G Model RESUS-5962; No. of Pages 7
ARTICLE IN PRESS Resuscitation xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia夽 Antonio Maria Dell’anna, Julia Bini Vinotti, Marjorie Beumier, Diego Orbegozo-Cortes, Katia Donadello, Sabino Scolletta, Jean-Louis Vincent, Fabio Silvio Taccone ∗ Department of Intensive Care, Erasme Hospital, Université Libre de Bruxelles, Route de Lennik, 808, 1070 Brussels, Belgium
a r t i c l e
i n f o
Article history: Received 28 December 2013 Received in revised form 31 March 2014 Accepted 3 April 2014 Keywords: Cardiac arrest Biomarker Inflammation Infection Prognosis
a b s t r a c t Aim: Prognostication of outcome after cardiac arrest (CA) is challenging. We assessed the prognostic value of daily blood levels of C-reactive protein (CRP), a cheap and widely available inflammatory biomarker, after CA. Methods: We reviewed the data of all patients admitted to our intensive care unit (ICU) after CA between January 2009 and December 2011 and who survived for at least 24 h. We collected demographic data, CA characteristics (initial rhythm; location of arrest; time to return of spontaneous circulation [ROSC]), occurrence of infection, ICU survival and neurological outcome at three months (good = cerebral performance category [CPC] 1–2; poor = CPC 3–5). CRP levels were measured daily from admission to day 3. Results: A total of 130 patients were admitted after successful resuscitation from CA and survived more than 24 h; 76 patients (58%) developed an infection and overall mortality was 56%. CRP levels increased from admission to day 3. CRP levels were higher in in-hospital than in out-of-hospital CA, especially on admission and day 1 (44.1 vs. 2.1 mg L−1 and 74.5 vs. 29.5 mg L−1 , respectively; p < 0.001), and in patients with non-shockable than in those with shockable rhythms. In a logistic regression model, high CRP levels on admission were independently associated with poor neurological outcome at 3 months. Conclusion: CRP levels increase in the days following successful resuscitation of CA. Higher CRP levels in patients with in-hospital CA, non-shockable rhythms and infection, suggest a greater inflammatory response in these patients. High CRP levels on admission may identify patients at high-risk of poor outcome and could be a target for future therapies. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Most patients successfully resuscitated from cardiac arrest (CA) remain comatose and are treated with mild therapeutic hypothermia (TH) to reduce the extent of post-anoxic brain injury.1,2 Despite advances in the treatment of such patients, substantial neurological deficits impact on long-term outcomes in survivors.3,4 As such, there has been increasing interest in the development of tools that could help assess prognosis in these patients and identify those at
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2014.04.003. ∗ Corresponding author at: Department of Intensive Care, Hôpital Erasme, Université Libre de Bruxelles (ULB), Route de Lennik 808, 1070 Brussels, Belgium. Fax: +32 25554698. E-mail address: [email protected] (F.S. Taccone).
high-risk of complications or those with extensive brain damage in whom the need for life-sustaining therapies could be considered.5 A systemic inflammatory response, which occurs after the return of spontaneous circulation (ROSC) because of the ischaemiareperfusion process, can significantly influence patient outcome.6 This so-called post-cardiac arrest syndrome (PCAS) shares several pathophysiological mechanisms with sepsis and may contribute to the burden of brain damage and impair cardiac or respiratory functions, eventually leading to multi-organ failure (MOF).7,8 Post-anoxic brain injury can be monitored clinically or by electrophysiological or imaging tools,9,10 but PCAS is more difficult to evaluate, especially at an early stage. Biomarkers of inflammation, such as procalcitonin (PCT) and Creactive protein (CRP), have been investigated in this setting. CRP is a pentraxin released into the bloodstream by the liver during the acute-phase response, as part of an inflammatory reaction.11 When the primary inflammatory stimulus is removed, CRP decreases
