Abstract
Background
The routine use of sedation and analgesia during post-cardiac arrest care and its association with clinical outcomes remain unclear. This study aimed to describe the use of sedatives and analgesics in post-cardiac arrest care, and evaluate associations with good functional outcome, survival, clinical seizures, and late awakening.
Methods
This was a post hoc analysis of the TTM2-trial, which randomized 1900 out-of-hospital cardiac arrest patients to either normothermia or hypothermia. In both groups, deep sedation (Richmond Agitation and Sedation Scale ≤ -4) was mandatory during the 40-h intervention. Cumulative doses of sedatives and analgesic drugs were recorded within the first 72 h from randomization. Outcomes were functional outcome (modified Rankin Scale) and survival status at 6 months, occurrence of clinical seizures during the intensive care stay, and late awakening (Full outline of unresponsiveness motor score of four 96 h after randomization). Cumulative propofol doses were divided into quartiles (Q1-Q4). Logistic regression models were used to assess associations between sedative doses and functional outcome and survival, clinical seizures, and late awakening, adjusting for the severity of illness and other clinical factors influencing sedation.
Results
A total of 1861 patients were analyzed. In a multivariable logistic regression model, higher propofol doses (Q3, 100.7–153.6 mg/kg) were associated with good functional outcome (OR 1.62, 95%CI 1.12—2.34) and (Q2 and Q3, 43.9–153.6 mg/kg) with survival (OR 1.49, 95%CI 1.05—2.12 and OR 1.84, 95%CI 1.27—2.65, respectively). Receiving fentanyl and remifentanil were associated with good functional outcome (OR 1.69, 95%CI 1.27—2.26 and OR 1.50, 95%CI 1.11—2.02) and survival (OR 1.80, 95%CI 1.35—2.40 and OR 1.56, 95%CI 1.16—2.10). Receiving fentanyl (OR 0.64, 95%CI 0.48—0.86) and higher propofol doses (Q2-4 (43.9–669.4 mg/kg) were associated with the occurrence of clinical seizures. The highest quartile of propofol dose (153.7–669.4 mg/kg, OR 3.19, 95%CI 1.91—5.42) was associated with late awakening.
Conclusions
In this study, higher doses of propofol and the use of remifentanil and fentanyl were associated with good functional outcome and survival, occurrence of clinical seizures, and late awakening.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13054-025-05461-0.
Keywords: Cardiac arrest, Targeted temperature management, Sedation, Seizures, Propofol, Midazolam
Introduction
The majority of patients unconscious after out-of-hospital cardiac arrest (OHCA) receive intensive care treatment and targeted temperature management (TTM) [1, 2]. TTM and fever management are recommended to mitigate hypoxic-ischemic brain injury and improve functional outcomes, with sedation and analgesia provided during post-cardiac arrest care to reduce discomfort induced by lowering the body temperature, facilitate therapy, reduce shivering, and prevent awareness during neuromuscular blockade [3–5]. About one third of cardiac arrest patients develop seizures, which are associated with poor outcomes, however, causal relationship with the degree of hypoxic-ischemic brain injury has not been established [6, 7]. Sedatives, some of which are also potent antiepileptic agents, may mitigate secondary brain injury and optimize neurological recovery by lowering metabolic rate and intracranial pressure [8, 9]. Thus, in clinical practice, sedation is often increased when a patient is having clinical seizures or shows signs of pain and agitation.
Sedation carries inherent risks, including compromised circulatory and respiratory functions, increased risk of delirium, prolonged mechanical ventilation, and intensive care stay [10–13]. Importantly, for patients with cardiac arrest, sedation may confound neurological prognostication, influencing the decision on withdrawal of life-sustaining therapies [14, 15].
This study investigates the association between the use of sedatives and analgesics and outcomes, including functional, mortality, clinical seizure, and time to awakening in a large multicenter trial. We hypothesized that there is an association between higher doses of sedatives and analgesics and long-term good functional outcome, survival, occurrence of clinical seizures, and late awakening.
Methods
This study adheres to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Supplement Table 1) [16].
Table 1.
Background and cardiac arrest characteristics
| Poor functional outcome | Good functional outcome | Overall | |
|---|---|---|---|
| n | 988 | 841 | 1861 |
| Male (%) | 737 (74.6) | 710 (84.4) | 1477 (79.4) |
| Age (mean (SD)) | 67.90 (12.01) | 59.22 (13.76) | 63.8 (13.6) |
| Frailty score (median (IQR)) | 3.24 (1.52) | 2.14 (0.91) | 2 (1) |
| Body mass index (mean (SD)) | 27.80 (6.34) | 27.12 (4.74) | 27.5 (5.7) |
| Circulatory shock on admission1 (%) | 359 (36.3) | 171 (20.3) | 536 (28.8) |
| Minutes to ROSC (mean (SD)) | 35.77 (21.21) | 24.80 (17.05) | 31 (20) |
| Initial shockable rhythm (%) | 585 (59.2) | 761 (90.5) | 1371 (73.7) |
| Bystander CPR (%) | 737 (74.6) | 725 (86.2) | 1487 (79.9) |
| Previous liver disease (%) | 31 (3.1) | 11(1.3) | 43 (2.3) |
| Previous renal disease (%) | 72 (7.3) | 19 (2.3) | 92 (4.9) |
| Previous cerebrovascular disease (%) | 90 (9.1) | 30 (3.6) | 120 (6.4) |
| Admission FOUR motor (%) | |||
| 0 No response to pain | 830 (91.6) | 607 (79.8) | 1463 (86.3) |
| 1 Extension response to pain | 19 (2.1) | 35 (4.6) | 54 (3.2) |
| 2 Flexion response to pain | 44 (4.9) | 68 (8.9) | 113 (6.7) |
| 3 Localizes pain | 13 (1.4) | 48 (6.3) | 61 (3.6) |
| 4 Awake and obeying commands | 0 (0.0) | 3 (0.4) | 5 (0.3) |
1 Circulatory shock at admission was defined as a systolic blood pressure of < 90 mmHg for > 30 min or end-organ hypoperfusion (cool arms and legs, urine output < 30 ml/hour, and heart rate < 60 beats/minute)
Abbreviations: SD standard deviation, BMI Body mass index, CPR Cardiopulmonary resuscitation, FOUR Full Outline of Unresponsiveness, ROSC Return of spontaneous circulation
Setting and participants
The TTM2-trial (Clinicaltrials.gov NCT02908308) was an international, multicenter, parallel group, investigator-initiated trial randomizing 1900 adult patients between November 2017 and January 2020, with an OHCA of presumed cardiac cause to targeted hypothermia (33°C) (hypothermia) or normothermia with early treatment of fever (< 37.8°C) (normothermia) [17]. As previously reported, unconscious adult (> 18 years) patients with a presumed cardiac cause and ROSC were eligible for inclusion. The main exclusion criteria were time from ROSC to screening of more than 180 min, unwitnessed cardiac arrest with asystole as the initial rhythm and limitations in care [17].
