Quick Links

Recommendations Summary

CI: Respiratory Quotient as a Method to Detect Measurement Error 2006

Click here to see the explanation of recommendation ratings (Strong, Fair, Weak, Consensus, Insufficient Evidence) and labels (Imperative or Conditional). To see more detail on the evidence from which the following recommendations were drawn, use the hyperlinks in the Supporting Evidence Section below.

  • Recommendation(s)

    CI: Respiratory Quotient

    If Respiratory Quotient  (RQ) is below 0.7 or above 1.0, then repeated measures are necessary under more optimal conditions. An RQ under 0.70 suggests hypoventilation (inadequate removal of metabolic carbon dioxide from the blood to the lung) or prolonged fasting. An RQ above 1.0, in the absence of overfeeding, suggests hyperventilation (removal of carbon dioxide from the blood to the lung in excess of the amount produced by metabolism) or inaccurate gas collection.

    Rating: Strong

    • Risks/Harms of Implementing This Recommendation

      No potential risks and harms are associated with the application of this recommendation.

    • Conditions of Application

      • Use rigorous adherence to gas collection device procedures to prevent air leaks
        • Air leaks will produce inaccurate RMR measures.
      • Indirect calorimetry equipment must be calibrated correctly and in good working order, based on manufacturer's instructions, for an accurate RQ measurement.

      Challenges to Implementation

      • The time required for machine calibration will add to the total time for RMR measurement
      • Ensuring availability of qualified staff (i.e., within scope of practice and trained to perform the procedure), to perform the measurement may be a barrier to implementing this recommendation.

    • Potential Costs Associated with Application

      If RQ is below 0.7 in the absence of prolonged fasting or if RQ is higher than 1.0 in the absence of overfeeding,  then a repeat measurement must be taken and the cost is in time.

    • Recommendation Narrative


      • Respiratory quotient is the ratio of vCO2 and vO2
      • Under typical metabolic conditions with stable respiratory function, the range of RQ in human metabolism is approximately 0.7 to 1.0, as reported by one positive quality prospective cohort (McClave et al, 2003),  one positive quality review (Peronnet and Massicotte, 1991),  four neutral quality reviews (Jequier et al,  1987; Brandi et al, 1997; Matarese, 1997; McClave et al, 2001),  one neutral quality meta-analysis (Elia and Livesey, 1992) and one neutral quality non-comparative descriptive study (Weissman et al,  1986)
      • Under atypical metabolic and respiratory conditions, RQ can be under 0.7 or over 1.0 and so RQ might aid in the assessment of the validity of some indirect calorimetry measurements of RMR (Brandi et al, 1989).

      Effect of Energy Consumption

      • A positive quality repeat-measures crossover study (Romyn et al,  1990) reported that prolonged fasting, recent or excessive food consumption and ethanol consumption prior to RMR measurement may affect RQ
        • Respiratory quotient ranged from 0.72 to 0.80 after 16 hours of fasting, but sometimes dropped below 0.70 in fasts lasting 22 hours (range 0.65 to 0.79).
      • Other neutral quality studies (one cohort study by Leff et al,  1987; one repeat-measures time series by Zauner et al,  2000; one repeat-measures crossover study by Johnstone et al,  2002), reported that group mean RQ remained above 0.70 in fasts lasting 36 to 84 hours (0.71±0.04), but based on group standard deviation, some individuals must have had an RQ of under 0.70 (possibly as low as 0.61)
      • Similarly, in trauma patients who fasted for 48 to 60 hours, group RQ remained above 0.7, but based on variation about the mean, some of the individual RQ values must have dropped below 0.7, as reported by two positive quality studies (repeat-measures crossover by Romyn et al,  1990; non-randomized trial by Jeevanandam et al,  1992) and four neutral quality studies (cohort study by Leff et al,  1987; case-control by Brandi et al,  1988; repeat-measures time series by Zauner et al,  2000; repeat-measures crossover by Johnstone et al,  2002).