http://dx.doi.org/10.1016/j.resuscitation.2014.04.003 0300-9572/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Dell’anna AM, et al. C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia. Resuscitation (2014), http://dx.doi.org/10.1016/j.resuscitation.2014.04.003
G Model RESUS-5962; No. of Pages 7
ARTICLE IN PRESS A.M. Dell’anna et al. / Resuscitation xxx (2014) xxx–xxx
2
with a variable half-life, estimated at around 20 h.12 Oppert et al. reported that CRP levels increased after CA, irrespective of the cause of arrest and the estimated time of anoxia, but did not reliably predict outcome. In their study, CRP levels did not differ in patients with and without infectious complications within the first days after hospital admission, but the study cohort included only 22 post-CA patients and 7 healthy volunteers.13 Some of these methodological limitations were also present in other studies that measured CRP levels in patients resuscitated from CA.14–16 PCT levels have also been correlated with poor prognosis after CA.17–19 The aim of the present study was to investigate the time profile of CRP levels in patients treated with TH after CA. Also, we assessed the diagnostic value for infection and the potential prognostic value in reconsidering or withdrawing therapies. 2. Patients and methods 2.1. Study population We retrospectively analyzed data for all patients admitted to the medical/surgical Dept. of Intensive Care of Erasme Hospital (Brussels, Belgium) between January 2009 and December 2011, who had been resuscitated from CA, survived at least 24 h, and remained comatose (Glasgow Coma score, GCS < 9). The local Ethical Committee approved the study, but waived the need for informed consent because the study was purely observational. 2.2. Post-resuscitation care All comatose patients were treated with TH (temperature set at 32–34 ◦ C) for 24 h, according to a standardized institutional protocol. Cooling is started immediately after ROSC has occurred (for in-hospital CA) or after hospital admission (for out-of-hospital CA [OHCA]) for patients who remain comatose. TH is induced with a cold fluid bolus (of either saline or a balanced solution given as a bolus of 20–30 ml kg−1 over 30 min) and maintained for 24 h using a circulating water blanket device (Medi-Therm II, Gamar, USA). Body temperature is initially measured using a rectal temperature probe, replaced by blood temperature monitoring as soon as an invasive haemodynamic monitoring device (PiCCO, Pulsion, Munich, Germany) is placed. Midazolam is administered for sedation (continuous infusion 0.03–0.1 mg kg−1 h−1 ) and morphine for analgesia (0.1–0.3 mg kg−1 h−1 ) during TH. Neuromuscular blocking agents (NMBAs) are administered to control shivering in the induction phase (cisatracurium as a bolus of 0.15 mg kg−1 ) and, if needed, as a continuous infusion thereafter (1–3 mcg kg−1 min−1 ). Rewarming (0.5 ◦ C h−1 ) is obtained passively after interruption of cooling, and sedation, analgesia and NMBAs are discontinued when body temperature is >37 ◦ C. Patients are kept in a 30◦ semi-recumbent position; ventilation is set to target PaCO2 between 35 and 45 mmHg and SpO2 >94%. Trans-oesophageal echocardiography is performed within the first 8–12 h after ICU admission. Blood glucose is kept between 110 and 150 mg dL−1 . Mean arterial pressure is maintained >65–70 mm Hg using volume resuscitation, dobutamine and/or norepinephrine, whenever needed. Intra-aortic balloon counterpulsation (IABP) or extracorporeal membrane oxygenation (ECMO) is used in cases of severe cardiogenic shock. After rewarming the patient and stopping sedative agents, repeated neurologic examination and standard or continuous EEG is performed. Withdrawal of care is an interdisciplinary decision based on bilateral absence of the N20 response to sensory evoked potentials (SEPs), persisting coma, or presence of status myoclonus or refractory status epilepticus.