Patient management
Patients assigned to hypothermia were cooled to 33°C, and this temperature was maintained for 28 h, followed by rewarming to 37°C for 12 h (intervention period). In the normothermia group, the aim was to maintain a temperature of 37.5°C or less. After the intervention period, normothermia was maintained until 72 h after randomization in all patients who remained unconscious, and sedation was discontinued or tapered according to clinical state. Extubation and wake-up test was attempted at the earliest time possible after the intervention period (40 h), based on standard protocols for discontinuation of mechanical ventilation [18].
Multimodal neurological prognostication was performed by a physician blinded to the intervention no earlier than 96 h after randomization, with strict criteria according to the TTM2-trial study protocol [18]. Decision of withdrawal of life-sustaining therapy (WLST) was at the discretion of the treating physician [18].
Neuron-specific enolase (NSE) sampling was optional and conducted at sites with relevant expertise. The highest NSE value was recorded. Bilirubin and glomerular filtration rate (GFR) were measured daily during the first week of intensive care.
Sedative and analgesic drugs
During the intervention period of 40 h, deep sedation, defined as Richmond Agitation and Sedation Scale (RASS) -4 to -5, was mandatory in both groups and provided similar treatment in both groups [19] (Supplement Table 2). The trial protocol recommended short-acting drugs, but to comply with the pragmatic nature of the trial was left to the treating physician's discretion or local protocol. Type of drug and cumulative dosing of each sedative and analgesic drug administered up to 72 h after randomization were recorded as follows: midazolam (mg), propofol (mg), dexmedetomidine (mcg), clonidine (mcg), esketamine (mg), ketamine (mg), fentanyl (mcg), morphine (mg), remifentanil (mcg) and oxycodone (mg). In patients in whom neurological prognostication was performed, time of discontinuation of sedatives was recorded.
Table 2.
Clinical variables and outcomes
| Poor functional outcome | good functional outcome | Overall | |
|---|---|---|---|
| n | 988 | 841 | 1861 |
| Normothermia (%) | 493 (49.9) | 418 (49.7) | 931 (50.0) |
| Early discontinuation of TTM (%)1 | 241 (25.2) | 55 (6.7) | 300 (16.6) |
| Shivering2 (%) | 270 (28.3) | 347 (42.1) | 632 (34.9) |
| Clinical seizures (%)3 | 388 (39.8) | 74 (8.8) | 464 (25.1) |
| Highest NSE3 (mean (SD)) | 146.29 (203.59) | 30.12 (62.95) | 88.5 (161.5) |
| Lowest GFR (mean (SD)) | 49.45 (25.89) | 66.81 (24.56) | 57.7 (26.8) |
| Highest bilirubin (mean (SD)) | 20.65 (31.11) | 20.05 (23.08) | 20.4 (27.6) |
| Time to extubation, days (median (IQR)) | 5.64 (7.58) | 4.79 (6.54) | 3.5 (1.9, 5.9) |
| Time to wake-up, days4 (median (IQR)) | 5.66 (5.42) | 3.55 (3.62) | 2.5 (1.8, 4.4) |
| Neurological prognostication performed according to protocol (%) | 472 (47.9) | 399 (47.4) | 883 (47.5) |
| Poor prognosis likely at time of neurological prognostication (%) | 256 (53.8) | 8 (2.0) | 264 (29.5) |
| Time of sedation discontinued before prognostication, hours (median (IQR)) | 27.20 (42.86) | 35.21 (38.88) | 14.0 (1.0, 48.0) |
| ICU length of stay, days (median (IQR)) | 6.38 (7.96) | 7.35 (6.85) | 4.9 (2.9, 8.1) |
| Good neurological outcome (mRS 4–6), n (%) | 0 (0.0) | 841 (100.0) | 873 (47.0) |
| Survival at six months, n (%) | 77 (7.8) | 841 (100.0) | 950 (50.8) |
1 Discontinuation of TTM any time during the intervention period of 40 h
2 The occurrence of shivering during up to 72 h
3 Highest NSE measurement during the ICU stay. Measured in 1018 out of 1861 patients
4 Data available in 979 patients out of 1861
Abbreviations: TTM Target temperature management, NSE Neuron specific enolase, GFR Glomerular filtration rate, ICU intensive care unit, SD standard deviation
Shivering was assessed daily during the intensive care stay using the bedside shivering assessment scale (BSAS) [20] (Supplement Table 3). Shivering was defined as BSAS > 1 anytime up to 72 h after randomization. Prophylactic acetaminophen/paracetamol was recommended. In response to a BSAS > 1, the first measure taken was to increase sedation, and secondarily to administer neuromuscular blocking agent at the discretion of the treating physician The goal was to maintain BSAS 0–1.
Table 3.
