Modes of Death in Patients with Cardiogenic Shock in the Cardiac Intensive Care Unit: A Report from the Critical Care Cardiology Trials Network.
Feb 21, 2024DOI: 10.1016/j.cardfail.2024.01.012Journal ArticleJCF
Written by: David D BergSachit SingalMichael PalazzoloVivian M Baird-ZarsFadel BofarragErin A BohulaSunit-Preet ChaudhryMark W DodsonDustin HillersonPatrick R LawlerShuangbo LiuConnor G O'BrienBarbara A PisaniLekha RacharlaRobert O RoswellKevin S ShahMichael A SolomonLakshmi SridharanAndrea D ThompsonSean VAN DiepenJason N KatzDavid A MorrowCCCTN Investigators
Find these authors on Roon
SC
Dr. Sunit-Preet Chaudhry
Advanced Heart Failure · Ascension St. Vincent Hospital
CO
Dr. Connor O'Brien
Cardiac Critical Care · UCSF Medical Center
Methods
The Critical Care Cardiology Trials Network (CCCTN) is a multinational research network of CICUs with oversight by the Thrombolysis In Myocardial Infarction (TIMI) Study Group (Boston, MA). Methods for the CCCTN Registry have been described.3 From 2021 to 2022, 32 centers contributed data on consecutive medical CICU admissions during annual 2-month collection periods. In addition, year-round capture of consecutive admissions was permitted.The present analysis included patients with CS assessed by site investigators using standardized definitions.1 Patients with coronavirus disease 2019 were excluded to ensure that the results were not overly influenced by an intercurrent event that might be unique to a limited time period. CS etiology was subdivided into AMI-CS, HF-CS, and secondary (nonmyocardial) CS (eg, tamponade). Subgroups defined by CS type, cardiac arrest (CA) before CICU admission, use of tMCS, and transition to comfort measures only (CMO) were evaluated.For the annual collection cycle from October 2021 to September 2022, site investigators classified the dominant mode of in-hospital death for all patients with CS who died using standardized definitions (Table 1). These consensus definitions were developed by CCCTN leadership for prospective application in the registry. Rather than identifying the proximate cause of death, investigators were prospectively trained to identify the direct reason for death (i.e., mode of exit), which may have been different from the initial diagnosis or indication for CICU admission. Mutually exclusive categories included 4 modes of cardiovascular death (CS, arrhythmia, stroke, or other cardiovascular death) and 4 modes of noncardiovascular death (respiratory failure, anoxic brain injury, noncardiogenic shock, or other noncardiovascular death). Classification of shock type (ie, cardiogenic vs noncardiogenic) was based on previously reported criteria.1
Results
Among 1068 CS cases, 337 (31.6%) died during the index hospitalization with a median time to death of 4.4 days (Q1–Q3, 1.2–10.9 days) from CICU admission. Clinical characteristics are summarized in Table 2. One-third of patients with CS (32.9%) who died had CA before CICU admission, 39.2% were managed with tMCS, and 67.1% were transitioned to CMO. In addition, 22.0% evolved to a secondary shock state, including 11.6% who evolved to mixed shock (i.e., vasodilatory CS) and 7.4% who evolved to pure distributive shock.Across all CS deaths, the primary mode of death was cardiovascular in 82.2% of cases. Persistent CS was the dominant mode of death (66.5% of deaths), followed by arrhythmia (12.8% of deaths). The leading modes of non-cardiovascular death were anoxic brain injury and respiratory failure, accounting for 6.2% and 4.5%, respectively, of deaths (Fig. 1).
There were no meaningful differences in modes of death between patients with AMI-CS and patients with HF-CS (Fig. 2). Compared with patients without preceding CA, those with CA before CICU admission were more likely to die from anoxic brain injury (17.1% vs 0.9%; P < .001) or arrhythmia (21.6% vs 8.4%; P < .001). Compared with patients managed without tMCS, those managed with tMCS were more likely to die from persistent shock (P < .01), including both cardiogenic (73.5% vs 62.0%) and noncardiogenic (6.1% vs 2.9%) shock. Last, compared with patients who were not transitioned to CMO before death, those who were transitioned to CMO were more likely to die from persistent CS (70.8% vs 57.7%; P = .016) and less likely to die from arrhythmia (5.3% vs 27.9%; P < .001) (Fig. 2).
