When white blood cells fall, food can become another source of anxiety. Should every person receiving chemotherapy avoid fresh produce? The evidence does not support one rule for every setting. It also does not justify dismissing a transplant team’s restrictions after reading a reassuring headline.
Evidence at a glance
Question: Does restricting foods such as raw produce prevent infection during neutropenia? Population: Evidence spans cancer treatments with very different risks, particularly acute leukemia and stem-cell transplantation. Bottom line: Routine restriction has not shown an overall benefit, but selected high-risk settings remain a concern. Food safety is essential regardless of diet.
Earlier trials were reassuring
A 2022 randomized trial enrolled 200 patients aged 1–60 receiving acute-leukemia induction chemotherapy in India. Major infections occurred in about 32% with a regular diet and 25% with a restrictive diet; the difference was not statistically significant. This was not proof that the two approaches were identical.
A 2023 multicenter randomized trial enrolled 222 adults, 79% receiving autologous transplantation, using their own stem cells. It found no significant difference in clinically important infections or deaths during neutropenia, while satisfaction favored the nonrestrictive diet. The predominance of autologous transplantation limits what it says about other high-risk settings.
A 2026 trial raised a safety concern
A trial published in the February 2026 JCO issue compared restrictive and liberalized diets in stem-cell transplant recipients and people with acute leukemia. The liberalized diet included fresh fruits and vegetables. The study stopped early after 214 evaluable participants because its predefined safety boundary was crossed.
Major infections occurred in 31.4% with liberalization versus 20.2% with restriction—an 11.2-percentage-point difference. The 95% confidence interval ranged from −0.4 to 22.9 points. Because that interval includes zero, the comparison does not conclusively establish a higher infection rate; importantly, it also failed to show noninferiority within the trial’s 10-point margin.
Eating more freely did not improve calorie intake, and nutritional problems remained common in both groups. This finding is a reason for caution in the studied setting, not proof that every outpatient with brief neutropenia needs the same restrictions.
The September update puts both sides together
A September 10, 2026 systematic review combined 15 studies with 4,320 patients. Overall, restrictive diets did not significantly reduce major infections or mortality. However, among six randomized studies restricted to blood-cancer or transplant populations, bacteremia or fungemia—bacteria or fungi in the bloodstream—was less frequent with restriction (relative risk 0.67; 95% CI 0.47–0.95). Other outcomes did not differ in that subgroup.
A subgroup result does not establish a universal rule, and differing diets, treatments and infection definitions complicate pooling. My reading is to distinguish short-lived, lower-risk neutropenia from prolonged or transplant-related immune suppression. These studies address supportive-care safety, not cancer control.
What remains practical
Follow the treating center’s instructions while asking how they apply to your treatment. MedlinePlus food-safety guidance for cancer treatment emphasizes safe cooking, washing permitted produce, and avoiding raw sprouts and unpasteurized products. “Liberalized” should never mean careless preparation. Ask for dietitian help if restrictions or symptoms make adequate nutrition difficult.
Practical questions for the care team
How deep and prolonged is my immune suppression expected to be? Which restrictions are specific to my treatment, and when can they end? Which foods are safe alternatives if appetite is poor? Has the team considered the newer evidence? What should I do if I cannot eat enough?
Our Evidence Explorer record provides additional context; this review incorporates the newer 2026 evidence.
Evidence checked October 19, 2026. Educational review, not individualized medical advice. Research doses and schedules are not personal prescriptions. Explore the Missing Script article library and the podcast—coming soon.