      Food Consumption

      • Food consumption increases the RQ, depending on the amount and composition of the meal
        • A mean RQ of 0.86±0.11 was recorded 60 minutes after a mixed meal providing 25% of energy needs in one neutral quality repeated-measures crossover study (Saltzman, Roberts,  1996)
        • In three neutral quality studies (a repeat-measures crossover by Delafosse et al,  1997; an RCT by Frankenfield et al,  1997 and a retrospective cohort by Kan et al,  2003), post-operative and critically ill adults consuming adequate kcalories resulted in RQ values in the middle of the physiologic range. 

      Excess Energy Consumption

      • Excess kcalorie intake, on the other hand, more often raises the RQ above 1.0
        • A neutral quality two-factor repeated-measures crossover study reported that when measured 10 minutes after consuming about 1, 200kcal as a high-carbohydrate or a high-fat meal, mean RQ in healthy subjects was 1.0±0.04 and 0.98±0.04, respectively (Surina et al,  1993).
      • At 2.5 hours, the group mean RQ for the high-carbohydrate meal remained at 0.99±0.08, but for the high-fat meal, it had dropped to 0.90±0.08
      • In one neutral quality non-randomized trial,  25 nutrition support patients with glucose infusion rates in excess of seven mg/kg per minute, 19 (76%) RQ values were greater than 1.0 (Ireton-Jones and Turner,  1987)
      • Similarly, in a neutral quality retrospective cohort study, a group mean RQ of 1.16±0.32 was reported in 19 hemodynamically stable, mechanically ventilated individuals receiving kcals at 110% above measured RMR, supplied in various specialized nutrition support (Kan et al,  2003)
        • The standard deviations of these mean values indicate that a large portion, though not all, of the subjects exceeded RQ of 1.0.
      • In another neutral quality prospective clinical cohort study of septic and non-septic patients receiving 25% of measured RMR supplied with TPN, mean RQ was 0.86±0.05 (Zauner et al,  2001).

      Ethanol Consumption 

      • Two neutral quality reviews (Elia and Livesey,  1992; Matarese, 1997; Fung, 2000) reported the RQ of ethanol oxidation is 0.67 and so may be expected to lower RQ during a test of RMR
      • However, as the RQ measured in humans is a net value of all metabolism, one neutral quality repeated-measures crossover study reported ethanol consumption does not decrease the RQ below 0.7 (Weststrate et al,  1990).

      Respiratory Changes

      • Respiratory changes can cause artifact change in the RQ
        • In a negative quality case presentation, air leak led to RQ values ranging from 0.32 to 1.02 in the same subject (McCamish et al,  1981) 
        • In a neutral quality narrative review of stable critically ill patients, intentional increase in minute ventilation resulted in an increase in RQ to above 1.0, due to hyperventilation (excess removal of carbon dioxide via gas exchange) while intentional decrease in minute ventilation resulted in a decrease in RQ to less than 0.73, due to hypoventilation (accumulation of carbon dioxide in the bloodstream rather than removal via gas exchange) (Brandi et al, 1997).

      Test Interpretation

      • The RQ can serve as a tool to detect some inaccurate measurements or RMR protocol violations (Compher et al, 2006)
        • If the RQ is below 0.7 or above 1.0, prolonged starvation or excessive recent caloric consumption should be suspected 
        • If metabolic reasons for an aberrant RQ do not exist, then air leaks, hypoventilation, hyperventilation or inaccurate gas sensors should be suspected
        • If recalibration of the calorimeter does not result in an RMR test result with an RQ between 0.7 and 1.0 and similar results are obtained in a different test subject, the gas sensors should be tested with an alcohol burn test or the machine should be sent to the manufacturer for repair
        • Within the range of 0.7 to 1.0, typical variation in RQ is high enough that RQ cannot be used to detect protocol violations or inaccurate measurements.

    • Recommendation Strength Rationale

      • Many studies of strong design were evaluated and they essentially agreed on the findings outlined above
      • Conclusion statements are Grade II.