Prediction of the neurological outcomes post-cardiac arrest: A prospective validation of the CAST and rCAST





Abstract


Introduction


The neurologic prognosis of out-of-hospital cardiac arrest (OHCA) patients in whom return of spontaneous circulation (ROSC) is achieved remains poor. The aim of this study was to externally and prospectively validate two scoring systems developed by us: the CAST score, a scoring system to predict the neurological prognosis of OHCA patients undergoing targeted temperature management (TTM), and a simplified version of the same score developed for improved ease of use in clinical settings, the revised CAST (rCAST) score.


Methods


This study was a prospective, multicenter, observational study conducted using the SOS KANTO 2017 registry, an OHCA registry involving hospitals in the Kanto region (including Tokyo) of Japan. The primary outcome was favorable neurological outcome (defined as Cerebral Performance Category score of 1 or 2) at 30 days and the secondary outcomes were favorable neurological outcome at 90 days and survival at 30 and 90 days. The predictive accuracies of the original CAST (oCAST) and rCAST scores were evaluated by using area under the receiver operating characteristic curve (AUC).


Results


Of 9909 OHCA patients, 565 showed ROSC and received TTM. Of these, we analyzed the data of 259 patients in this study. The areas under the receiver operating characteristic curve (AUCs) of the oCAST and rCAST scores for predicting a favorable neurological outcome at 30 days were 0.86 and 0.87, respectively, and those for predicting a favorable neurological outcome at 90 days were 0.87 and 0.88, respectively. The rCAST showed a higher predictive accuracy for the neurological outcome as compared with the NULL-PLEASE score. The patients with a favorable neurological outcome who had been classified into the high severity group based on the rCAST tended to have hypothermia at hospital arrival and to not show any signs of loss of gray-white matter differentiation on brain CT. Neurological function at 90 days was correlated with the rCAST ( r = 0.63, p < 0.001).


Conclusions


rCAST showed high predictive accuracy for the neurological prognosis of OHCA patients managed by TTM, comparable to that of the oCAST score. The scores on the rCAST were strongly correlated with the neurological functions at 90 days, implying that the rCAST is a useful scale for assessing the severity of brain injury after cardiac arrest.


Highlights




  • •

    oCAST and rCAST scores are useful to predict with a high accuracy, the neurological outcome of PCAS patients treated by TTM.


  • •

    rCAST showed better predictive accuracy than the NULL-PLEASE score.


  • •

    The predictive accuracies of rCAST and oCAST scores for the neurological outcome were comparable.


  • •

    We observed a strong correlation between the score on rCAST and the neurological function at 90 days in PCAS patients.



1


Introduction


Out-of-hospital cardiac arrest is one of the leading causes of mortality in the field of emergency and critical care, occurring in an estimated 100,000 people each year in Japan [ ]. Even in patients showing return of spontaneous circulation (ROSC) and receiving neurocritical care, including targeted temperature management (TTM), most suffer from post-cardiac arrest syndrome (PCAS) and only a small percentage of cases are able to return into society [ , ].


Prediction of the neurological prognosis with high accuracy in the early phase after ROSC is important in PCAS patients, because it not only helps in providing objective information to the family during counselling, but also serves to guide both medical providers and family members in arriving at a joint decision about future treatment of the patients. Previously, using the data of 151 OHCA patients who underwent TTM, we developed the post-Cardiac Arrest Syndrome for Therapeutic hypothermia (CAST) score as a scoring system for predicting the neurological prognosis of post-cardiac arrest patients [ , ]. The CAST score is calculated based on the following eight variables, all of which are available before the initiation of TTM: the initial rhythm, time to ROSC from witness, blood pH, blood lactate levels, motor score on the Glasgow Coma Scale (GCS) at the time of ROSC, serum levels of albumin, serum levels of hemoglobin, and gray matter: white matter attenuation ratio (GWR) on brain CT [ ]. Furthermore, for ease of clinical use, we also developed a simplified version of the original CAST (oCAST) score, the revised CAST score (rCAST) [ ]. The rCAST score is calculated based on only 5 variables (initial rhythm, witness/ROSC time, lactate, pH, motor score on the GCS at the time of ROSC) rather than the 8 variables needed to calculate the oCAST score.


So far, most studies have reported high predictive accuracies of both the oCAST and rCAST scores for the neurological prognosis, but these validation studies have been retrospective [ ], and there was several concerns such as biases in retrospective population selection and data collection. There are no reports to date of a prospective external validation study conducted for the two scores. The aim of this study was to externally and prospectively validate the predictive ability of the oCAST and rCAST scores for the neurological prognosis in PCAS patients undergoing TTM.


