The authors reply
First, as arterial to venous carbon dioxide difference (AVCO2) is a surrogate of low cardiac output syndrome (LCOS), “poor outcome” in our study was a composite outcome of mortality and morbidity related to clinically important LCOS—with goal to retrospectively determine whether AVCO2 can estimate or predict LCOS-related morbidity. Prospective confirmation is required to determine whether AVCO2 truly measures real-time LCOS. We agree that therapy-based metrics such as Inotrope Score (IS) have limitations as measures of morbidity and disease severity. However, most cardiac intensive care physicians would agree that very high doses of the vasoactive medications comprising the IS are almost exclusively prescribed when hemodynamic support is needed to support significant cardiovascular dysfunction or LCOS. A priori, we defined “high” IS as greater than 15. Subsequently, Gaies et al (3) demonstrated in a prospective, multiinstitutional study that IS greater than or equal to 20 is strongly and significantly associated with poor outcomes in infants after cardiac surgery. As stated in our article, a patient could have more than one bad outcome and although seven of our patients met IS greater than 15 inclusion criteria, only two of them had that criteria alone; further, both of those subjects had IS greater than 20—even without the addition of milrinone (which would have added minimum five more points to the IS). When these two patients were removed from our analysis, the significant association with poor outcomes remains (median AVCO2, 8.3 mm Hg in patients with poor outcomes vs 5.5 mm Hg in patients without poor outcomes; p = 0.002). Therefore, with conceded limitations detailed in our article, we stand by the validity of our definition of “poor outcome.”
Second, as mentioned in our article, we are in agreement with Inata et al (1) that the findings of our study may not be generalizable to older populations with cardiac disease or similar cohorts at different institutions. We specifically chose neonates less than 90 days old in order to capture more complex surgeries with higher frequency of LCOS and morbidity, and thus increased proportion of patients in which AVCO2 monitoring may have clinical utility.
Third, study inclusion criteria required an admission venous blood gas, which at our institution was drawn by protocol in patients with higher Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery Congenital Heart Surgery Mortality Categories category and as clinically indicated in unstable patients. Thus by definition, there was a selection bias toward higher risk patients, and therefore mortality will be high in our study cohort. The overall unadjusted mortality at our program during the study period (1.95%) is below the recent reported Society of Thoracic Surgeons (STS) aggregate (3.7%); furthermore, our infant mortality (0.79%) and neonatal mortality (6.1%) are below the STS unadjusted aggregate mortality (3% and 10.1%) (4). Therefore, although we may not be a statistical outlier with respect to increased postsurgical mortality as suggested by Inata et al (1), we concede that AVCO2 may have decreased utility in programs or patient populations with very low mortality and/or morbidity.
Fourth, we found no clear cutoff for AVCO2 to predict poor outcomes. However, this does not weaken the strength of association of high AVCO2 with poor outcomes. Small sample size, variability of the data, and the retrospective nature of this study all may have contributed to this fact. Additionally, we presume high AVCO2 estimates LCOS, and it is probable that not all LCOS leads to downstream morbidity.