Discussion

This study is the first to date to examine the outcomes of paediatric cardiac patients requiring cardiopulmonary resuscitation for more than 30 minutes until extracorporeal membrane oxygenation initiation. The primary finding of our study is that our cohort of patients had an overall survival to hospital discharge of 43.8%, with 75% of these patients having either normal neurologic function or only mild disability. Our reported survival is comparable to previous paediatric cardiac extracorporeal cardiopulmonary resuscitation studies that reported a survival of 20–46%,^{Reference Morris, Ittenbach and Godinez2,Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Shah, Shankar and Churchwell18–Reference Guo, Yang and Zhang24} but differs from other studies with more favourable extracorporeal cardiopulmonary resuscitation survival rates of 51–79%.^{Reference Kane, Thiagarajan and Wypij6,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25–Reference Shakoor, Pedroso, Jacobs, Okochi, Zenilman, Cheung and Middlesworth30} Many prior studies of extracorporeal cardiopulmonary resuscitation include patients with wide ranges of duration of chest compressions prior to cannulation, some of them including patients receiving less than 10 minutes of cardiopulmonary resuscitation prior to placement on mechanical support.^{Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6,Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Shah, Shankar and Churchwell18–Reference Shakoor, Pedroso, Jacobs, Okochi, Zenilman, Cheung and Middlesworth30} Our focus on those patients receiving prolonged cardiopulmonary resuscitation provides information about a more discrete and potentially higher risk population.

We focused on patients with primary cardiac disease because prior studies demonstrated increased survival with extracorporeal cardiopulmonary resuscitation and conventional cardiopulmonary resuscitation compared to non-cardiac patients.^{Reference Morris, Ittenbach and Godinez2–Reference Tajik and Cardarelli4,Reference Kane, Thiagarajan and Wypij6,Reference Matos, Watson and Nadkarni31,Reference Ortmann, Prodhan and Gossett32} Proposed mechanisms for this cardiac survival benefit are the lower cardiopulmonary reserve and baseline hypoperfusion which may lead to less arrest-related multi-organ dysfunction,^{Reference Morris, Ittenbach and Godinez2–Reference Tajik and Cardarelli4,Reference Kane, Thiagarajan and Wypij6,Reference Matos, Watson and Nadkarni31} or to the readily available deployment of extracorporeal cardiopulmonary resuscitation in the cardiac ICU.^{Reference Tajik and Cardarelli4,Reference Conrad and Rycus33} We compared our findings with previous literature that only focus on paediatric cardiac extracorporeal cardiopulmonary resuscitation or paediatric cardiac extracorporeal membrane oxygenation with subgroup analysis of extracorporeal cardiopulmonary resuscitation patients.^{Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6,Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Shah, Shankar and Churchwell18–Reference Shakoor, Pedroso, Jacobs, Okochi, Zenilman, Cheung and Middlesworth30} The Supplementary Table provides a review of paediatric extracorporeal cardiopulmonary resuscitation studies in the cardiac ICU literature. Several factors may account for differing results between our study and previous ones. First, several of these studies concentrated on different populations, such as post-cardiotomy or cath lab patients.^{Reference Kumar, Zurakowski and Dalton11,Reference Shah, Shankar and Churchwell18,Reference Alsoufi, Al-Radi and Gruenwald20,Reference Erek, Aydın and Suzan22,Reference Guo, Yang and Zhang24,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25,Reference Polimenakos, Rizzo, El-Zein and Ilbawi29} Second, others relied on subgroup analysis of larger reviews of cardiac extracorporeal membrane oxygenation support^{Reference Morris, Ittenbach and Godinez2,Reference Kumar, Zurakowski and Dalton11,Reference Shah, Shankar and Churchwell18,Reference Alsoufi, Al-Radi and Gruenwald20,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25,Reference Thourani, Kirshbom and Kanter26} rather than our study, with its specific focus on cardiac extracorporeal cardiopulmonary resuscitation. Third, our cohort focused on prolonged extracorporeal cardiopulmonary resuscitation and as such had a longer median duration of cardiopulmonary resuscitation (56 minutes) compared to the previous cardiac extracorporeal cardiopulmonary resuscitation reports where median duration ranged from 16 to 47.5 minutes^{Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6,Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Shah, Shankar and Churchwell18–Reference Alsoufi, Al-Radi and Gruenwald20,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25–Reference Polimenakos, Rizzo, El-Zein and Ilbawi29} with only four studies reporting a prolonged median duration similar to our study.^{Reference Huang, Wu and Chen21,Reference Guo, Yang and Zhang24,Reference Thourani, Kirshbom and Kanter26,Reference Shakoor, Pedroso, Jacobs, Okochi, Zenilman, Cheung and Middlesworth30} This factor would be expected to portend worse survival to our patients as prolonged cardiopulmonary resuscitation duration has been noted to be a predictor of mortality in paediatric cardiac patients undergoing extracorporeal cardiopulmonary resuscitation.^{Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Erek, Aydın and Suzan22}