2.3. Data collection Demographics, co-morbidities, data on cardiopulmonary resuscitation (CPR: first rhythm, bystander CPR, time to ROSC) were collected in all patients from hospital (if in-hospital) or ambulance (if OHCA) records. CRP levels were routinely measured daily using an immunological test (Cobas, Roche Diagnostics, Mannheim, Germany). Values were collected from ICU admission until day 3. The highest (peak) CRP level during this time period was recorded for each patient. Changes in CRP levels over the first day were measured as the difference in levels on admission and on the first day of ICU stay (“delta CRP”). The presence of infection was defined as the combination of a clinical or radiological sign of an infectious process associated with a positive culture from a normally sterile site (e.g., blood culture, broncho-alveolar lavage [BAL] or endotracheal aspirate) and the need for antibiotic administration. The site of infection was diagnosed according to the CDC/NHSN criteria.20 These data were reviewed by two authors (MB and FST), who were blinded to the CRP levels. In case of disagreement, a final decision was resolved by consensus (MB, AMD and FST). Only infections occurring during the first 3 days following ICU admission were considered. Neurological evaluation at hospital discharge and at 3 months, assessed using the cerebral performance category score (CPC; 1 = no neurological disability, 2 = mild neurological disability, 3 = severe neurological impairment, 4 = vegetative state, 5 = death), was recorded. A favourable outcome was defined as a CPC of 1 or 2. 2.4. Statistical analysis Statistical analyses were performed using IBM SPSS Statistics 21 for Windows. Descriptive statistics were computed for all study variables. A Kolmogorov–Smirnov test was used, and histograms and normal-quartile plots were examined to verify the normality of distribution of continuous variables. Data are presented as count (percentage) or median [25th–75th percentiles]. Differences between groups were assessed using a Fisher’s exact test for categorical variables and a Wilcoxon rank test for continuous variables. Data from repeated measures were analyzed using a two-way Friedman ANOVA because the data were not all normally distributed, and differences between groups for each time point were explored by a Mann–Whitney U test for non-parametric data. Correlation between continuous variables was evaluated using the Spearman rank correlation, with 95% confidence intervals. Multivariate binomial logistic regression was used to assess whether admission CRP levels or peak CRP levels were independently associated with a favourable neurological outcome at 3 months. The discriminative ability of variables identified by multivariate analysis as being associated with neurological outcome was evaluated using receiver operating characteristic (ROC) curves with the corresponding area under the curve (AUC) and related sensitivity and specificity. A p < 0.05 was considered as statistically significant. 3. Results 3.1. Patient characteristics and CRP measurement During the study period, 166 patients were admitted to the ICU after successfully resuscitated CA; 36 patients were excluded from the study because they survived less than 24 h. Of the 130 patients included in the final cohort, 40 had in-hospital and 90 out-of-hospital CA (Table 1). In the 76 patients with infection, the median time to antibiotic initiation was 1 [0–2] day. CRP samples were collected from all 130 patients on admission and on day 1, and
Please cite this article in press as: Dell’anna AM, et al. C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia. Resuscitation (2014), http://dx.doi.org/10.1016/j.resuscitation.2014.04.003
G Model RESUS-5962; No. of Pages 7
ARTICLE IN PRESS A.M. Dell’anna et al. / Resuscitation xxx (2014) xxx–xxx
3
Table 1 Characteristics of patients on ICU admission. Data are expressed as median [interquartile range] and counts (percentage). Favourable neurological outcome was defined as a CPC value of 1 or 2, unfavourable a CPC of 3–5.
Age, years Male gender, n (%) OHCA, n (%) Witnessed, n (%) Bystander CPR, n (%) VF/VT, n (%) Asystole/PEA, n (%) Infection, n (%) Time to ROSC, min Admission CRP, mg L−1 Peak CRP, mg L−1
All patients (n = 130)
Survivors (n = 57)
Non-survivors (n = 73)
p-Value
Favourable neurological outcome (n = 46)
Unfavourable neurological outcome (n = 84)
p-Value
62 [51–75] 102 (78.5%) 90 (69%) 99 (76%) 63 (48.5%) 56 (43%) 74 (56%) 76 (58%) 16.50 [10–25] 6.30 [1.3–42.5] 150 [98.5–210]
54 [48.5–68.5] 51 (89%) 40 (70%) 51 (89%) 36 (63%) 38(67%) 19 (33%) 40 (70%) 15 [9–25] 3.4 [0.9–21.5] 170 [120–215]
67 [54–76] 51 (70%) 50 (68%) 48 (66%) 26 (36%) 18 (25%) 55 (75%) 36 (49%) 20 [10–25] 10.2 [1.5–49.5] 120 [61.5–200]
0.005 0.009 0.85 0.002 0.002
ARTICLE IN PRESS Resuscitation xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia夽 Antonio Maria Dell’anna, Julia Bini Vinotti, Marjorie Beumier, Diego Orbegozo-Cortes, Katia Donadello, Sabino Scolletta, Jean-Louis Vincent, Fabio Silvio Taccone ∗ Department of Intensive Care, Erasme Hospital, Université Libre de Bruxelles, Route de Lennik, 808, 1070 Brussels, Belgium
a r t i c l e
i n f o
