The February issue of Journal of the National Cancer Institute featured two articles on coffee consumption and cancer: a Japanese group claimed that drinking caffeinated coffee could prevent half of hepatocellular carcinoma (liver cancer) in the Japanese cohort. Previous studies have shown that drinking coffee could reduce the risk of gallstone and liver cirrhosis. These might inspire the Japanese group to investigate the protective effect of coffee on liver cancer.

On the other hand, Willett from Harvard U reported that drinking coffee was not related to colon rectal cancer in his Nurse and Physician cohorts, contrary to earlier German study which suggested a protective effect of coffee. Nevertheless, Willett did hint that drinking decaffeinated coffee might prevent rectal cancer.

Although the above information is encouraging, it reminds me an old story about coffee drink and pancreatic cancer. A report by Charles Hennkin in 1980s showed an increased risk of pancreatic cancer among those who drank coffee regularly. That study was criticized due to its case control study nature and the neglect of adjusting for confounders. Subsequent studies didn’t reach any consensus.

We should notice that all these controversial results are from observational studies. Although they are widely used in etiological inference, they can’t determine the causal link, a fact many news media grossly ignored. I am not saying results from random controlled trials are better. Ill designed trials are worse than observational studies. Recent scandal on Vioxx is such an example. Nevertheless, observational studies do suffer many weaknesses.

Take the Japanese coffee study as an example. The study only measured coffee consumption once, at baseline, which was many years ago. People do change habits over the time. In addition, the study didn’t have information on hepatic virus infection (in particular HBV), a recognized risk factor for liver cancer. Hepatic virus infection is very high in the East Asian. It is very likely that those who had HBV infection would abstain from coffee. In addition, those who drink coffee were more likely living in a modern lifestyle than those who don’t drink. Studies have shown that immigrants from Japan to the US had reduced risk of liver cancer and stomach cancer than those native Japanese (NiHonSan study). The change of diet is likely the reason, but no one knows which diet factor cause the reduction. Furthermore, liver cancer tends to cluster in areas or in specific population, probably due to the cluster of HBV infection, or genetic susceptibility. Therefore, even with many statistical adjustments in their models, all of their claims are inconclusive. In fact, most these claims will be unvalidated in future studies.

Are there any good stuff in coffee? Not really. There is no mystery why coffee makes you feel high. It is caffeine, the substance many people addict to. The magic antidotes—antioxidants such as flavonoids— are widely available in other fruits. If coffee can help you, many other fruits will help you too.

I don’t drink either coffee or tea. Instead, I live in a primitive way. I achieve the same stimulating effect by flushing myself with two big bottles of plain water a day. It’s much safer than drinking coffee. Anyway, coffee hurts my esophagus and stomach.