N-terminal pro-B-type natriuretic (NT-proBNP) is expressed in the heart and brain, and serum levels are elevated in acute heart and brain diseases. We aimed to assess the possible association between serum levels and neurological outcome and death in comatose patients resuscitated from out-of-hospital cardiac arrest (OHCA). Of the 939 comatose OHCA patients enrolled and randomized in the Targeted Temperature Management (TTM) trial to TTM at 33°C or 36°C for 24 hours, 700 were included in the biomarker substudy. Of these, 647 (92%) had serum levels of NT-proBNP measured 24, 48, and 72 hours after return of spontaneous circulation (ROSC). Neurological outcome was evaluated by the Cerebral Performance Category (CPC) score and modified Rankin Scale (mRS) at 6 months. Six hundred thirty-eight patients (99%) had serum NT-proBNP levels ≥125 pg/ml. Patients with TTM at 33°C had significantly lower NT-proBNP serum levels (median 1,472 pg/ml) than those in the 36°C group (1,914 pg/ml) at 24 hours after ROSC, p <0.01 but not at 48 and 72 hours. At 24 hours, an increase in NT-proBNP quartile was associated with death (P logrank <0.0001). In addition, NT-proBNP serum levels > median were independently associated with poor neurological outcome (odds ratio, OR CPC 2.02, CI 1.34 to 3.05, p <0.001; OR mRS 2.28, CI 1.50 to 3.46, p <0.001) adjusted for potential confounders. The association was diminished at 48 and 72 hours after ROSC. In conclusion, NT-proBNP serum levels are increased in comatose OHCA patients. Furthermore, serum NT-proBNP levels are affected by level of TTM and are associated with death and poor neurological outcome.
In this post hoc analysis of data from the Targeted Temperature Management (TTM) trial, we aimed to assess the association between N-terminal pro-B-type natriuretic peptide (NT-proBNP) with death and neurological outcome in patients resuscitated but comatose after out-of-hospital cardiac arrest (OHCA). Second, we assessed the effect of TTM at 33°C (TTM33) versus 36°C (TTM36) on NT-proBNP concentrations, as the level of TTM has previously been found to influence hemodynamics in OHCA patients. Third, we examined whether NT-proBNP could add prognostic information in addition to neuron-specific enolase (NSE), a glycolytic enzyme expressed in neuronal and neuroendocrine cells and a strong and independent prognostic marker of neurological outcome in comatose OHCA patients.
Methods
The TTM trial was a multicenter, randomized, parallel-group, assessor-blinded, monitored clinical trial investigating the potential protective effect of TTM of 33°C versus 36°C in comatose patients resuscitated from OHCA. Patients were randomized, stratified for site from November 2010 to January 2013. Comatose (Glasgow Coma Scale <8) adults (age ≥18) hospitalized after OHCA with presumed cardiac cause were included. Eligible patients had sustained return of spontaneous circulation (ROSC) of more than 20 minutes. A full list of exclusion criteria has been published elsewhere. Written informed consent was obtained or waived as previously described. Patients were randomized in a 1:1 ratio to either TTM33 or TTM36 maintained for 24 hours. All patients were sedated, intubated, and mechanically ventilated during the intervention period. Active treatment was maintained for a predefined period of minimum 108 hours after ROSC, and strict criteria for withdrawal of life support were predefined. Prehospital cardiac arrest data and factors traditionally associated with mortality were systematically collected in accordance with the Utstein guidelines. Furthermore, data on factors with potential influence on NT-proBNP serum concentrations (age, gender, kidney function, body mass index, acute ischemic heart disease, and co-morbidities including heart failure) were collected.
The present biomarker substudy included patients enrolled at 29 of the 36 participating sites, and patients were included if NT-proBNP were measured 24 hours after ROSC. Fifty-three patients lacked NT-proBNP measurements 24 hours after ROSC, leading to 647 patients (92%) comprising the study cohort. About 318 of these patients (49%) were randomized to TTM33 and 329 patients (51%) to TTM36. Blood samples were drawn 24, 48, and 72 hours after ROSC. All samples were processed at the different sites, aliquoted, and frozen to −80°C before shipment to a core biobank before analysis. NT-proBNP values were not available to the treating physicians during the trial. Determination of NT-proBNP in serum was performed using an e601 module on a COBAS 6000 line with an Electrochemiluminescent Immunoassay (ECLIA) kit from Roche Diagnostics (Rotkreuz, Switzerland). The measuring range extended from 5.0 to 35,000 pg/ml. Samples with values above the measuring range were diluted accordingly. Functional sensitivity was at 50 pg/ml and expected normal values were <125 pg/ml. Between-run precision at concentrations of 138 and 4,420 pg/ml was respectively 3.6% and 3.0%. Serum levels of NSE were determined as previously described.
