77 cancer patients in a Brazilian hospital were each given a meal plan. On a typical day they ate about 900 kcal less than it, and the men who ate least were a kilo lighter by day six

When someone with cancer is admitted to hospital, a dietitian usually works out how much energy and protein they need and the kitchen sends up meals to match. Whether the food on the tray ends up inside the patient is another matter, and it is rarely measured day by day. A study published in Supportive Care in Cancer on 23 October set out to do exactly that on an ordinary hospital ward.

Researchers at the Hospital de Clínicas de Porto Alegre in southern Brazil followed 77 adults with gastrointestinal, lung or head and neck cancer, the types most often linked to malnutrition, through their first six days in hospital. All were eating normal food by mouth. Their average age was 62.6, 54.5% were men, seven in ten had a gastrointestinal cancer and, of those with a recorded stage, 84.6% had stage III or IV disease. Within 48 hours of admission, 80.6% were already classed as moderately or severely malnourished on a cancer-specific assessment.

On days 2, 4 and 6 a researcher asked each patient how much of their hospital meals they had eaten in the previous 24 hours, using a four-step picture scale: almost all, half, a little or nothing. That share was applied to the energy and protein in the meals prescribed. The prescriptions came to a median of 2,008 to 2,196 kcal and 98 to 109 g of protein a day, about 34 to 36 kcal and 1.6 to 1.7 g of protein per kilogram of body weight. Estimated intake was a median of 1,185 to 1,387 kcal and 55 to 59 g of protein. The typical shortfall was 867 kcal on day 2, 932 on day 4 and 1,012 on day 6, and 35 to 50 g of protein. Only 44.2% of patients averaged three quarters or more of their meals.

The patients who ate least were mostly the ones who said they had no appetite: 88.4% of them rated it low, against about a quarter of those who ate well. Loss of appetite, dry mouth and changes in taste and smell were all more common among the low eaters. Lung cancer stood out: 78.9% of lung cancer patients ate less than three quarters of their meals, against 46.3% of those with gastrointestinal cancer. Chemotherapy in the previous three months also went with eating less (70.0% against 46.8%), but the stage of the cancer did not.

Over just four days, men who ate less than three quarters of their meals lost an estimated 1.02 kg (from 66.25 to 65.23 kg) and their upper-arm circumference shrank by 0.21 cm, while men who ate well held steady. Women showed no clear change in weight or arm size, and grip strength didn't move in either sex. The authors say the sex difference may come down to the short follow-up and the small numbers once the group was split.

There are real limits. Intake was estimated from the fraction of the tray eaten, assuming people ate every item in the same proportion, when patients often leave some foods and finish others. Food or supplements from outside the hospital menu could have been missed, weight can swing with fluid balance, and only the first six days were covered. There was no comparison group and nothing was changed, so the study shows how big the gap is rather than how to close it. The authors call for symptom checks and closer tracking of intake from admission, with energy- and protein-dense food or supplements when needed. The work was funded by Brazilian public research bodies and the hospital's research fund; the authors declared no competing interests.