Retrospective Analysis of Arterial Carbon Dioxide Level and Arterial pH Level at the Time of Initiation of Respiratory ECMO and Outcome

Introduction

Utility of extracorporeal technologies in critical care unit is expanding, especially the continuous renal replacement therapy (CRRT) and respiratory extracorporeal membrane oxygenation (ECMO). Respiratory ECMO was initially limited to neonatal and pediatric population but after Cesar trial (2006) and pandemic of H1N1(2009) ECMO in adult cases also got the boost and in coronavirus disease 2019 (COVID-19) era it got well popularized. This can be well understood by growing number of total respiratory ECMO cases in Extracorporeal Life Support Organization (ELSO) registry. In April 2022 statistics has more than 88,000 respiratory ECMO cases (∼ 51% of total cases registered) with survival to discharge around 58 to 73%.¹ Many factors affect the outcome of ECMO, one among them is pre-ECMO pH and arterial partial pressure of carbon dioxide (paCO2) level. A significant proportion of patients has hypercapnia before ECMO initiation. EOLIA (Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Distress Syndrome) trial also had 47% of patients with a paCO2 of more than 55 mm Hg.² Schmidt et al in 2014 analyzed ELSO Registry data and found a higher pre-ECMO paCO2 was associated with increased mortality in patients receiving ECMO.³ Hypercapnia and respiratory acidosis are likely a marker of underlying disease severity. Our center planned to analyze our data to find out correlation of pre-ECMO paCO2 and pH level with outcome.

Method

We designed a retrospective observational study of all respiratory ECMO (neonatal, pediatric, and adult) patient managed by Riddhi Vinayak Multispecialty Hospital team from January 1, 2010 to December 31, 2021. We extracted data on all patients who received ECMO primarily for respiratory failure from our institutional ECMO database. All ECMO modalities were included: venovenous, venoarterial, or hybrid. We created two groups of survivor versus nonsurvivor which were compared by bivariate analysis with regard to demographics and clinical and ECMO course characteristics. Differences between the two groups were analyzed and p-value obtained by using chi-squared two variant calculators. p-Value less than 0.05 is considered significance. Data was analyzed for primary outcome goal of survival to discharge home and secondary outcome goal of number of ECMO days and the incidence of neurological complications. Data analysis was done with pre-ECMO paCO2 level more than 45 and less than 45 with primary and secondary outcome goal. Similarly, data was analyzed for pre-ECMO pH level less than 7.20 or more than 7.20.

Results and Analysis

Total patients who received respiratory ECMO in our institute during specified period were 256. We have subclassified (on the basis of diagnosis) the data in three broad categories of COVID-19, H1N1, and other causes of respiratory ECMO (non-COVID-19/ H1N1). As our primary goal was survival to discharge home, we have excluded five patients who have been shifted for transplant facilities. So, overall patients data analyzed was of 251. Results from our analyzed data are as follows.

• Out of 256 respiratory ECMO, 23% patients were COVID-19, 36% H1N1, and 41% non-COVID-19/ H1N1. Overall survival in respiratory ECMO was around 38% with maximum in non-COVID-19/H1N1 respiratory failure approximately 43% and least in COVID-19 approximately 24% (see Table 1).

• Most of the patient had type II respiratory failure (paCO2 > 45 in 82.5%) and mixed or predominant respiratory acidosis (pH < 7.4 in ∼ 84%) at the time of initiation of ECMO.

• Survival to discharge chances were better in group with paCO2 less than 45 (42.2 vs. 31.06); however, on statistical analysis it was not significant (p-value: 0.15) (see Tables 2 and 3). There was not much difference in average paCO2 of survival to discharge (63.3) and in nonsurvivor (66.8) (Fig. 1).

• Survival to discharge chances were better in group with pre-ECMO pH less than 7.2 (42.2 vs. 31.06) and were also statistically significant (p-value: 0.038) (Table 4).

• In meta-analysis, there was no difference in outcome of H1N1 patient with paCO2 level. However, in COVID-19 patient (25 vs. 11.11%) and non-COVID-19/H1N1 (54.54 vs. 36.14) had better survival with paCO2 less than 45 mmHg (see Table 2).

• There was no significant difference in number of ECMO days, in patient with paCO2 less than 45 or more than 45 (11.1 vs. 15.7 days) and similarly with pH of less than 7.2 or more than 7.2, also there was no significance difference (11.6 vs. 15.8 days)

• Total adverse neurologic events diagnosed by neuroimaging and clinical findings occurred in 28 of 251 (11.15%)patients, predominantly diagnosed was intracranial haemorrhage in 14 of 251 i.e. 5.57% of patients and PCO2 too remained high as shown in Figure 2 (Table 5). However, the true prevalence of neurologic injuries in survivors and nonsurvivors was unknown secondary to the elective nature of neuroimaging.

• Incidence of neurological complications was much low not only in pre-ECMO paCO2 of less than 45 (4.4 vs. 12.6) (Table 5) but also in pre-ECMO pH less than 7.2 (17.39 vs. 7.55, p-value 0.034) (Table 6).

Close

Fig. 1

Average pre-extracorporeal membrane oxygenation partial pressure of carbon dioxide in survival and nonsurvival. COVID-19, coronavirus disease 2019.

Export to PPT

Conclusion

In our analysis, we found most (> 80%) of the patients who required respiratory ECMO were in type II respiratory failure and in respiratory or mixed acidosis. These numbers are significantly higher as compared to the numbers in EOLIA trial (47% only). Also, both higher paCO₂ level and lower pH at the time of initiation of ECMO are related to higher mortality and risk of neurological complications. However, only pre-ECMO lower pH was found statistically significant for both mortality and neurological complications. So, early initiation of ECMO with pre-ECMO paCO2 less than 45 and pH more than 7.2 has better survival chances.

Discussion

Many of the times patients who require respiratory ECMO are hypercapnic and/or in acidosis. This is because of our routine ventilatory management practice of permissive hypercapnia. Respiratory acidosis and pre-ECMO elevations of paCO2 have been associated with poor outcomes from ECMO.⁴, ⁵ In older children, Mehta et al found that higher pH (pH > 7.2) before ECMO initiation was associated with increased survival.⁴ Also, pre-ECMO elevated paCO2 (>50 mmHg) is associated with ICH in neonates rescued with ECMO.⁵, ⁶ In other studies of pediatric ECMO, duration of mechanical ventilation and serum pH less than 7.29 before ECMO initiation has been associated with increased mortality.⁷

There is no clear mechanism for elevated paCO2 to “actively” contribute to increased mortality except through the physiologic effect of paCO2 on the brain. Rapid corrections of patient's paCO2 and pH could adversely affect cerebral blood flow and perhaps contribute to increased neurologic injury and poor outcome.⁸, ⁹, ¹⁰ Careful attention to the rate of correction of paCO2 has been advocated in these patients.

Summary

Our study has the limitations inherent to the retrospective design and single-center experience. But many other retrospective studies also have similar experience of poor outcome associated with pre-ECMO hypercapnia and acidosis (respiratory or mixed) in respiratory ECMO. Also, rapid corrections post-initiation of ECMO leads to neurological complications like IC bleed. Authors strongly recommend early initiation of ECMO and slow corrections paCO₂ post-initiation of ECMO for a better outcome.

Close

Fig. 2

Neurological complications and partial pressure of carbon dioxide level. COVID-19, coronavirus disease 2019.

Export to PPT