The Fick principle (cardiac output [Q c ] = oxygen uptake [V o 2 ]/arteriovenous oxygen difference) can be used to calculate Q c , with VO 2 frequently estimated by derived equations. To compare the accuracy of measured versus estimated VO 2 , data were analyzed from 2 studies in which VO 2 at rest was measured using the Douglas bag technique. One study comprised adults with diabetes, and the other was an exercise study of healthy adults. VO 2 at rest was estimated as VO 2 (ml/min) = 125 ml/min/m 2 × body surface area (m 2 ), with sensitivity analyses evaluating 2 other commonly used equations. Mean absolute difference (milliliters per minute) and ordinary least products regression were used to assess agreement between measured and estimated VO 2 . Overall, mean measured versus estimated VO 2 differed significantly (307.2 ± 75.2 vs 259.9 ± 36.7 ml/min, p <0.0001), with a mean absolute difference of 52.9 ± 43.2 ml/min (p <0.0001); 20% of the estimates differed by >25% from the measured VO 2 . Mean absolute difference increased from 36.7 ml/min in the lowest body mass index group (<25 kg/m 2 ) to 91.7 ml/min in the highest group (≥40 kg/m 2 ) (p for trend = 0.001) and was significantly higher in men than in women (65.6 vs 33.9 ml/min, p = 0.001); error was similar by median-split age (p = 0.65) and race (p = 0.34). Similar results were obtained when evaluating each of the other 2 estimating equations. Estimation of VO 2 at rest is inaccurate, especially in men and with increasing adiposity. In conclusion, when clinical hemodynamic assessment is performed, VO 2 should be measured, not estimated.
Accurate determination of cardiac output is essential for hemodynamic monitoring in the intensive care unit and for the assessment of severity of heart failure, pulmonary arterial hypertension, and valvular heart disease in the cardiac catheterization laboratory. In contemporary clinical practice, cardiac output can be calculated by thermodilution, dye dilution, and the Fick method (cardiac output [Qc] = oxygen uptake [VO 2 ]/arteriovenous oxygen difference, with the Fick method considered the gold standard. Although the Fick method is predicated on direct measurement of VO 2 , in practice, VO 2 is more commonly estimated using derived equations, because of the complexity of measurement with either mass spectrometry analysis of timed Douglas bag collections of exhaled air or breath-by-breath analysis of exhaled air using a metabolic cart. However, these estimating equations have been statistically derived from relatively small and highly selected cohorts, and their accuracy applied to general clinical practice has not been rigorously assessed. We sought to assess the overall accuracy of estimation of VO 2 at rest compared to direct measurement using gold-standard techniques in a contemporary, ethnically diverse, adult patient population with a high prevalence of obesity, with further analyses stratified by body mass index (BMI), gender, age and race.
Methods
Baseline data were analyzed from participants in 2 previous clinical studies in 2005 and 2006 in which baseline demographic, anthropometric, and clinical characteristics were recorded and VO 2 at rest was directly measured. One study included patients with type 2 diabetes with prevalent atherosclerotic vascular disease or additional cardiac risk factors, and the other study included lean and obese subjects without documented cardiovascular disease participating in an exercise intervention study. The 2 study protocols were approved by the University of Texas Southwestern Medical School at Dallas Institutional Review Board, and all participants provided written informed consent.
VO 2 at rest was estimated according to the derived formula commonly used in our institution’s (Parkland Hospital) cardiac catheterization laboratories: VO 2 (ml/min) = 125 ml/min/m 2 × body surface area (m 2 ), with body surface areaw calculated according to the formula of Dubois and Dubois as 0.007184 × weight (kg) 0.425 × height (cm) 0.725 . For sensitivity analyses, we also evaluated estimates of VO 2 at rest derived from 2 other commonly used equations.
VO 2 at rest was measured in all patients using the gold-standard technique of Douglas, analyzing 2 consecutive 3-minute collections of exhaled air through a mouthpiece fitted with a 3-way valve, with VO 2 at rest reported as the average of the 2 collections. Concentrations of oxygen, carbon dioxide, and nitrogen were determined by mass spectrometry (MGA 1100; Marquette Electronics, Milwaukee, Wisconsin), calibrated before every measurement, and volumes were determined using a Tissot spirometer. For the 27 subjects from the diabetes study, VO 2 at rest was obtained in a standing position before commencing maximal exercise treadmill testing. For the remaining 48 subjects, VO 2 at rest was recorded while subjects sat on a recumbent cycle ergometer before exercise testing.