Total doses of sedatives, analgesics, and neuromuscular blockade up to 72 h
| Good | Poor | p-value | |
|---|---|---|---|
| 841 | 988 | ||
| Propofol, mg/kg (mean (SD)) | 115.0 (72.2) | 98.0 (79.3) | < 0.001 |
| Propofol, n (%) | 767 (91%) | 815 (82%) | < 0.001 |
| Midazolam, mg/kg (mean (SD)) | 2.9 (3.4) | 2.7 (6.2) | 0.638 |
| Midazolam, n (%) | 324 (38%) | 372 (38%) | 0.737 |
| Midazolam and propofol, n (%) | 272 (32%) | 271 (27%) | 0.025 |
| Midazolam only, n (%) | 52 (6%) | 101 (10%) | 0.002 |
| Remifentanil, mcg/kg (mean (SD)) | 1.2 (3.92) | 0.7 (2.33) | 0.042 |
| Remifentanil, n (%) | 297 (35%) | 335 (34%) | 0.561 |
| Fentanyl, mcg/kg (mean (SD)) | 0.4 (3.94) | 0.3 (2.91) | 0.544 |
| Fentanyl, n (%) | 494 (59%) | 456 (46%) | < 0.001 |
| Dexmedetomidine, mcg/kg (mean (SD)) | 0.02 (0.04) | 0.02 (0.02) | 0.644 |
| Dexmedetomidine, n (%) | 113 (14%) | 26 (3%) | < 0.001 |
| Oxycodone, mg/kg (mean (SD)) | 0.3 (0.37) | 0.3 (0.41) | 0.972 |
| Oxycodone, n (%) | 77 (9%) | 35 (4%) | < 0.001 |
| Morphine, mg/kg (mean (SD)) | 0.8 (1.5) | 1.2 (2.3) | 0.224 |
| Morphine, n (%) | 101 (12%) | 120 (12%) | 0.986 |
| Any neuromuscular blockade, n (%) | 483 (57%) | 529 (53%) | 0.105 |
Primary outcome
A good functional outcome was defined as a modified Rankin Scale (mRS) 0–3 at six months. The mRS ranges from 0 to 6, with 0 representing no symptoms, 1 no clinically significant disability, 2 slight disability, 3 moderate disability, 4 moderately severe disability, 5 severe disability, and 6 death [21].
Secondary outcomes
Survival
Survival status was recorded by six months follow-up.
Clinical seizures
Clinical seizures were defined as myoclonus or tonic–clonic seizures, which were assessed daily and recorded as present or absent during the intensive care stay up to seven days after randomization. Seizures were managed according to local protocols, at the discretion of the treating physician.
Time to awakening
Awakening was defined as obeying commands, i.e. Full Outline of Unresponsiveness (FOUR) (Supplement Table 4) score motor component of four [22]. FOUR score was recorded daily during the first seven days of intensive care. Time to awakening was recorded, with late awakening defined as occurring after 96 h.
Table 4.
Multivariable logistic regression and chi-square analysis of clinical factors, sedation, and analgesics on functional outcome, survival, clinical seizures, and late awakening Association of clinical factors, TTM, and total dose of propofol up to 72 h with good functional outcome (mRS 0-3) at six months follow up, survival at six months, clinical seizures, and late awakening in multivariate regression model.
| Good functional outcome | Survival | |||||
|---|---|---|---|---|---|---|
| Number of complete cases | 1477 | 1495 | ||||
| Reference variable | Poor functional outcome | Mortality | ||||
| Variable | OR | CI | p | OR | CI | p |
| Normothermia | 0.84 | 0.65—1.08 | 0.176 | 0.98 | 0.76—1.26 | 0.872 |
| Propofol dose (mg/kg)* | 0.01 | 0.001 | ||||
| Propofol quartile 1 (0.01 – 43.8 mg/kg) | – | – | – | – | – | – |
| Propofol quartile 2 (43.9–100.6 mg/kg) | 1.38 | 0.96 – 2.00 | 0.08 | 1.49 | 1.05—2.12 | 0.025 |
| Propofol quartile 3 (100.7–153.6 mg/kg) | 1.62 | 1.12—2.34 | 0.01 | 1.84 | 1.27—2.65 | 0.001 |
| Propofol quartile 4 (153.7–669.4 mg/kg) | 1.04 | 0.71—1.51 | 0.825 | 1.28 | 0.88—1.85 | 0.192 |
| Midazolam | 1.01 | 0.77—1.33 | 0.943 | 1.10 | 0.84—1.45 | 0.49 |
| Fentanyl | 1.69 | 1.27—2.26 | < 0.001 | 1.80 | 1.35—2.40 | < 0.001 |
| Remifentanil | 1.50 | 1.11—2.02 | 0.008 | 1.56 | 1.16—2.10 | 0.003 |
| Clinical seizures | Late awakening | |||||
|---|---|---|---|---|---|---|
| Number of complete cases | 1505 | 826 | ||||
| Reference variable | No clinical seizures | Early awakening | ||||
| Normothermia | 0.75 | 0.58–0.96 | 0.022 | 0.57 | 0.40–0.81 | 0.002 |
| Propofol dose (mg/kg)* | < 0.001 | < 0.001 | ||||
| Propofol quartile 1 (0.01 – 43.8 mg/kg) | – | – | – | – | – | – |
| Propofol quartile 2 (43.9–100.6 mg/kg) | 1.53 | 1.06–2.2 | 0.022 | 0.78 | 0.46—1.34 | 0.369 |
| Propofol quartile 3 (100.7–153.6 mg/kg) | 1.56 | 1.06–2.29 | 0.023 | 1.33 | 0.80—2.23 | 0.279 |
| Propofol quartile 4 (153.7–669.4 mg/kg) | 2.82 | 1.94–4.11 | < 0.001 | 3.19 | 1.91–5.42 | < 0.001 |
| Midazolam | 1.99 | 1.52–2.61 | < 0.001 | 1.98 | 1.38–2.86 | < 0.001 |
| Fentanyl | 0.64 | 0.48–0.86 | 0.003 | 1.26 | 0.86–1.86 | 0.246 |
| Remifentanil | 1.19 | 0.89–1.59 | 0.245 | 1.20 | 0.82–1.77 | 0.352 |
Bolded values indicate significant associations (p < 0.05)
The multivariable regression model is including variables: age, male sex, time to return of spontaneous circulation, witnessed arrest, shock on admission, shockable rhythm, normothermia, shivering or neuromuscular blockade, lowest glomerular filtration rate, and highest bilirubin. Abbreviations: OR Odds ratio, conf Confidence interval, p p-value
*Analysed using Analysis of Variance (ANOVA)
Statistical analyses
Patient and cardiac arrest characteristics were used to describe the study cohort. Clinical variables such as temperature management strategy, sedation and analgesia use, intensive care interventions, and outcomes, including a figure describing numbers of patients awake, dead, and comatose from day 1 to 7 post randomization, were also presented to describe the post–cardiac arrest care. Cumulative total doses of sedatives and analgesics were adjusted for body weight, expressed as mg or mcg per kg as appropriate. Continuous variables are presented as median and interquartile range (IQR) or mean and standard deviation (SD). Categorical variables are reported using numbers and percentages. Propofol doses were divided into quartiles (Q1-Q4) to explore dose–response relationships, given observed non-linearity, while administration of midazolam, fentanyl, and remifentanil were evaluated as binary variables (yes/no).