We assessed the association between NT-proBNP and mortality and neurological status at 6 months defined by the Cerebral Performance Category score (CPC) and modified Rankin Scale (mRS). The CPC scale ranges from 1 to 5; 1 represents good cerebral performance or minor disability; 2, moderate disability; and 3, severe disability; 4, coma or vegetative state; and 5, death. The mRS range from 0 to 6, with 0 representing no symptoms; 1, no clinically significant disability; 2, slight disability; 3, moderate disability; 4, moderately severe disability; and 5, severe disability; with 6 representing death. Good neurological function was defined as CPC 1 to 2/mRS 0 to 3 and poor as CPC 3 to 5/mRS 4 to 6. Presumed cause of death was determined clinically by the treating physicians.
The TTM trial is registered at ClinicalTrials.gov (Identifier: NCT01020916 ) and ethical committees in each participating country approved the protocol.
Data are presented as mean ± SD for normally distributed data, and median and 25th and 75th percentile for skewed data. Categorical data are presented as numbers and percentage. Differences in baseline demographic variables were compared using the one-way analysis of variance and chi-square (ChiSq) tests as appropriate. The NT-proBNP levels were compared between the TTM groups using repeated-measures mixed models with an unstructured covariance structure. The models included the TTM allocation group, time, and the interaction term between TTM group and time. The fixed type-3 effect of the interaction between the TTM group and time was reported. Kaplan–Meier plots and log-rank tests were used for survival analysis in patients stratified in quartiles by serum levels of NT-proBNP 24 hours after ROSC and TTM allocation group. Logistic regression was used when assessing the association levels of NT-proBNP on neurological outcome. Variables adjusted for in the multivariable models (model 1) were NT-proBNP levels >/< median, TTM allocation (33°C vs 36°C), age, gender, time to ROSC, lactate on admission, bystander cardiopulmonary resuscitation, initial rhythm (shockable vs nonshockable), creatinine on day 1, ST-elevation myocardial infarction, shock on admission, body mass index, co-morbidities (0 vs ≥1 co-morbidity), reported heart rate at day 1, and the cardiovascular subgroup of the Sequential Organ Failure Assessment score. In addition, we added NSE levels >/< median, measured 24 hours after ROSC, to the models (model 2). As model control, we found a linear relation between the logarithmic transformed variables (NT-proBNP and NSE) and the outcome in the logistic regression analysis. When examining the predictive value of NT-proBNP and NSE on causes of death, we used C-statistics described by area under the receiver-operating characteristic curves (AUCs). Comparison of AUC was made using the formula ChiSq = (AUC NT-proBNP − AUC NSE ) 2 /(standard error [SE] NT-proBNP 2 − SE NSE 2 ). As NSE is a strong predictor of mortality in comatose OHCA patients, we stratified patients by NSE quartiles and subsequent by NT-proBNP median serum levels to examine any potential additive prognostic information. Data were analyzed assuming that missing data were missing at random, and thus, in the univariable and multivariable logistic regression analyses, data were handled by multiple imputation by chained equations to minimize bias and increase statistical power. Data analysis was performed using SAS software (Enterprise, version 5.1; SAS institute Inc., Cary, North Carolina). R 3.1.2 ( http://www.R-project.org/ ) was used for multiple imputation, using the “mice” ( http://cran.r-project.org/web/packages/mice/citation.html ) and “Hmisc” (Harrell FE [2014]: Hmisc: A package of miscellaneous R functions. Programs available from http://biostat.mc.vanderbilt.edu/Hmisc ) packages. A 2-sided p value <0.05 was considered statistical significant.
