Clinical use of the respiratory quotient obtained from indirect calorimetry.

Abstract

The respiratory quotient (RQ) obtained from indirect calorimetry (IC), defined by the ratio carbon dioxide production (VCO2)/oxygen consumption (VO2), is affected by extremes of substrate use by the body. Underfeeding, which promotes use of endogenous fat stores, should cause decreases in the RQ, whereas overfeeding, which results in lipogenesis, should cause increases in the RQ. Marked increases in VCO2 (with subsequent increases in RQ) in response to overfeeding may cause respiratory compromise in patients with limited pulmonary reserve. Thus, variation in the RQ in response to the feeding regimen may indicate inappropriate feeding and serve as a marker for patient intolerance. This prospective, multicenter study was designed to determine the clinical use of RQ for monitoring adequacy and tolerance of nutrition support. Patients in any 1 of 30 long-term acute care Kindred hospitals made nil orally (NPO) and placed on total parenteral or enteral feeding were eligible for this study. Arterial blood gas, serum ketones, 24-hour collection of urine urea nitrogen, and IC measurements were obtained on all. Actual volume of enteral/parenteral feeding infused over the 24 hours before performance of IC was documented. A total of 263 patients (mean age, 70.2 years, 57.4% male) were entered in the study. Of the 263 study patients, 88.6% required mechanical ventilation, and 92.0% received enteral tube feeding only. Overall, 41.5% of patients were overfed, receiving >110% of required calories, whereas 34.2% were underfed, receiving <90% of required calories. The ratio of calories provided/required correlated significantly with overall measured RQ (p < .0001; R2 = .16). Correcting for the metabolism of protein by calculating a nonprotein RQ (NPRQ) from a 24-hour urine urea nitrogen did not improve this correlation (p < .0001, R2 = .32). Using a measured NPRQ >1.0 to identify overfeeding had an acceptable specificity of 85.1% but a low sensitivity of 38.5%. Similarly, use of a NPRQ <0.85 to determine underfeeding had a specificity of 72.2% and a sensitivity of 55.8%. Comparing the measured NPRQ with a predicted reference RQ (based on percent infusion of carbohydrate/fat and the Lusk table) did not improve the overall use of RQ. In the majority of patients (67.7%), comparison of the measured NPRQ to the predicted value failed to differentiate appropriate (meeting 100% +/- 10% of requirements) from inappropriate feeding. Deviation of the measured NPRQ from predicted failed to identify factors unrelated to substrate use purported to affect the RQ (such as acid/base disturbances or hyper/hypoventilation). Increasing measured RQ did correlate significantly with increasing respiratory rate (p = .002, R2 = .04) and decreasing tidal volume (p = .002, R2 = .04), suggesting reduced tolerance with development of shallow rapid respirations and ventilatory compromise. Although changes in the overall and nonprotein RQ correlate to percent calories provided/required, low sensitivity and specificity limit its efficacy as an indicator of over- or underfeeding. The RQ should not be used to finely adjust the nutrition support regimen. Elevation of overall measured RQ > or = 1.0 may be associated with reduced tolerance and mild respiratory compromise. The clinical use of RQ is limited to a marker of test validity (to confirm measured RQ values are in physiologic range) and a marker for respiratory tolerance of the nutrition support regimen.