2


Method


2.1


Study design


The study was a prospective, multicenter, observational study conducted using the SOS KANTO 2017 registry [ ]. SOS KANTO is an out-of-hospital cardiac arrest prospective registry established with the participation of 42 hospitals in the Kanto region (including Tokyo) of Japan. The need to obtain consent from individual patients for participation in the study was waived, as only anonymized information was handled. This study was conducted with the approval of the Institutional Review Boards of all participating hospitals. This study was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR) prior to the start of the registry (Registration number is UMIN000043930).


2.2


Patients


For this study, we analyzed the data of post-cardiac arrest patients registered in the SOS-KANTO registry between September 2019 and March 2021 who showed ROSC following OHCA and underwent TTM. Patients were excluded if they were under 18 years of age, had a traumatic cardiac arrest, were on dialysis, had dementia, had received ECMO, had not undergone CT imaging (data from brain CT is needed for calculation for the oCAST score).


2.3


Variables


The information prospectively collected by the SOS KANTO registry has been reported previously [ ]. In brief, the registry contains data on the patient characteristics, including the age, gender and Charlson Comorbidity Index, information about the circumstances of the cardiac arrest, including place of cardiac arrest, presence/absence of a witness/bystander, the initial electrocardiographic (ECG) rhythm at the time of the cardiac arrest, information on the various time pointers (such as time of emergency medical services call, time of initiation of cardiopulmonary resuscitation (CPR), time at hospital arrival, time of ROSC), clinical findings at hospital arrival, including the GCS score, body temperature and pupillary diameters, hospital examination findings, treatments administered, including extracorporeal membrane oxygenation, intra-aortic balloon pump, coronary angiography, percutaneous coronary intervention, and TTM, medications used, blood test results, and outcomes, including the Cerebral Performance Category (CPC) score and survival at 30 days and 90 days. There were some patients in whom the body temperature was unmeasurable at hospital arrival, and we regarded that the body temperature in these patients was too low to measure. We calculated the oCAST and rCAST scores of the patients according to the methods described in our previous reports [ , , ]. The formula is shown in Supplementary Fig. 1 . We also calculated the NULL-PLEASE score [ ], the other predictive score that has been used for post-cardiac arrest patients, with the aim of comparing the predictive accuracy of this score with that of our scores.


2.4


Outcome


The primary outcome was favorable neurological outcome at 30 days, defined as a CPC score of 1 or 2 [ ]. The secondary outcomes were favorable neurological outcome (CPC score ≤ 2) at 90 days, and survival at 30 and 90 days.


2.5


Statistical analysis


Continuous variables are reported as median values (with interquartile ranges) or as means ± standard deviation, and categorical variables are reported as n (%). Continuous variables were compared by the Mann-Whitney U test, and categorical variables were compared by Fisher’s exact test. The Cochran-Armitage test was performed to determine if there was a trend in the frequency of favorable neurological outcomes or survival in patients classified into the three severity scales based on the rCAST scores. To evaluate the predictive ability of the scoring system, we calculated the area under the curve (AUC) by plotting receiver operating characteristic (ROC) curves. As for the analysis for the outcome at 90 days, we could only include those patients for whom data on the outcome at 90 days were available. The Delong test was used to compare the AUCs among the different scoring systems.


The cutoff values used to calculate sensitivity, specificity, positive predictive value, negative predictive value for each scoring system were calculated based on Youden index. We also calculated the Spearman correlation ratio between CPC score at 90 days as an index of neurological function and the scores on the variables used to calculate rCAST using the data of patients who survived until 90 days. For this analysis, we excluded patients with CPC 5 at 90 days, because these patients had possibly died due to some hemodynamic reasons, and we considered it impossible to evaluate their neurological function. All analysis was performed using EZR version 4.1.2.


3


Results


3.1


Patient characteristics


A total of 9909 OHCA patients were registered in SOS KANTO 2017, and 565 showed ROSC and underwent TTM ( Fig. 1 ). Of these, 214 patients were excluded from this study for the following reasons: pediatric cardiac arrest ( n = 17); history of dementia ( n = 22); history of hemodialysis ( n = 16); traumatic cardiac arrest (n = 2); ECMO use ( n = 118); and brain CT not performed ( n = 39). Furthermore, 92 were excluded due to missing data on the outcomes and/or on the variables needed to calculate the rCAST score. Finally, 259 patients for whom the rCAST scores could be calculated were included in this study. Among the 259 patients, the oCAST score was available for 189 patients.




Fig. 1


Flow diagram for target patient enrollment.

HD, hemodialysis; ECMO, extracorporeal membrane oxygenation; CT, computed tomography; rCAST, revised version of the post-Cardiac Arrest Syndrome for Therapeutic hypothermia score; oCAST, original version of the post-Cardiac Arrest Syndrome for Therapeutic hypothermia score.