Discussion
In this analysis of modes of death in a broad CS population from a contemporary multinational registry, we found that most in-hospital deaths in patients with CS are cardiovascular and most are specifically related to persistent CS. These observations confirm the findings of prior analyses of patients with AMI-CS from both clinical trials and registries,4,5 and extend them to patients with HF-CS. In addition, among patients with CS after resuscitated CA, arrhythmic deaths and deaths owing to anoxic brain injury were found to be relatively more common, and among patients with CS managed with tMCS before death, persistent shock (both cardiogenic and noncardiogenic) was found to account for a greater majority of the deaths.
Defining modes of death is important for understanding the natural history of CS and for identifying opportunities to improve outcomes in this highly mortal syndrome. In the context of accumulating literature highlighting the extracardiac complications of CS and CS-related therapies, it is notable that persistent CS remains the dominant mode of death in CS. However, for several reasons, these results should not necessarily be interpreted as suggesting that more robust forms of hemodynamic support are critical to improving outcomes. First, two-thirds of CS deaths were preceded by a CMO transition, and, in that group, persistent CS was an especially common reason for death. In many of these cases, death was likely the result of an active transition in goals of care in the setting of persistent CS without a viable destination, rather than the inability to adequately support the patient with tMCS. Second, our data indicate the important contribution of noncardiogenic shock (eg, refractory vasoplegia) to CS mortality among those managed with tMCS, which is not an uncommon clinical evolution with tMCS. Finally, randomized trials of tMCS to date have uniformly failed to demonstrate improved survival with these devices.6,7 Nevertheless, our data emphasize the importance of reversing underlying drivers of CS and interrupting the vicious “cardiogenic shock spiral” that underpins mortality in CS,8 which should remain the focus of future interventional trials. At the same time, the fact that one-third of patients with CS died from a mode other than persistent CS underscores, in part, what makes intervening in CS so challenging.
Our data also highlight clinically important heterogeneity in how patients die with coincident CS and CA, namely, the greater prevalence of arrhythmic deaths and deaths owing to anoxic brain injury compared with those without preceding CA. These observations underscore the distinct insults and clinical sequelae associated with CA, which introduce competing risks that may affect the interpretation of clinical trials. Indeed, subgroup analyses of CS trials have suggested possible (albeit nonsignificant) heterogeneity in the treatment effects of certain therapies (eg, tMCS devices) according to CA status.6,7 To this end, there have been calls to stratify analyses by CA status in CS trials, and even to analyze these subgroups separately.2 Notably, the prevalence of preceding CA among patients with CS who died was lower than reported in recent trials.7 This finding is likely a function of trial entry criteria (eg, requirement for elevated serum lactate), which can lead to biased sampling of patients with CS enrolled in clinical trials. Our observations, thus, provide a more generalizable view of CS mortality and emphasize the importance of enrolling a sufficient number of patients without preceding CA in CS trials.
Several limitations of our study should be acknowledged. First, death classification was not adjudicated formally by a central clinical events committee. Nevertheless, site investigators were trained prospectively on standardized criteria for classifying modes of death. Second, accurately classifying mode of death in CS is complicated, not only because it is often multifactorial in patients with multiorgan system dysfunction, but because assigning a single mode of death may not adequately account for the clinical nuances involved in a patient’s death. For example, a patient with ongoing or worsening hemodynamic instability and multiorgan system dysfunction for whom care is withdrawn may be classified as dying from persistent CS, but noncardiac organ injury may have also contributed to the decision to withdraw care. These challenges are not unique to this analysis, but underscore the importance of continued efforts to describe and standardize death classification in CS. Finally, our analysis is limited to modes of in-hospital death, which may differ from modes of death among those who survive to hospital discharge.
Conclusions
Most in-hospital deaths in patients with CS are related to direct cardiovascular causes, particularly persistent CS. However, there is potentially important heterogeneity in modes of in-hospital death across subgroups defined by preceding CA and use of tMCS, which may be relevant to developing new therapeutic approaches, designing randomized trials, and ultimately improving outcomes in CS.
SC
Dr. Sunit-Preet Chaudhry
Cardiology · Advanced Heart Failure & TransplantationSt Vincent Hospitalsunitpreet