In our study, we observed moderate to severe neurological disability (Pediatric Cerebral Performance Categories 3 and 4) in 25% of survivors. We also observed a strong correlation between poor neurological outcomes and evidence of injury on brain MRI (87.5% of survivors with moderate to severe neurological disabilities had severe hypoxic-ischemic injury on brain MRI after decannulation). Data on neurological outcomes of cardiac extracorporeal cardiopulmonary resuscitation survivors are limited and based on different diagnostic modalities.^{Reference Joffe, Lequier and Robertson1,Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Erek, Aydın and Suzan22,Reference Philip, Burgman and Bavare23,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25,Reference Polimenakos, Rizzo, El-Zein and Ilbawi29} A meta-analysis by Joffe et al described 9 paediatric extracorporeal cardiopulmonary resuscitation studies with a total of 287 survivors and a 22% prevalence of neurological disability.^{Reference Joffe, Lequier and Robertson1} Four of these studies exclusively evaluated cardiac extracorporeal cardiopulmonary resuscitation patients and reported variable neurological outcomes using different assessment tools: Morris et al and Allan et al demonstrated moderate clinical neurological deficits in 29 and 20% of patients, respectively, while Kane et al and Sivarajan et al described Pediatric Cerebral Performance Categories scores of 3–4 in 21 and 67% of survivors, respectively.^{Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12,Reference Allan, Thiagarajan, Armsby, del Nido and Laussen25} Philip et al reported Pediatric Overall Performance Category scores of 3–4 (correlating with moderate to severe neurological impairment) in 32% of their cardiac extracorporeal cardiopulmonary resuscitation survivors. Interestingly, 75% of these patients with Pediatric Overall Performance Category 3–4 had cardiopulmonary resuscitation duration to extracorporeal membrane oxygenation initiation >30 minutes.^{Reference Philip, Burgman and Bavare23} Other studies demonstrated lower incidence of severe neurologic deficits (18^{Reference Huang, Wu and Chen21} and 10%^{Reference Wolf, Kanter, Kirshbom, Kogon and Wagoner28}) in their cardiac extracorporeal cardiopulmonary resuscitation survivors but did not clearly identify their diagnostic tools.^{Reference Erek, Aydın and Suzan22,Reference Polimenakos, Rizzo, El-Zein and Ilbawi29}

Our study describes several risk factors associated with increased mortality in our cohort. Cardiopulmonary resuscitation lasting greater than 50 minutes was found to be a predictor of mortality. Three of our patients survived with mild neurological deficits (Pediatric Cerebral Performance Categories ≤ 2) after very prolonged cardiopulmonary resuscitation episodes (118, 134, and 139 minutes) before extracorporeal membrane oxygenation cannulation. Data on duration of cardiopulmonary resuscitation before extracorporeal membrane oxygenation initiation as a predictor for survival in paediatric cardiac patients are conflicting. While some investigators identified cardiopulmonary resuscitation duration as a predictor of mortality,^{Reference Kumar, Zurakowski and Dalton11,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12} others did not find this association.^{Reference Morris, Ittenbach and Godinez2,Reference Kane, Thiagarajan and Wypij6} Two previous paediatric studies demonstrated “adequate” neurological outcomes after cardiopulmonary resuscitation episodes lasting >90 minutes before extracorporeal membrane oxygenation initiation.^{Reference Raymond, Cunnyngham and Thompson5,Reference Kelly, Porter, Meier, Myers and Thomas14} However, the upper limit of cardiopulmonary resuscitation duration before extracorporeal membrane oxygenation cannulation resulting in acceptable neurological outcomes continues to remain unknown.