Results
Of the 647 patients, 638 (99%) had elevated NT-proBNP serum concentrations ≥125 pg/ml 24 hours after ROSC. Increasing quartiles of NT-proBNP measured were associated more unfavorable prearrest (more co-morbidities), intraarrest (e.g., longer time to ROSC and initial shockable rhythm), and postarrest conditions (increasing total and cardiovascular Sequential Organ Failure Assessment score) were associated with increasing NT-proBNP quartiles ( Table 1 ). As the target temperature of TTM affected the NT-proBNP levels, patients were stratified into quartiles based on the NT-proBNP levels within allocated TTM group ( Figures 1 and 2 ).
| BNP Q1 (466 pg/ml) | BNP Q2 (1140 pg/ml) | BNP Q3 (2308 pg/ml) | BNP Q4 (6584 pg/ml) | ||
|---|---|---|---|---|---|
| TTM33 (pg/ml); min – max range | 52 – 696 | 703 – 1433 | 1512 – 3054 | 3058 – 89382 | P |
| TTM36 (pg/ml); min – max range | 70 – 971 | 977 – 1905 | 1914 – 3744 | 3752 – 38541 | |
| Distribution | 161 (25%) | 162 (25%) | 163 (25%) | 161 (25%) | |
| Randomization Target Temperature Management (36°C) | 82 (51%) | 82 (51%) | 83 (51%) | 82 (51%) | ns |
| Sex (male) | 145 (90%) | 130 (80%) | 134 (82%) | 115 (71%) | <0.001 |
| Age (years), mean (SD) | 56.9 (13.1) | 62.7 (11.1) | 66.8 (10.5) | 68.5 (5) | <0.0001 |
| Body mass index (Kg/m 2 ), mean (SD) | 26.2 (4.0) | 26.7 (4.7) | 26.8 (4.6) | 26.6 (4.7) | ns |
| Shockable rhythm | 132 (85%) | 132 (84%) | 131 (84%) | 113 (72%) | <0.01 |
| Unknown initial rhythm | 6 (4%) | 4 (3%) | 7 (4%) | 3 (2%) | ns |
| ST-Elevation Myocardial Infarction | 69 (43%) | 84 (52%) | 73 (45%) | 72 (45%) | ns |
| Congestive Heart Failure | 2 (1%) | 5 (3%) | 11 (7%) | 21 (13%) | <0.0001 |
| Ischemic Heart Disease | 29 (18%) | 34 (21%) | 48 (30%) | 68 (43%) | <0.0001 |
| Previous Myocardial Infarction | 21 (13%) | 27 (17%) | 36 (22%) | 45 (28%) | <0.01 |
| Hypertension | 53 (33%) | 64 (40%) | 74 (45%) | 70 (44%) | ns |
| Previous Transient Ischemic Attack or stroke | 10 (6%) | 8 (5%) | 11 (7%) | 20 (12%) | ns |
| Epilepsy | 4 (3%) | 0 (0%) | 4 (3%) | 3 (2%) | ns |
| Diabetes | 21 (13%) | 25 (16%) | 20 (12%) | 33 (21%) | ns |
| Asthma or Chronic Obstructive Pulmonary Disease | 12 (8%) | 17 (11%) | 19 (12%) | 18 (11%) | ns |
| Chronic dialysis | 1 (0.6%) | 0 (0.0%) | 0 (0.0%) | 4 (3%) | <0.05 |
| Cirrhosis | 0 (0.0%) | 0 (0.0%) | 1 (0.6%) | 0 (0.0%) | ns |
| Malignancy | 3 (2%) | 5 (3%) | 5 (3%) | 12 (8%) | <0.05 |
| Previous Percutaneous Coronary Intervention | 15 (9%) | 16 (10%) | 20 (12%) | 31 (19%) | <0.05 |
| Previous Coronary artery bypass grafting | 3 (2%) | 12 (7%) | 18 (11%) | 25 (16%) | <0.001 |
| Alcoholism | 8 (5%) | 5 (3%) | 8 (5%) | 5 (3%) | ns |
| Place of cardiac arrest (public) | 76 (47%) | 79 (49%) | 75 (46%) | 48 (30%) | <0.01 |
| Witnessed arrest | 145 (90%) | 148 (91%) | 145 (89%) | 141 (88%) | ns |
| Bystander Cardiopulmonary Resuscitation | 123 (76%) | 113 (70%) | 121 (74%) | 113 (70%) | ns |
| Bystander Defibrillation | 23 (14%) | 16 (10%) | 14 (9%) | 16 (10%) | ns |
| Time to Advanced Life Support (minutes), median (25; 75) | 9 (6; 12) | 9 (5; 12) | 9 (6; 13) | 9 (6; 13) | ns |
| Time to Return of Spontaneous Circulation (minutes), median (25; 75) | 20 (15; 30) | 25 (15; 35) | 25 (18; 40) | 30 (17; 48) | 0.0001 |
| Shock on admission | 10 (6%) | 15 (9%) | 24 (15%) | 32 (20%) | <0.01 |
| Arterial lactate concentration (mM), mean (SD) | 6.2 (4.0) | 6.8 (4.9) | 6.4 (4.3) | 7.3 (4.3) | ns |
