The validity of three malnutrition screening markers among older patients with cancer.

BACKGROUND

Malnutrition is common in older adults with cancer and is associated with adverse clinical outcomes. We assessed and compared the validity of three tools commonly used to screen for malnutrition: The Mini Nutritional Assessment (MNA), weight loss and body mass index (BMI).

METHODS

In this retrospective study, we reviewed patients over age 65 with a diagnosis of cancer who were treated at the MD Anderson Cancer Center between 1 January 2013 and 31 March 2017. All patients in this study were evaluated by a trained geriatrician as part of a comprehensive geriatric assessment (CGA). Malnutrition was diagnosed by both CGA and clinical examination. The sensitivity, specificity and Cohen's κ of each tool was also compared with the clinical diagnosis.

RESULTS

A total of 454 older patients with cancer who had malnutrition information available were included in the analyses. The median age was 78%, and 42% (n=190) were clinically diagnosed with malnutrition at baseline. When the MNA was performed, 105 out of 352 patients (30%) were malnourished, and 122 (35%) at risk of malnutrition. Weight loss >3 kg was seen in 183 out of 359 (51%) patients, and BMI <20 kg/m2 was found in 30 of the 454 (7%) patients. MNA had the highest validity (area under curve (AUC)=0.83) and reliability (κ=0.67), weight loss had moderate validity (AUC=0.73) and reliability (κ=0.46), while BMI had the lowest validity (AUC=0.55) and reliability (κ=0.55).

CONCLUSIONS

For clinical practice, MNA should be incorporated for standard assessment/screening for these older patients with cancer.