The beta-carotene trial that overturned a prevention promise

One word interrupts the abstract of the Finnish beta-carotene trial published on April 14, 1994: “Unexpectedly.” Researchers had tested a plausible way to prevent lung cancer. The participants assigned beta-carotene developed more of it.[1]

The surprise becomes clearer when the paper is read as a sequence of questions. What had earlier observations suggested? What did the researchers actually give people? Which comparison produced the result? Following those questions reveals how easily a promising dietary association can acquire a much larger claim when it moves into a capsule.

Demetrius Albanes, an ATBC investigator, photographed outdoors wearing a dark jacket and tie. — Demetrius Albanes, principal investigator of the ATBC Study, in a portrait published by the National Cancer Institute. Photo: NCI.[3]

What went into the capsule

The Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study—ATBC—recruited 29,133 men aged 50–69 in southwestern Finland between 1985 and 1988. All smoked at least five cigarettes a day. The trial therefore began with a defined population already exposed to a powerful cause of lung cancer.[2]

Its four groups received vitamin E, beta-carotene, both, or placebo. The daily doses were 50 milligrams of vitamin E and 20 milligrams of beta-carotene. This arrangement, called a factorial design, let investigators compare everyone assigned beta-carotene with everyone not assigned it, while also examining vitamin E. Reading the beta-carotene result as merely the combined-vitamin group against placebo misses that design.[2]

The intervention ended on April 30, 1993, after five to eight years. Cancer detection relied on Finland's national registry, with medical-record review and examination of available tissue or cell specimens during the trial.[2] The outcome was diagnosed cancer over years of observation. A favorable blood measurement alone could not count as successful prevention.

A dietary clue becomes an intervention

There were good reasons to ask the question. Antioxidants can neutralize reactive chemicals capable of damaging cells, and laboratory work had suggested ways this might matter for cancer. But the National Cancer Institute's evidence review distinguishes that biological rationale from proof that supplements improve human outcomes.[4]

A dietary association contains several possible explanations. People who eat differently may also smoke differently; foods contain mixtures of substances; blood concentrations can reflect processes beyond a single nutrient's intake. NCI's lung-cancer review specifically notes how difficult it is to separate fruit-and-vegetable associations from smoking.[6]

Random allocation addressed a narrower question: what happens when this supplement is added under these conditions? It reduced the biases of comparisons between people's existing habits. The observational work supplied a hypothesis; the trial tested the intervention.[4]

Read the increase with its comparison attached

The 1994 report recorded 876 new lung cancers. Incidence was 18% higher among men assigned beta-carotene than among those not assigned it, with a 95% confidence interval of 3%–36% higher. Total mortality was also higher: 8%, with a confidence interval of 1%–16%.[1]

The first percentage is a relative increase. It does not mean that 18 extra men in every hundred developed lung cancer, or that 18% of participants did. The interval expresses uncertainty about the size of the increase; it lies entirely on the side of harm.[1]

That distinction changes the reading of the trial. This was more troubling than a supplement simply failing to deliver its promised benefit. The abstract's cautious conclusion allowed for harmful effects, even though the experiment had begun with prevention in view.[1]

A second trial strengthened the warning

The US Beta-Carotene and Retinol Efficacy Trial, CARET, tested a different regimen: 30 milligrams of beta-carotene plus 25,000 IU of retinol daily, compared with placebo. Its 18,314 participants included smokers, former smokers and workers exposed to asbestos.[5]

The report published in May 1996 found a lung-cancer relative risk of 1.28 (95% CI 1.04–1.57) after an average four years of follow-up. All-cause mortality was higher too, and the intervention was stopped 21 months earlier than planned.[5]

CARET administered two agents together, so its main comparison cannot allocate the entire effect to either ingredient alone.[5] ATBC's factorial comparison helped isolate the concern about beta-carotene.[2] Read together, the trials strengthened the warning while preserving the distinction between the regimens.

Keep the limits beside the lesson

Neither trial assigned people to eat more carrots or leafy vegetables. NCI points out that purified supplements and foods containing complex mixtures may act differently.[4] The capsule results therefore do not establish that carotene-rich vegetables cause lung cancer. Nor do they retrospectively prove that beta-carotene was the protective component of an observed dietary pattern.

The population matters as well. NCI's review notes that evidence from other trials, including the Physicians' Health Study, does not show the same increased lung-cancer risk in nonsmokers.[6] That limits a universal claim of harm; it supplies no reason to recover the original promise of prevention.

The US Preventive Services Task Force's 2022 recommendation advises against beta-carotene supplements for preventing cancer or cardiovascular disease in community-dwelling, nonpregnant adults. Its scope excludes people with known nutritional deficiencies and several other groups whose clinical needs require separate assessment.[7] Prevention in generally healthy adults and treatment of a deficiency are different questions.

The lasting achievement of ATBC was to let an outcome overrule a persuasive explanation. A nutrient's role in biology made it worth studying. Only the comparison between assigned groups could establish what taking the capsule actually did.

cronfeed.work