Associations between sedative and analgesic use and outcomes: Justification for variables in the multivariable model
We evaluated propofol doses using chi-square statistics to test the associations with outcomes (good functional outcome (mRS 0–3) at 6 months, survival at 6 months, clinical seizures, and late awakening). The association of propofol quartiles, midazolam, fentanyl, remifentanil, and outcomes were also evaluated in an univariable and multivariable logistic regression model with odds ratio (OR with 95% confidence intervals (CI)). The overall test of propofol chi-square test and the multiple logistic regression model, including propofol dose, midazolam, fentanyl, and remifentanil, were adjusted for baseline severity of illness, clinically important variables, and design variables of the TTM2-trial to account for severity of illness and clinical factors potentially influencing the choice and dosage of sedative and analgesics. These included: age, sex, witnessed arrest, shockable rhythm, minutes to ROSC, shock on admission, body mass index (BMI), TTM allocation, shivering or administration of any neuromuscular blockade agent (recorded as yes/no), lowest GFR, and highest bilirubin.
Sedation/analgesia by temperature group
Sedative and analgesic doses and proportion of patients in the hypothermia versus the normothermia group were analyzed using Wilcoxon rank sum test or Wilcoxon rank rum exact test and using Pearson’s chi-square test.
Sensitivity analysis
Patients comatose at 96 h without clinical seizures were included in the sensitivity analysis, as they would be the population most likely to benefit from optimized sedation strategies. Patients with clinical seizures during intensive care stay were excluded, as their convulsions could significantly impact sedation management and confound the analysis. For the patients who underwent neurological prognostication, the time of sedation discontinuation was recorded, allowing for the calculation of the duration of sedation and the average dose of sedatives and analgesics as dose per kilogram per hour.
Results
The TTM2 trial enrolled 1900 patients, with 37 patients withdrawing or unable to provide consent, and 2 patients undergoing randomization twice. Flowchart of included and excluded patients are shown in Fig. 1. This resulted in 1861 patients being included in the intention-to-treat analyses. Detailed patient and cardiac arrest characteristics are summarized in Table 1.
Fig. 1.
Flow chart of included and excluded patients in this post-hoc study of the TTM2-trial [18]
Clinical variables and outcomes are presented in Table 2, while Supplement Fig. 1 describes the proportion of patients awake, comatose, and dead during the first seven days after cardiac arrest. The rate of good functional outcomes (mRS 0–3) at six months was 47.0%, while the six-month survival rate was 50.8%. Clinical seizures were observed in 25% of patients during the intensive care stay. The median time until awakening was 2.5 (IQR 1.8- 4.4) days.
Total sedative and analgesic doses up to 72 h post-randomization are outlined in Table 3. Propofol was the most used sedative, with higher mean doses (115.0 mg/kg vs. 98.0 mg/kg) and proportions of use (91% vs. 82%) in patients with good versus poor outcome, followed by midazolam, which showed similar usage between the two groups (mean dosages 2.9 mg/kg vs. 2.7 mg/kg and both 38%, p = 0.737). Among analgesics, fentanyl was the most frequently, with more frequent use in good outcome patients (mean dosages 0.4 mcg/kg vs 0.3 mcg/kg, proportion of use 60% vs. 47%), followed by remifentanil, which also had a higher mean dosage in good outcome patients (mean dosages 1.2 mcg/kg vs. 0.7 mcg/kg, proportion of use 35% vs. 34%).
There were no significant differences in average doses or proportion of patients receiving any sedative or analgesics between the hypothermia and the normothermia group, as shown in Supplement Table 5. However, significantly more patients received a neuromuscular blockade in the hypothermia group, 614 (66.0%) compared to 418 (44.9%) (p < 0.001).
Associations between sedatives and outcomes
Exploratory analyses indicated a non-linear relationship between propofol dose and outcomes (good functional outcome, survival, clinical seizures, and delayed awakening) in logistic regression models, as illustrated in Fig. 2. Consequently, propofol doses were categorized into quartiles for subsequent analyses to better assess associations with outcomes. Proportion of patients alive at 72 h in each propofol quartile group were: 67.7% (Q1), 88.0% (Q2), 97.0% (Q3), and 99.3% (Q4).
Fig. 2.
