Nutritional risk in hospitalized elderly patients

We have read the original article by López-Gómez et al.¹ on nutritional risk of elderly patients, and want to make some practical comments on it. From a methodological viewpoint, the study shows a clear bias in patient selection, because it only included patients for whom their treating physicians had requested nutritional assessment and support. Because of this, most patients enrolled were malnourished or at nutritional risk (96.5% using the Nutritional Risk Index [NRI] and 81.3% using the Geriatric Nutritional Risk Index [GNRI]) and were at a moderate or severe risk (94.7% with the NRI and 60.9% with the BNRI).¹ Such bias probably was one of the reasons why no correlation was found between these indices and mean stay or nutrition duration. On the other hand, it would be interesting to know in how many patients actual weight was measured and in how many their weight was estimated for body mass index (BMI) of 22 kg/m², because several estimates (rather than actual data) may be used to calculate GNRI, including estimation of actual weight on the one hand and estimation of ideal weight using the Lorentz formula (not adapted to the Spanish population) on the other, which may have partially accounted for the apparent lack of value of the GNRI in the study sample.

The availability of nutritional risk assessment tools such as NRI and GNRI, based on only 3 parameters such as albumin, actual weight, and standard or ideal weight, is highly attractive, first of all because of its fast implementation (as compared to the 5–15 min required for the Mini Nutritional Assessment [MNA]),² the few parameters to be recorded, and its easy completion. Such indices have, however, limitations for the nutritional assessment of elderly inpatients: (1) the NRI is valid for the detection of surgical risk in the young, rather than in elderly, patients.³ (2) It is not always easy to measure the current weight of acute bedridden patients. (3) Both risk indices include albumin, which is not the best nutritional parameter in elderly patients because their levels decrease 0.8 g/dL by decade from 60 years of age, may be altered by non-nutritional causes (acute disease, sarcopenia, renal, hepatic or heart failure, nephrotic syndrome, etc.), and are not an early marker of visceral protein reserve⁴ (MNA detects malnutrition risk before significant weight or albumin changes occur).² (4) NRI and GNRI have not been recommended for nutritional risk screening by scientific societies such as the European Society of Parenteral and Enteral Nutrition (ESPEN).⁵

MNA has a high sensitivity (96%) and specificity (98%) and good inter-observer agreement, and is not only valid in a community setting or in nursing homes and long-term facilities, as stated by the authors, but its use has been extended to all kinds of care environments (including primary, home, and hospital care).^6,7 In addition, MNA does not lose efficacy in elderly patients admitted to hospital or those with acute disease. In fact, low MNA scores in inpatients predict for an unfavorable course (prolonged stay, adverse effects, institutionalization after discharge, increased mortality).⁷ The problem with MNA is that it is less applicable in inpatients as compared to applicability of other nutritional screening methods (66.1% in the Bauer et al. study⁸) because it requires four anthropometric measures (weight, height, arm and calf circumference) and a degree of cooperation from the patient and/or relatives that is sometimes difficult to obtain in acute patients.²

We do not think that the significant correlation of MNA with NRI and GNRI is due to the terms they share, because weight (used to calculate BMI) is the only one of the 18 items of the MNA also used to calculate NRI and GNRI.

From the viewpoint of daily practice in a hospital ward, we doubt that additional calculation of NRI or GNRI provides supplemental information to MNA. If a MNA score ranging from 17 to 23.5 (malnutrition risk) or less than 17 (malnutrition) is found, a detailed dietary history should be obtained, measures should be taken to improve nutritional status (increased energy intake, nutritional supplements, water intake, etc.), and referral to the nutrition department should be considered. In the event of malnutrition (MNA less than 17), other causes of malnutrition should also be investigated (increased metabolic requirements, disease, etc.) and prompt nutritional intervention should be started.⁴ If MNA already assesses nutritional risk and malnutrition, allows for nutritional intervention aimed at MNA areas with lower scores, and is useful for following the course of nutritional status,^6,7 it is not clear what helpful information regarding nutritional screening and patient intervention is added by the calculation of another supplemental nutritional risk marker such as NRI or GNRI.

Because of the care burden in hospitalization areas, we think that it is more practical to use a single nutritional screening tool validated for that care setting (hospital) and age group (elderly patients), which may be applied with the available means, and recommended by scientific societies (MNA or Nutritional Risk Score, NRS-2002).⁵