The degree of agreement between directly measured and estimated VO 2 at rest was assessed by evaluation of mean absolute difference and ordinary least products regression analyses. Mean absolute difference (milliliters per minute) was calculated as an average of the absolute value of the raw difference between measured and estimated VO 2 determined for each subject. The degree of disagreement between measured and estimated VO 2 at rest was calculated as a percentage error, dividing the absolute difference by the corresponding measured oxygen uptake, and multiplying by 100. Ordinary least products regression was used to assess fixed and proportional bias for the overall cohort. Mean absolute differences were compared between measured and estimated VO 2 at rest, with statistical significance determined using 1-sample Student’s t tests, in the overall cohort, and in subjects stratified by (1) BMI using clinical categories (18 to 24.9, 25 to 29.9, 30 to 34.9, 35 to 39.9, and ≥40 kg/m 2 ), (2) gender, (3) age stratified by median split, and (4) race. Linear trend analysis (Jonckheere-Terpstra) for mean absolute difference across BMI categories, along with unpaired Student’s t tests for gender, age, and race were used as appropriate. All testing was 2 tailed at a significance level of 0.05, with analyses performed using SAS version 9.1.3 (Cary, North Carolina), and no corrections were made for multiple comparisons.
Results
Detailed patient characteristics for the overall cohort and selected strata are listed in Table 1 . Mean absolute difference by grouping is listed in Table 2 . Measured VO 2 for the overall cohort ranged from 170 to 510 ml/min, and using the estimating formula used at our institution’s cardiac catheterization laboratory, estimated VO 2 ranged from 175.3 to 339.6 ml/min. Overall, VO 2 was significantly underestimated by the estimating formula, with a mean absolute difference of 52.9 ± 43.2 ml/min (p <0.0001; Table 2 , Figure 1 ) . On the basis of the regression analysis ( Figure 1 ), for the few subjects with measured VO 2 ranging from 150 to 225 ml/min, the equation tended to overestimate VO 2 , while for those with measured VO 2 >250 ml/min, there was systematic underestimation using the equation. Further analysis of agreement in the overall cohort using ordinary least products regression methods demonstrated significant fixed error (reflected by the y intercept >0 [95% confidence interval 96.5 to 125.3]) and proportional error (reflected by the slope of <1.0 [95% confidence interval 0.43 to 0.56]) in the estimated VO 2 . The magnitude of the disagreement between measured and estimated VO 2 expressed as percentage error for the overall cohort is shown in Figure 2 . Nearly 50% of the cohort had a level of disagreement between measured and estimated VO 2 ranging from 10% to 25%, while 20% of the cohort had differences >25%. Sensitivity analyses of mean absolute difference using 2 other commonly used estimating equations are listed in Table 2 and yielded qualitatively similar results.
| Variable | Overall | Men | Women | BMI (kg/m 2 ) | ||||
|---|---|---|---|---|---|---|---|---|