| Creatinine on admission (mg/dL), mean (SD) | 0.83 (0.42) | 0.92 (0.44) | 1.14 (0.61) | 1.60 (0.94) | <0.0001 |
| Sequential Organ Failure Assessment score day 1, mean (SD) | 9.3 (2.1) | 9.8 (2.0) | 10.4 (2.2) | 10.6 (2.2) | <0.0001 |
| Cardiovascular Sequential Organ Failure Assessment subscore day 1, mean (SD) | 2.3 (1.4) | 2.6 (1.3) | 2.7 (1.5) | 2.7 (1.4) | <0.05 |
NT-proBNP serum levels increased from 24 to 72 hours after ROSC in both TTM allocation groups ( Figure 1 ). After 24 hours during TTM intervention, patients in the TTM33 group had significantly lower median NT-proBNP serum level (1,473 [703; 3,054] pg/ml) than those in the TTM36 group (1,914 [977; 3,744], p <0.01). After 48 and 72 hours, the differences between the 2 TTM allocation groups were no longer statistically significant. Creatinine levels and heart rate on day 1 after ROSC were significantly lower in patients in the TTM33 group (mean [SD]): creatinine: 1.03 (0.59) mg/dl; heart rate: 67 (17) per minute than those in the TTM36 group (creatinine: 1.21 [0.78] mg/dl; heart rate: 77 [20] per minute), p creatinine <0.001; p heart rate <0.0001. When adjusting for creatinine and heart rate during TTM, the effect on NT-proBNP by the level of TTM was no longer statistically significant (data not shown).
Patients dying during follow-up (n = 298) had significantly higher NT-proBNP levels 24 hours after ROSC (2,514 [1,264; 5,226] pg/ml) than those surviving (n = 349, 1,127 [579; 2,118] pg/ml), p <0.0001. In addition, patients in shock on admission (n = 81) had significantly higher serum levels of NT-proBNP (1,914 [977; 3,744] pg/ml) than patients without shock (1,473 [703; 3,054] pg/ml), p <0.001. Kaplan–Meier curves showed that the probability of survival decreased significantly with increasing quartiles of NT-proBNP serum levels measured 24 hours after ROSC in both TTM allocation groups, p logrank <0.0001 ( Figure 2 ). Patients dying from a neurological cause (n = 183) had lower NT-proBNP levels measured 24 hours after ROSC: 2011 (1,088; 3,752) pg/ml, compared to 4,647 (2,331; 10,682) pg/ml in patients dying from a cardiovascular cause (n = 97), p <0.0001, 24 hours after ROSC. Levels of NT-proBNP 24 hours after ROSC had higher predictive value of cardiovascular cause of death (AUC NT-proBNP = 0.79) than NSE (AUC NSE = 0.63), p <0.0001. In contrast, NSE serum levels 24 hours after ROSC had higher predictive value of neurological cause of death (AUC NSE = 0.77) than NT-proBNP serum levels (AUC NT-proBNP = 0.58), p <0.0001 ( Figure 3 ). Forty-eight hours after ROSC, predictive value of NT-proBNP serum levels on cardiovascular death (AUC NT-proBNP = 0.78) remained significantly higher than NSE serum levels (AUC NSE = 0.64), p <0.001, whereas NSE serum levels remained more predictive on neurological cause of death (AUC NSE = 0.89) than NT-proBNP serum levels (AUC NT-proBNP = 0.42), p <0.0001. At 72 hours, the signal remained unaltered; cardiovascular cause of death (AUC NT-proBNP = 0.74 vs AUC NSE = 0.65, p = 0.07) and neurological cause of death (AUC NSE = 0.88 vs AUC NT-proBNP = 0.55, p <0.0001).
When stratifying patients into quartiles of NSE, patients with NT-proBNP levels < the median level had significantly lower risk of death than those with levels > the median, except in the highest quartile ( Figure 4 ). In addition, we found a statistical significant interaction between NT-proBNP and NSE, p <0.01.
