For patients undergoing major oncologic surgery, the best definition of malnutrition used to assess postoperative risk varies by cancer type, results of a retrospective study suggest.
The current, one-size-fits-all approach to nutritional status leads to both undertreatment and overtreatment of malnutrition, as well as inaccurate estimations of postoperative risk, reported lead study author Nicholas P. McKenna, MD, of the Mayo Clinic in Rochester, Minn., and colleagues.
“Assessing nutritional status is important because it impacts preoperative planning, particularly with respect to the use of prehabilitation,” the investigators wrote. Their report is in the Journal of the American College of Surgeons. They noted that while prehabilitation has been shown to reduce postoperative risk among those who need it, identification of these patients is an area that needs improvement.
With this in mind, Dr. McKenna and colleagues analyzed 205,840 major oncologic operations, with data drawn from the American College of Surgeons National Surgical Quality Improvement (NSQIP) database.
The researchers evaluated patients’ nutritional status using three techniques: the NSQIP method, the European Society for Clinical Nutrition and Metabolism (ESPEN) definitions, and the World Health Organization body mass index (BMI) classiﬁcation system.
Combining these three assessments led to seven hierarchical nutritional status categories:
- Severe malnutrition – BMI less than 18.5 kg/m2 and greater than 10% weight loss
- ESPEN 1 – BMI 18.5-20 kg/m2 (if younger than 70 years) or less than 22 kg/m2 (if 70 years or older) plus greater than 10% weight loss
- ESPEN 2 – BMI less than 18.5 kg/m2
- NSQIP – BMI greater than 20 kg/m2 (if younger than 70 years) or 22 kg/m2 (if 70 years or older) plus greater than 10% weight loss
- Mild malnutrition – BMI 18.5-20 kg/m2 (if younger than 70 years) or less than 22 kg/m2 (if 70 years or older)
- Obese – BMI at least 30 kg/m2
- No malnutrition.
The study’s primary outcomes were 30-day mortality and 30-day morbidity. The latter included a variety of complications, such as deep incisional surgical site infection, septic shock, and acute renal failure. Demographic and clinical factors were included in multivariate analyses.
Most of the operations involved patients with colorectal cancer (74%), followed by pancreatic (10%), lung (9%), gastric (3%), esophageal (3%), and liver (2%) cancer.
Across all patients, 16% fell into one of five malnutrition categories: mild malnutrition (6%), NSQIP (6%), ESPEN 2 (2%), ESPEN 1 (1%), or severe malnutrition (0.6%). The remainder of patients were either obese (31%) or had normal nutritional status (54%).
Malnutrition was most common among patients with pancreatic cancer (28%) and least common among those with colorectal cancer (14%).
Aligning with previous research, this study showed that nutritional status was associated with postoperative risk. Mortality risk was highest among patients with severe malnutrition, and morbidity was most common in the severe and ESPEN 1 groups (P less than .0001 for both).
While the spectrum of classifications appeared accurate across the population, multivariable models for mortality and morbidity revealed an interaction between cancer type and malnutrition definition (P less than .0001 for both), which suggested the most accurate definition of malnutrition differed from one type of cancer to another.
Specifically, a classification of severe malnutrition was most predictive of mortality among patients with esophageal or colorectal cancer. ESPEN 1 was most predictive of mortality for patients with gastric or lung cancer, and NSQIP was most predictive for those with liver cancer.
For predicting morbidity, severe malnutrition was most accurate among patients with colorectal cancer, whereas ESPEN 1 was better suited for gastric and lung cancer.