The relationship between cumulative propofol dose and good functional outcome, survival, clinical seizures, and late awakening. A logistic regression was applied to visualize the relationship between cumulative propofol dose (mg/kg) and the probability of good functional outcome, survival, clinical seizures, and late awakening. Each point represents an individual patient (jittered for visibility), with the blue line indicating the predicted probability and the shaded area representing the 95% confidence interval. The increasing curve suggests a non-linear association, with higher cumulative doses associated with a greater probability of the outcome across the observed range
Chi-square analyses showed a significant association between propofol dose and good functional outcome (p < 0.01) and survival at six months (p = 0.001), as well as clinical seizures (p < 0.001), and late awakening (p < 0.001), also after adjusting for confounding factors. The results of the multivariable logistic regression models are presented in Table 4. In a multivariable regression model, higher propofol doses (Q3, 100.7–153.6 mg/kg) were associated with good functional outcome (OR 1.62 CI 1.12—2.34), and Q2 and Q3 (43.9–153.6 mg/kg) with survival (OR 1.49 95%CI 1.05—2.12 and OR 1.84 95%CI 1.27—2.65, respectively). Fentanyl and remifentanil were associated with good functional outcome (OR 1.69 95% CI 1.27—2.26 and OR 1.50 95% CI 1.11—2.02) and survival at six months follow-up (OR 1.80 95%CI 1.35—2.40 and OR 1.56 95%CI 1.16—2.10). Additionally, fentanyl (OR 0.64 95%CI 0.48—0.86) and higher propofol doses (Q2-4, 43.9–669.4 mg/kg) were associated with clinical seizures. Higher doses of propofol (Q4, 153.7–669.4 mg/kg) were associated with late awakening (OR 3.19 95%CI 1.91—5.42). Furthermore, midazolam was associated with clinical seizures (OR 1.99, CI 95% 1.52–2.61) and with late awakening (OR 1.98, CI 95% 1.38–2.86).
In the sensitivity analysis including 463 patients unconscious at 96 h and without any clinical seizures, 359 (78%) patients had neurological prognostication performed, 90 (25%) patients had a poor neurological prognosis, and median duration without sedatives before neurological prognostication was 6.0 h (IQR 1.0, 46.5). Background, cardiac arrest characteristics, clinical variables and outcomes are presented in supplement Tables 6 and 7, and doses of sedative and analgesic are outlined in Supplement Table 8. Chi-square analyses showed no significant association of propofol dose and good functional outcome was observed, while high dose propofol were significantly associated with survival (Supplement Table 9). In a multivariable regression model, higher propofol doses (Q3 and Q4, 1.86–38.86 mg/kg) were associated with good functional outcome (OR 3.15 95%CI 1.29—8.06 and OR 2.78 95%CI 1.15—6.99, respectively), and survival (OR 3.43 95%CI 1.43—8.59 and OR 3.27 95%CI 1.38 – 8.00, respectively). Receiving fentanyl and remifentanil were associated with good functional outcome (OR 2.16 95% CI 1.10—4.39 and OR 2.25 95% CI 1.14—4.52, respectively) and survival (OR 2.65 95%CI 1.35—5.40 and OR 2.25 95%CI 1.15—4.49, respectively).
Discussion
In this post-hoc analysis of a large multicentered trial comparing hypothermia with normothermia in out of hospital cardiac arrest, including 40 h of protocolized deep sedation, we found a significant association between higher total doses of propofol and good functional outcome and survival at 6 months. Notably, proportion of patients alive at 72 h increased across quartiles of total propofol dose, suggesting a dose–response relationship between higher propofol exposure and early survival. Additionally, higher total doses of propofol over the first 72 h after randomization were associated with a higher frequency of clinical seizures, with high total doses also associated with late awakening. In addition, we found that the use of the two analgesics remifentanil and fentanyl was associated with good functional outcome and survival at 6 months follow-up. In a subgroup of patients without clinically observed seizures and remaining comatose at the time of earliest neurological prognostication (96 h), we found that higher hourly doses of propofol were associated with good functional outcome and survival at 6 months follow-up. Our results also indicate that the use of fentanyl and remifentanil is associated with good functional outcomes and survival, further emphasizing that patients with milder injuries tend to need more sedatives and analgesics for therapeutic comfort.
Seizures are often observed after cardiac arrest as a result of neuronal excitation due to brain injury [6]. Seizures are strongly associated with poor outcomes and if left untreated, can potentially exacerbate brain injury by increasing metabolic demand, disruption of cerebral autoregulation, and excitotoxicity [23–27]. Since the causality cannot be known, it is unclear whether a higher dose of propofol is neuroprotective and reduces the risk of seizures or if this only reflects the clinical practice of increasing the sedation when a patient is having clinical seizures or shows signs of pain and agitation. Thus, in patients without seizures, higher doses of propofol were associated with improved outcomes, aligning with the observation that patients with less severe brain injuries often require more sedatives for comfort. Alternatively, this may suggest that sedation itself has neuroprotective effects. However, the direction of this relationship remains uncertain, highlighting the need for further investigation.
Shivering is another common reason for increased sedation in clinical practice, and in the main trial, it was recommended that shivering be treated with increased sedation as first-line therapy. A shivering response may indicate preserved thermoregulation and less severe brain injury and has been previously associated with improved outcomes after cardiac arrest [28–30]. Higher doses of sedatives and analgesics observed in patients with good functional outcomes may be partly explained by the higher incidence of shivering in patients with good functional outcome and survival [31–33].
In concordance with our findings that midazolam was associated with late awakening, a small randomized clinical trial comparing sedation regimens during hypothermia after cardiac arrest found that propofol and remifentanil significantly reduced time to extubation compared to midazolam and fentanyl [34]. Another trial with 460 participants demonstrated that the use of propofol and remifentanil compared to midazolam and fentanyl resulted in significantly earlier awakening and more ventilator-free days, however, with no differences in survival nor neurological outcomes at hospital discharge [35].
There were no statistically significant differences in sedation or analgesia dosing between patients treated with hypothermia or normothermia. This suggests that doses to keep the patients at deep sedation during TTM were similar. However, after adjusting for sedation and analgesia, severity of illness, and clinical factors, we found normothermia to be associated with a reduced risk of late awakening compared to hypothermia. Hypothermia has previously been shown to be associated with late awakening [36]. Hypothermia decreases drug metabolism and elimination and increases the risk of lingering sedation, which consequently may delay awakening [37–40].
Current guidelines favor the use of short-acting sedatives and analgesics during TTM, specifically propofol, remifentanil, and fentanyl, over midazolam and morphine, and recommend stopping sedatives as soon as possible to assess the level of consciousness and to facilitate neurological prognostication [3–5]. However, limited evidence supports these recommendations, reflected in the variability in clinical practice [41–44]. In this study propofol was the most used sedative agent, however, midazolam was also commonly used. Our multivariable regression model showed no association between midazolam use and good functional outcomes or survival. Instead, midazolam use was associated with seizures and delayed awakening, even after adjusting for illness severity and clinical factors. Other studies have similarly reported significant use of long-acting drugs [45–48]. While causality cannot be established in this study, the use of long-acting sedatives may reflect patient-specific challenges, such as hemodynamic instability or clinical seizures, which influence sedative choice. Alternatively, the selection of sedatives may depend on physician preference or institutional practices [48, 49].