The patient characteristics are shown in Table 1 . The median age was 65.0 years (51.5–76.0), and 191 (73.7%) were male. The cardiac arrest status was as follows: presence of a witness; 181 (69.9%) cases; presence of a bystander(s), 142 (54.8%) cases; and a shockable initial rhythm, 113 (43.6%) cases. The median time to ROSC was 29 (21–40) mins. The survival rates at 30 and 90 days were 57.5% (149/259) and 46.9% (115/245), respectively. The percentages of patients with a favorable neurological prognosis at 30 and 90 days were 31.7% (82/259) and 31.8% (78/245), respectively.



Table 1

Characteristics of included patients.

























































Variables All patients ( n = 259)
Age, median (IQR) 65.0 (51.5–76.0)
Sex, Male, n (%) 191 (73.7)
Witness, n (%) 181 (69.9)
Bystander, n (%) 142 (54.8)
Initial rhythm, shokable, n (%) 113 (43.6)
Time to ROSC, min, median (IQR) 29.0 (21.0–40.0)
PCI, n (%) 67 (25.9)
pH, mean ± SD 7.03 ± 0.22
Lactate, mmol/L, mean ± SD 10.3 ± 7.29
GCS M > 1, n (%) 27 (10.4)
Outcome at 30 days
Survival, n (%) 149 (57.5)
Favorable neurological function, n (%) 82 (31.7)
Outcome at 90 days ⁎
Survival, n (%) 115 (46.9)
Favorable neurological function, n (%) 78 (31.8)

IQR, interquartile range; ROSC, return of spontaneous circulation; PCI, percutaneous coronary intervention; SD, standard deviation; GCS, glasgow coma scale.

⁎ Only 245 patients for whom outcome at 90 days were available were analyzed.



3.2


Predictive accuracies of the oCAST and rCAST scores


The AUCs of the oCAST score for a favorable neurological outcome and survival at 30 days were 0.86 (0.80–0.92) and 0.75 (0.68–0.82), respectively ( Fig. 2 A ), while those for a favorable neurological outcome and survival at 90 days were 0.87 (0.82–0.93) and 0.83 (0.77–0.89), respectively ( Fig. 2 B).




Fig. 2


Receiver operating characteristic curve analysis of the oCAST and rCAST scores for predicting a favorable neurological outcome. A, Receiver operating characteristic curve of the original CAST score for predicting a favorable neurological outcomes and survival at 30 days. B, Receiver operating characteristic curve of oCAST for predicting a favorable neurological outcome and survival at 90 days. C, Receiver operating characteristic curve of rCAST for predicting a favorable neurological outcome and survival at 30 days. D, Receiver operating characteristic curve of rCAST for predicting a favorable neurological outcome and survival at 90 days. As for the curves for the outcomes at 90 days, we analyzed only 180 patients for the oCAST score and 245 for the rCAST score, for whom the outcome data at 90 days were available.

oCAST, original version of the post-Cardiac Arrest Syndrome for Therapeutic hypothermia score; rCAST, revised version of the post-Cardiac Arrest Syndrome for Therapeutic hypothermia score.


The AUCs of the rCAST score for a favorable neurological outcome and survival at 30 days were 0.87 (0.83–0.92) and 0.77 (0.71–0.83), respectively ( Fig. 2 C), while those for a favorable neurological outcome and survival at 90 days were 0.88 (0.84–0.93) and 0.83 (0.78–0.88), respectively ( Fig. 2 D). The correlation ratio between oCAST and rCAST scores was 0.79 (0.73–0.82) with a statistical significance ( p < 0.01).


The patients were classified into severity categories according to the rCAST score, as follows [ ]: low severity category, n = 22; moderate severity category, n = 106; and high severity category, n = 131. The rates of a favorable neurological prognosis at 30 and 90 days were 86.4% (19/22) and 85.7% (18/21) for the low severity group, 50.9% (54/106) and 53.5% (54/101) for the moderate severity group, and 6.9% (9/131) and 6.5% (8/123) for the high severity group, respectively. The survival rates at 30 and 90 days were 100% (22/22) and 100% (21/21) in the low severity group, 70.8% (75/106) and 65.3% (66/101) in the moderate severity group, and 39.7% (52/131) and 22.8% (28/123) in the high severity group, respectively. The Cochran-Armitage test showed that increase in the severity category based on the rCAST scores were associated with statistically significant decreases of both the rates of favorable neurological outcomes and survival at 30 and 90 days ( Table 2 ).


Tags:
Mar 29, 2024 | Posted by in EMERGENCY MEDICINE | Comments Off on Prediction of the neurological outcomes post-cardiac arrest: A prospective validation of the CAST and rCAST

Full access? Get Clinical Tree
Get Clinical Tree app for offline access