Our findings demonstrate increased mortality when chest compressions were paused >6 times during cannula insertion. The American Heart Association 2015 guidelines for “excellent cardiopulmonary resuscitation” include as a quality metric for improved survival the reduction of pauses during chest compressions.^{Reference Atkins, Berger and Duff34} Adult cardiopulmonary resuscitation studies have demonstrated that minimising interruptions during chest compressions is associated with increased return of spontaneous circulation and survival.^{Reference Christenson, Andrusiek and Everson-Stewart35,Reference Gundersen, Kvaløy, Kramer-Johansen, Steen and Eftestøl36} We believe this is the first paediatric study demonstrating reduced interruptions of chest compression during extracorporeal membrane oxygenation cannulation being associated with survival.

Additionally, we observed that calcium chloride or gluconate were more frequently used among non-survivors during resuscitation. This finding could be related to patients with poor cardiovascular function requiring higher calcium levels to maintain cardiac output, or to higher calcium levels damaging myocardium and impairing heart function until cannulation. Calcium is essential for myocardial excitation–contraction coupling, and as such for myocardial contractility, especially in neonates due to limited intracellular stores.^{Reference Kimura, Iwasaki and Oe37,Reference Srinivasan, Morris, Helfaer, Berg and Nadkarni38} However, calcium also has detrimental effects on myocardial function after ischemia-reperfusion injury (such as during bypass and cardiopulmonary resuscitation), likely due to the reduced capacity of the injured myocytes to handle intracellular calcium load.^{Reference Kimura, Iwasaki and Oe37,Reference Srinivasan, Morris, Helfaer, Berg and Nadkarni38} Our study adds data to the previous literature demonstrating increased mortality with calcium use during paediatric cardiopulmonary resuscitation.^{Reference De Mos, Van Litsenberg, McCrindle, Bohn and Parshuram7,Reference Sivarajan, Best, Brizard, Shekerdemian, d’Udekem and Butt12}

Finally, we demonstrated that both the need for continuous renal replacement therapy and abnormal pupillary response during extracorporeal membrane oxygenation predicted mortality in our cohort. Previous studies have shown that renal and neurological injury, as part of post-arrest multi-organ dysfunction, are important predictors for poor outcomes in children undergoing extracorporeal cardiopulmonary resuscitation.^{Reference Thiagarajan, Laussen, Rycus, Bartlett and Bratton3,Reference Tajik and Cardarelli4,Reference Kane, Thiagarajan and Wypij6,Reference Chan, Thiagarajan, Frank and Bratton19,Reference Huang, Wu and Chen21,Reference Shakoor, Pedroso, Jacobs, Okochi, Zenilman, Cheung and Middlesworth30} We did not find a relationship between abnormal neuroimaging during extracorporeal membrane oxygenation and mortality in our cohort; however, abnormal pupillary response had a significant strong correlation with poor outcome.

Our study has a few limitations. First, it is a retrospective database review of a single-centre experience with a limited sample, so the results cannot be extrapolated to all extracorporeal cardiopulmonary resuscitation patients. Second, we were not able to adjust for factors associated with mortality, such as severity of illness, comorbidities, and extracorporeal membrane oxygenation-related complications. Third, the absence of data on quality of cardiopulmonary resuscitation (specifically, lack of end tidal carbon dioxide data and time to first dose of epinephrine), as well as absence of standardised extracorporeal cardiopulmonary resuscitation guidelines in our patients, may be an important confounder. Finally, our neurodevelopmental outcome assessments had a limited follow-up to hospital discharge. These limitations leave us unable to draw conclusions regarding the long-term impact of prolonged extracorporeal cardiopulmonary resuscitation in paediatric cardiac patients.

In conclusion, survival after prolonged extracorporeal cardiopulmonary resuscitation in our paediatric cardiac patients was 43.8%, with 75% of these patients having either normal neurologic function or only mild disability. Mortality is associated with increased use of calcium during resuscitation, longer cardiopulmonary resuscitation duration, increased number of interruptions of chest compressions during cannulation, the need for continuous renal replacement therapy, and abnormal pupillary response during extracorporeal membrane oxygenation support. Prospective studies evaluating the factors associated with prolonged extracorporeal cardiopulmonary resuscitation in paediatric cardiac patients may help further define these relationships.