Strengths and limitations
This study provides the to date most comprehensive evaluation of sedation and analgesia management after out-of-hospital cardiac arrest, with robust adjustments for illness severity. By examining individual patient data, the study identifies factors influencing outcome. The large, diverse patient cohort from international centers further strengthens the generalizability of the findings. Nevertheless, as with all observational analyses, particularly post-hoc studies, there is a risk of confounding bias. Although we adjusted for key clinical variables and illness severity in the multivariable analysis, there may still be residual confounding. For instance, while use of neuromuscular blockade was included in the multivariable models to account for their potential impact, the depth and duration of neuromuscular blockade were not recorded, limiting our ability to assess their relationship with sedation depth, neurological status, or outcomes. Additionally, EEG recordings were not uniformly available across participating centers and were therefor not included in this study which may have introduced heterogeneity in seizure detection. Moreover, variation in sedation and analgesia practices across the 61 participating centers may represent a source of residual confounding not fully captured by the current multivariable model. Although the TTM2-trial protocol was standardized across sites, recommending a target for sedation depth (RASS –4 to –5) and the use of short-acting agents, differences in clinical practice may still have influenced drug choice and dosing, and are not fully accounted for in this analysis.
Furthermore, sedation and analgesia management were not primary outcomes of the original TTM2 trial. As such, there is a risk of post-randomization bias, particularly related to differences in clinical management not captured by available data. To mitigate potential information bias, data was collected prospectively using standardized case report forms across all sites. However, with the post-hoc nature of the study, further investigation to establish the causality of the associations found, and findings should be interpreted as hypothesis-generating.
Additionally, while sedation was mandatory up to 40 h, data on sedative and analgesic cumulative doses were collected at 72 h. The reasons for prolonged sedation or titration of sedation were not recorded, such as the adjustment of sedative and analgesic dosages in the context of WLST. To address this, we conducted subgroup analyses, analyzing those comatose at 96 h and with known sedation duration, to improve our understanding of sedation and analgesia management in post-cardiac arrest care.
Future aspects The findings in this study and the possible impact of sedation strategies on patient outcomes, will be further investigated in the Sedation after Cardiac Arrest and Resuscitation (SED-CARE) trial (clinicaltrials.gov no NCT05564754, 2022–10-03). The SED-CARE trial is a part of the CARE platform trial and is essential in addressing the knowledge gap in a randomized and prospective manner, allowing for a more robust assessment of the causal effects of sedation strategies on seizures, neurological outcome, and survival.
Conclusions
In this post-hoc analysis of a large multicenter cardiac arrest trial, we found that higher doses of propofol were significantly associated with good functional outcomes and survival at six months, clinical seizures, and late awakening. Remifentanil and fentanyl were both associated with good functional outcomes and survival, while midazolam was associated with clinical seizures and delayed awakening. These findings partially support our hypothesis, suggesting that the type and dosage of sedatives and analgesics during post-cardiac arrest care may reflect the severity of illness, with higher doses in patients with less severe brain injury and better outcomes. In comparison, higher doses of sedatives in the presence of seizures are not associated with improve outcomes. Clinical trials are needed to establish causality and optimizing sedation strategies.
Supplementary Information
Author contributions
All authors have substantially contributed to the conception of the work and substantially contributed to the conception, interpretation of data, and substantially revised the manuscript. All authors also have approved the submitted version and agreed both to be personally accountable for their contribution and to ensure accuracy and integrity of all parts of the work. A.C. has drafted the work, substantially contributed to the design of the work, the analysis, and the interpretation of data.
Funding
Open access funding provided by Lund University. Supported by independent research grants from nonprofit or governmental agencies (the Swedish Research Council [Vetenskapsrådet], Swedish Heart–Lung Foundation, Stig and Ragna Gorthon Foundation, Knutsson Foundation, Laerdal Foundation, Hans-Gabriel and Alice Trolle-Wachtmeister Foundation for Medical Research, and Regional Research Support in Region Skåne) and by governmental funding of clinical research within the Swedish National Health Service.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Data is provided within the manuscript or supplementary information files.