| (n = 75) | (n = 45) | (n = 30) | 18–24.9 | 25–29.9 | 30–34.9 | 35–39.9 | ≥40 | |
| (n = 13) | (n = 12) | (n = 20) | (n = 12) | (n = 18) | ||||
| Age, median (years) | 39 | 42 | 33 | 29 | 47 | 41 | 42 | 42 |
| Height (in) | 67.4 ± 4.2 | 69.6 ± 3.3 | 64.1 ± 2.8 | 66.1 ± 2.8 | 69.0 ± 4.1 | 68.3 ± 4.4 | 66.4 ± 3.6 | 67.0 ± 5 |
| Weight (kg) | 97.2 ± 25.5 | 105.4 ± 24.8 | 84.9 ± 21.6 | 62.8 ± 10 | 81.5 ± 10 | 98.4 ± 13.3 | 106.1 ± 13.3 | 125.3 ± 20.3 |
| BMI (kg/m 2 ) | 33.0 ± 7.6 | 33.7 ± 7.3 | 32.1 ± 8.2 | 22.1 ± 2.2 | 26.5 ± 1.3 | 32.5 ± 1.3 | 37.2 ± 1.5 | 43.1 ± 2.1 |
| Women | 30 (40%) | — | — | 8 (62%) | 3 (25%) | 7 (35%) | 6 (50%) | 6 (33%) |
| Nonwhite | 41 (55%) | 22 (49%) | 19 (63.3%) | 4 (31%) | 5 (42%) | 14 (70%) | 8 (75%) | 10 (56%) |
| Hypertension | 19 (25%) | 12 (27%) | 7 (23.3%) | 0 | 6 (50%) | 3 (15%) | 5 (42%) | 5 (28%) |
| Hyperlipidemia | 17 (23%) | 12 (27%) | 5 (16.7%) | 0 | 5 (42%) | 3 (15%) | 3 (25%) | 6 (33%) |
| Prior cardiovascular disease | 8 (11%) | 7 (16%) | 1 (0.1%) | 0 | 3 (25%) | 3 (15%) | 0 | 2 (11%) |
| Diabetes | 27 (36%) | 20 (44%) | 7 (23.3%) | 0 | 6 (50%) | 6 (30%) | 5 (42%) | 10 (56%) |
| Duration of diabetes ⁎ , median (years) | 5.0 | 6.0 | 5.0 | 0 | 4.0 | 7.5 | 4.0 | 5.5 |
| Parkland Hospital Formula ⁎ | LaFarge Formula † | Bergstra Formula ‡ | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Group | Measured VO 2 (ml/min) | Estimated VO 2 (ml/min) | Absolute Difference | p Value | Estimated VO 2 (ml/min) | Absolute Difference | p Value | Estimated VO 2 (ml/min) | Absolute Difference | p Value |
| Overall | 307.2 ± 75.2 | 259.9 ± 36.7 | 52.9 ± 43.2 | <0.0001 | 247.6 ± 48.4 | 61.3 ± 42.2 | <0.0001 | 277.9 ± 43.1 | 42.2 ± 35.6 | <0.0001 |
| Men | 338.8 ± 72.8 | 275.7 ± 33.0 | 65.6 ± 47.8 | <0.0001 | 275.0 ± 37.6 | 64.2 ± 44.8 | <0.0001 | 292.9 ± 41.4 | 52.1 ± 40.4 | <0.0001 |
| Women | 259.8 ± 50.1 | 237.2 ± 28.8 | 33.9 ± 26.2 | <0.0001 | 204.6 ± 26.4 | 55.4 ± 32.6 | <0.0001 | 255.2 ± 35.5 | 27.5 ± 19.7 | <0.0001 |
| BMI (kg/m 2 ) | ||||||||||
| 18–24.9 | 247.7 ± 50.7 | 213.9 ± 20.7 | 36.7 ± 34.3 | 0.001 | 202.7 ± 34.1 | 46.0 ± 30.2 | <0.0001 | 225.5 ± 23.8 | 31.1 ± 29.7 | 0.001 |
| 25–29.9 | 271.5 ± 64.9 | 247.0 ± 23.0 | 41.5 ± 28.4 | 0.001 | 247.1 ± 22.9 | 41.9 ± 33.8 | 0.001 | 258.9 ± 29.9 | 34.8 ± 21.0 | 0.001 |
| 30–34.9 | 302.9 ± 43.9 | 265.6 ± 27.5 | 38.9 ± 38.8 | <0.0001 | 253.0 ± 40.2 | 50.7 ± 26.4 | <0.0001 | 284.2 ± 30.0 | 29.0 ± 25.6 | <0.0001 |
| 35–39.9 | 313.9 ± 55.2 | 268.2 ± 24.5 | 46.8 ± 38.5 | 0.002 | 248.8 ± 37.2 | 65.1 ± 36.9 | <0.0001 | 288.5 ± 39.4 | 33.7 ± 29.8 | 0.002 |
| ≥40 | 374.4 ± 86.7 | 289.9 ± 35.0 | 91.7 ± 52.4 | <0.0001 | 282.4 ± 52.7 | 94.6 ± 52.9 | <0.0001 | 314.2 ± 41.0 | 75.6 ± 41.7 | <0.0001 |
⁎ VO 2 (ml/min) = 125 × body surface area.
† VO 2 (ml/min) = 138.1 − (11.49 × log e [age]) + (0.378 × heart rate) × body surface area (men); VO 2 (ml/min) = 138.1 − (17.04 × log e [age]) + (0.378 × heart rate) × body surface area (women).
‡ VO 2 (ml/min) = 157.3 × body surface area + 10 − (10.5 × log e [age]) + 4.8 (men); VO 2 (ml/min) = 157.3 × body surface area − (10.5 × log e [age]) + 4.8 (women).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
