. pylori vaccine (if and when developed) might reduce or obviate the need for antibiotic treatment.

Despite the findings from the current report, there are abundant reasons to be optimistic about the prevention of gastric cancer. The way forward will require continued etiologic research to identify new modifiable factors, such as bioactive food components or other alterable environmental factors, as well as a search for beneficial associations with drugs. However, this overall effort should continue to emphasize randomized controlled trials as the most powerful and valid approach for testing specific prevention strategies and follow lessons and leads from the first generation of cancer prevention interventions ( 19 ). Importantly, the next trials of nutrients should match the intervention to the population, emphasizing nutrient repletion (i.e., physiologic as opposed to pharmacologic doses) in high-risk populations with poor nutrition.

References