Declarations
Ethics approval and consent to participate
The TTM2-trial was approved by Regional Ethical Review Board in Lund, Sweden (Nr 2015/228 and 2017/36) and in all participating countries and was carried out in accordance with the World Medical Association’s Declaration of Helsinki. Written informed consent was waived, deferred, or obtained from a legal surrogate, depending on the circumstances, and was obtained from each patient who regained mental capacity. For this sub analysis, no further ethical approval was required.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Thomassen AWM. Prevalence and prognostic significance of coma after cardiac arrest outside intensive care and coronary units. Acta Anaesthesiol Scand. 1979;23(2):143–8. [DOI] [PubMed] [Google Scholar]
- 2.Dragancea I, Horn J, Kuiper M, et al. Neurological prognostication after cardiac arrest and targeted temperature management 33 degrees C versus 36 degrees C: Results from a randomised controlled clinical trial. Resuscitation. 2015;93:164–70. [DOI] [PubMed] [Google Scholar]
- 3.Sandroni C, Nolan JP, Andersen LW, et al. ERC-ESICM guidelines on temperature control after cardiac arrest in adults. Intensive Care Med. 2022;48(3):261–9. [DOI] [PubMed] [Google Scholar]
- 4.Nolan JPSC, Bottiger BW, et al. European Resuscitation Council and European Society of Intensive Care Medicine guidelines 2021: post-resuscitation care. Intensive Care Med. 2021;47(4):369–421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hirsch KG, Abella BS, Amorim E, et al. Critical Care Management of Patients After Cardiac Arrest: A Scientific Statement From the American Heart Association and Neurocritical Care Society. Circulation. 2024;149(2):e168–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Nielsen NHJ, Nilsson F. Outcome, timing and adverse events in therapeutic hypothermia after out-of-hospital cardiac arrest. Acta Anaesthesiol Scand. 2009;53:926–34. [DOI] [PubMed] [Google Scholar]
- 7.Rossetti AO, Tovar Quiroga DF, Juan E, et al. Electroencephalography Predicts Poor and Good Outcomes After Cardiac Arrest: A Two-Center Study. Crit Care Med. 2017;45(7):e674–82. [DOI] [PubMed] [Google Scholar]
- 8.Oddo M, Crippa IA, Mehta S, et al. Optimizing sedation in patients with acute brain injury. Crit Care. 2016;20(1):128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Dunn LK, Naik BI, Nemergut EC, Durieux ME. Post-Craniotomy Pain Management: Beyond Opioids. Curr Neurol Neurosci Rep. 2016;16(10):93. [DOI] [PubMed] [Google Scholar]
- 10.Treggiari MM, Romand JA, Yanez ND, et al. Randomized trial of light versus deep sedation on mental health after critical illness. Crit Care Med. 2009;37(9):2527–34. [DOI] [PubMed] [Google Scholar]
- 11.Kress JPPA, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med. 2000;342(20):1471–7. [DOI] [PubMed] [Google Scholar]
- 12.Shehabi Y, Chan L, Kadiman S, et al. Sedation depth and long-term mortality in mechanically ventilated critically ill adults: a prospective longitudinal multicentre cohort study. Intensive Care Med. 2013;39(5):910–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.SRLF Trial Group. Impact of oversedation prevention in ventilated critically ill patients: a randomized trial-the AWARE study. Ann Intensive Care. 2018;8(1):93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dale Craig M, et al. Understanding early decisions to withdraw life-sustaining therapy in cardiac arrest survivors a qualitative investigation. Annals Am Thorac Soc. 2016;13(7):1115–22. [DOI] [PubMed] [Google Scholar]
- 15.Mulder MGH, Smith SW, Dhaliwal R, Scott NL, Sprenkle MD, Geocadin RG. Awakening and withdrawal of life-sustaining treatment in cardiac arrest survivors treated with therapeutic hypothermia*. Crit Care Med. 2014;42(12):2493–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.von Elm EAD, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370:1453–7. [DOI] [PubMed] [Google Scholar]
- 17.Dankiewicz JCT, Lilja G, Jakobsen JC, Levin H, Ullén S, Rylander C, Wise MP, Oddo M, Cariou A, Bělohlávek J, Hovdenes J, Saxena M, Kirkegaard H, Young PJ, Pelosi P, Storm C, Taccone FS, Joannidis M, Callaway C, Eastwood GM, Morgan MPG, Nordberg P, Erlinge D, Nichol AD, Chew MS, Hollenberg J, Thomas M, Bewley J, Sweet K, Grejs AM, Christensen S, Haenggi M, Levis A, Lundin A, Düring J, Schmidbauer S, Keeble TR, Karamasis GV, Schrag C, Faessler E, Smid O, Otáhal M, Maggiorini M, Wendel Garcia PD, Jaubert P, Cole JM, Solar M, Borgquist O, Leithner C, Abed-Maillard S, Navarra L, Annborn M, Undén J, Brunetti I, Awad A, McGuigan P, Bjørkholt Olsen R, Cassina T, Vignon P, Langeland H, Lange T, Friberg H, Nielsen N. TTM2 Trial investigators: hypothermia versus normothermia after out-of-hospital cardiac arrest. N Engl J Med. 2021;384(24):2283–94. [DOI] [PubMed] [Google Scholar]
- 18.Dankiewicz JCT, Lilja G, Jakobsen JC, Bělohlávek J, Callaway C, Cariou A, Eastwood G, Erlinge D, Hovdenes J, Joannidis M, Kirkegaard H, Kuiper M, Levin H, Morgan MPG, Nichol AD, Nordberg P, Oddo M, Pelosi P, Rylander C, Saxena M, Storm C, Taccone F, Ullén S, Wise MP, Young P, Friberg H, Nielsen N. Targeted hypothermia versus targeted Normothermia after out-of-hospital cardiac arrest (TTM2): A randomized clinical trial-Rationale and design. Am Heart J. 2019;217:23–31. [DOI] [PubMed] [Google Scholar]
- 19.Sessler CN, Gosnell MS, Grap MJ, et al. The richmond agitation-sedation scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med. 2002;166(10):1338–44. [DOI] [PubMed] [Google Scholar]
- 20.Badjatia N, Strongilis E, Gordon E, et al. Metabolic impact of shivering during therapeutic temperature modulation: the Bedside Shivering Assessment Scale. Stroke. 2008;39(12):3242–7. [DOI] [PubMed] [Google Scholar]
- 21.Van Swieten JC, et al. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19(5):604–7. [DOI] [PubMed] [Google Scholar]
- 22.Wijdicks EF, Bamlet WR, Maramattom BV, Manno EM, McClelland RL. Validation of a new coma scale: the FOUR score. Ann Neurol. 2005;58(4):585–93. [DOI] [PubMed] [Google Scholar]