In the multivariable model 1, NT-proBNP levels > median measured 24 hours after ROSC were associated with an increased risk of poor neurological outcome ( Table 2 ). When NSE was added to the model (model 2), NT-proBNP levels > median remained independently associated with increased odds of poor neurological outcome.
| Cerebral Performance Category Score | Modified Rankin Scale | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 1 | Model 2 | ||||||||||||
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | ||||
| NT-proBNP > median | 2.02 | (1.34-3.05) | <0.001 | 1.80 | (1.18-2.76) | <0.01 | 2.28 | (1.50-3.46) | <0.001 | 2.03 | (1.32-3.12) | <0.01 | |||
| Target Temperature Management (36°C) | 0.79 | (0.53-1.18) | ns | 0.83 | (0.54-1.26) | ns | 0.92 | (0.61-1.39) | ns | 0.95 | (0.62-1.46) | ns | |||
| Age per year | 1.06 | (1.04-1.08) | <0.0001 | 1.06 | (1.04-1.08) | <0.0001 | 1.06 | (1.04-1.08) | <0.0001 | 1.06 | (1.04-1.08) | <0.0001 | |||
| Male sex | 0.62 | (0.37-1.04) | ns | 0.61 | (0.36-1.04) | ns | 0.67 | (0.40-1.12) | ns | 0.67 | (0.39-1.15) | ns | |||
| Time to Return of Spontaneous Circulation | 1.03 | (1.01-1.04) | <0.0001 | 1.02 | (1.01-1.03) | <0.01 | 1.03 | (1.01-1.04) | <0.0001 | 1.02 | (1.01-1.03) | <0.01 | |||
| Lactate | 1.03 | (0.98-1.08) | ns | 1.01 | (0.96-1.06) | ns | 1.04 | (0.99-1.10) | ns | 1.03 | (0.98-1.08) | ns | |||
| Bystander Cardiopulmonary Resuscitation | 0.57 | (0.36-0.89) | <0.05 | 0.61 | (0.39-0.97) | <0.05 | 0.52 | (0.33-0.82) | <0.05 | 0.56 | (0.35-0.88) | <0.05 | |||
| Shockable rhythm | 0.19 | (0.10-0.35) | <0.0001 | 0.21 | (0.11-0.40) | <0.0001 | 0.19 | (0.10-0.35) | <0.0001 | 0.20 | (0.11-0.38) | <0.0001 | |||
| Creatinine | 2.04 | (1.39-3.00) | <0.001 | 1.83 | (1.24-2.69) | <0.01 | 2.01 | (1.37-2.96) | <0.001 | 1.78 | (1.21-2.61) | <0.01 | |||
| ST-Elevation Myocardial Infarction | 0.94 | (0.63-1.39) | ns | 0.76 | (0.50-1.16) | ns | 0.92 | (0.62-1.37) | ns | 0.76 | (0.50-1.16) | ns | |||
| Heart Rate at 28 hours | 1.01 | (1.00-1.02) | ns | 1.01 | (1.00-1.02) | ns | 1.01 | (1.00-1.02) | ns | 1.01 | (0.99-1.02) | ns | |||
| Shock on admission | 1.20 | (0.64-2.26) | ns | 1.07 | (0.56-2.04) | ns | 0.96 | (0.51-1.79) | ns | 0.86 | (0.45-1.64) | ns | |||
| Body Mass Index | 1.00 | (0.95-1.06) | ns | 1.01 | (0.96-1.06) | ns | 1.01 | (0.96-1.06) | ns | 1.01 | (0.96-1.06) | ns | |||
| Comorbidities ≥ 1 | 1.12 | (0.72-1.73) | ns | 1.18 | (0.75-1.87) | ns | 1.14 | (0.73-1.78) | ns | 1.18 | (0.74-1.87) | ns | |||
| Cardiovascular Sequential Organ Failure Assessment score day 1 | 0.89 | (0.77-1.03) | ns | 0.84 | (0.72-0.98) | <0.05 | 0.90 | (0.77-1.06) | ns | 0.84 | (0.72-0.98) | <0.05 | |||
| Neuron Specific Enolase > median | – | – | – | – | 3.38 | (2.15-5.31) | <0.0001 | – | – | – | – | 3.39 | 2.15 | 5.33 | <0.0001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