- 23.Rossetti AOOM, Logroscino G, Kaplan PW. Prognostication after cardiac arrest and hypothermia: a prospective study. Ann Neurol. 2010;67:301–7. [DOI] [PubMed] [Google Scholar]
- 24.Bouwes Aline, et al. Acute posthypoxic myoclonus after cardiopulmonary resuscitation. BMC neurology. 2012;12:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Seder DBSK, Rubertsson S. Neurologic outcomes and postresuscitation care of patients with myoclonus following cardiac arrest. Crit Care Med. 2015. 10.1097/SA.0000000000000181. [DOI] [PubMed] [Google Scholar]
- 26.Elmer JRJ, Faro J. Clinically distinct electroencephalographic phenotypes of early myoclonus after cardiac arrest. Ann Neurol. 2016;80:175–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Backman SWE, Dragancea I, Friberg H, Rundgren M, Ullen S. Cronberg T Electroencephalographic characteristics of status epilepticus after cardiac arrest. Clin Neurophysiol Off J Int Fed Clin Neurophysiol. 2017;128:681–8. [DOI] [PubMed] [Google Scholar]
- 28.Lee DH, Lee BK, Cho YS, et al. High heat generation is associated with good neurologic outcome in out-of-hospital cardiac arrest survivors underwent targeted temperature management at 33 degrees C. Resuscitation. 2020;153:187–94. [DOI] [PubMed] [Google Scholar]
- 29.Nair SU, Lundbye JB. The occurrence of shivering in cardiac arrest survivors undergoing therapeutic hypothermia is associated with a good neurologic outcome. Resuscitation. 2013;84(5):626–9. [DOI] [PubMed] [Google Scholar]
- 30.Hovdenes J, Roysland K, Nielsen N, et al. A low body temperature on arrival at hospital following out-of-hospital-cardiac-arrest is associated with increased mortality in the TTM-study. Resuscitation. 2016;107:102–6. [DOI] [PubMed] [Google Scholar]
- 31.Lee DH, Lee BK, Cho YS, et al. High heat generation is associated with good neurologic outcome in out-of-hospital cardiac arrest survivors underwent targeted temperature management at 33 °C. Resuscitation. 2020;153:187–94. [DOI] [PubMed] [Google Scholar]
- 32.May TL, Riker RR, Fraser GL, et al. Variation in sedation and neuromuscular blockade regimens on outcome after cardiac arrest. Crit Care Med. 2018;46(10):e975–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Geller BJ, Maciel CB, May TL, Jentzer JC. Sedation and shivering management after cardiac arrest. Eur Heart J Acute Cardiovasc Care. 2023;12(8):518–24. [DOI] [PubMed] [Google Scholar]
- 34.Bjelland TW, Dale O, Kaisen K, et al. Propofol and remifentanil versus midazolam and fentanyl for sedation during therapeutic hypothermia after cardiac arrest: a randomised trial. Intensive Care Med. 2012;38(6):959–67. [DOI] [PubMed] [Google Scholar]
- 35.Paul M, Bougouin W, Dumas F, et al. Comparison of two sedation regimens during targeted temperature management after cardiac arrest. Resuscitation. 2018;128:204–10. [DOI] [PubMed] [Google Scholar]
- 36.Lybeck A, Cronberg T, Aneman A, et al. Time to awakening after cardiac arrest and the association with target temperature management. Resuscitation. 2018;126:166–71. [DOI] [PubMed] [Google Scholar]
- 37.van den Broek MPGF, Egberts AC, Rademaker CM. Effects of hypothermia on pharmacokinetics and pharmacodynamics: a systematic review of preclinical and clinical studies. Clin Pharmacokinet. 2010;49(5):277–94. [DOI] [PubMed] [Google Scholar]
- 38.Leslie KSD, Bjorksten AR, Moayeri A. Mild hypothermia alters propofol pharmacokinetics and increases the duration of action of atracurium. Anesth Analg. 1995;80(5):1007–14. [DOI] [PubMed] [Google Scholar]
- 39.Tortorici MA, Kochanek PM, Poloyac SM. Effects of hypothermia on drug disposition, metabolism, and response: a focus of hypothermia-mediated alterations on the cytochrome P450 enzyme system. Crit Care Med. 2007;35(9):2196–204. [DOI] [PubMed] [Google Scholar]
- 40.Varghese JM, Roberts JA, Lipman J. Pharmacokinetics and pharmacodynamics in critically ill patients. Curr Opin Anaesthesiol. 2010;23(4):472–8. [DOI] [PubMed] [Google Scholar]
- 41.Celis-Rodríguez E, Díaz Cortés JC, Cárdenas Bolívar YR, et al. Evidence-based clinical practice guidelines for the management of sedoanalgesia and delirium in critically ill adult patients. Med Intensiva (Engl Ed). 2020;44(3):171–84. [DOI] [PubMed] [Google Scholar]
- 42.DAS-Taskforce 2015; Baron R, Binder A, Biniek R et al: Evidence and consensus based guideline for the management of delirium, analgesia, and sedation in intensive care medicine. Revision 2015 (DAS-Guideline 2015) - short version. Ger Med Sci 2015 Nov 12, 13(Coc19). [DOI] [PMC free article] [PubMed]
- 43.Devlin JW, Skrobik Y, Gelinas C, et al. Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU. Crit Care Med. 2018;46(9):e825–73. [DOI] [PubMed] [Google Scholar]
- 44.Gonzalez D, Dahiya G, Mutirangura P, et al. Post cardiac arrest care in the cardiac intensive care unit. Curr Cardiol Rep. 2024;26(2):35–49. [DOI] [PubMed] [Google Scholar]
- 45.Orban JC, Cattet F, Lefrant JY, et al. The practice of therapeutic hypothermia after cardiac arrest in France: a national survey. PLoS ONE. 2012;7(9): e45284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Deye N, Cariou A, Girardie P, et al. Endovascular versus external targeted temperature management for patients with out-of-hospital cardiac arrest: a randomized. Control Study Circ. 2015;132(3):182–93. [DOI] [PubMed] [Google Scholar]
- 47.Kamps MJ, Horn J, Oddo M, et al. Prognostication of neurologic outcome in cardiac arrest patients after mild therapeutic hypothermia: a meta-analysis of the current literature. Intensive Care Med. 2013;39(10):1671–82. [DOI] [PubMed] [Google Scholar]
- 48.Ceric A, May TL, Lybeck A, et al. Cardiac arrest treatment center differences in sedation and analgesia dosing during targeted temperature management. Neurocrit Care. 2023;38(1):16–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Dolmans RGF, Nahed BV, Robertson FC, Peul WC, Rosenthal ES, Broekman MLD. Practice-pattern variation in sedation of neurotrauma patients in the intensive care unit: an international survey. J Intensive Care Med. 2023;38(12):1143–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Data is provided within the manuscript or supplementary information files.