Article history: Received 28 December 2013 Received in revised form 31 March 2014 Accepted 3 April 2014 Keywords: Cardiac arrest Biomarker Inflammation Infection Prognosis
a b s t r a c t Aim: Prognostication of outcome after cardiac arrest (CA) is challenging. We assessed the prognostic value of daily blood levels of C-reactive protein (CRP), a cheap and widely available inflammatory biomarker, after CA. Methods: We reviewed the data of all patients admitted to our intensive care unit (ICU) after CA between January 2009 and December 2011 and who survived for at least 24 h. We collected demographic data, CA characteristics (initial rhythm; location of arrest; time to return of spontaneous circulation [ROSC]), occurrence of infection, ICU survival and neurological outcome at three months (good = cerebral performance category [CPC] 1–2; poor = CPC 3–5). CRP levels were measured daily from admission to day 3. Results: A total of 130 patients were admitted after successful resuscitation from CA and survived more than 24 h; 76 patients (58%) developed an infection and overall mortality was 56%. CRP levels increased from admission to day 3. CRP levels were higher in in-hospital than in out-of-hospital CA, especially on admission and day 1 (44.1 vs. 2.1 mg L−1 and 74.5 vs. 29.5 mg L−1 , respectively; p < 0.001), and in patients with non-shockable than in those with shockable rhythms. In a logistic regression model, high CRP levels on admission were independently associated with poor neurological outcome at 3 months. Conclusion: CRP levels increase in the days following successful resuscitation of CA. Higher CRP levels in patients with in-hospital CA, non-shockable rhythms and infection, suggest a greater inflammatory response in these patients. High CRP levels on admission may identify patients at high-risk of poor outcome and could be a target for future therapies. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Most patients successfully resuscitated from cardiac arrest (CA) remain comatose and are treated with mild therapeutic hypothermia (TH) to reduce the extent of post-anoxic brain injury.1,2 Despite advances in the treatment of such patients, substantial neurological deficits impact on long-term outcomes in survivors.3,4 As such, there has been increasing interest in the development of tools that could help assess prognosis in these patients and identify those at
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2014.04.003. ∗ Corresponding author at: Department of Intensive Care, Hôpital Erasme, Université Libre de Bruxelles (ULB), Route de Lennik 808, 1070 Brussels, Belgium. Fax: +32 25554698. E-mail address: [email protected] (F.S. Taccone).
high-risk of complications or those with extensive brain damage in whom the need for life-sustaining therapies could be considered.5 A systemic inflammatory response, which occurs after the return of spontaneous circulation (ROSC) because of the ischaemiareperfusion process, can significantly influence patient outcome.6 This so-called post-cardiac arrest syndrome (PCAS) shares several pathophysiological mechanisms with sepsis and may contribute to the burden of brain damage and impair cardiac or respiratory functions, eventually leading to multi-organ failure (MOF).7,8 Post-anoxic brain injury can be monitored clinically or by electrophysiological or imaging tools,9,10 but PCAS is more difficult to evaluate, especially at an early stage. Biomarkers of inflammation, such as procalcitonin (PCT) and Creactive protein (CRP), have been investigated in this setting. CRP is a pentraxin released into the bloodstream by the liver during the acute-phase response, as part of an inflammatory reaction.11 When the primary inflammatory stimulus is removed, CRP decreases
http://dx.doi.org/10.1016/j.resuscitation.2014.04.003 0300-9572/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Dell’anna AM, et al. C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia. Resuscitation (2014), http://dx.doi.org/10.1016/j.resuscitation.2014.04.003
G Model RESUS-5962; No. of Pages 7
ARTICLE IN PRESS A.M. Dell’anna et al. / Resuscitation xxx (2014) xxx–xxx
2
with a variable half-life, estimated at around 20 h.12 Oppert et al. reported that CRP levels increased after CA, irrespective of the cause of arrest and the estimated time of anoxia, but did not reliably predict outcome. In their study, CRP levels did not differ in patients with and without infectious complications within the first days after hospital admission, but the study cohort included only 22 post-CA patients and 7 healthy volunteers.13 Some of these methodological limitations were also present in other studies that measured CRP levels in patients resuscitated from CA.14–16 PCT levels have also been correlated with poor prognosis after CA.17–19 The aim of the present study was to investigate the time profile of CRP levels in patients treated with TH after CA. Also, we assessed the diagnostic value for infection and the potential prognostic value in reconsidering or withdrawing therapies. 2. Patients and methods 2.1. Study population We retrospectively analyzed data for all patients admitted to the medical/surgical Dept. of Intensive Care of Erasme Hospital (Brussels, Belgium) between January 2009 and December 2011, who had been resuscitated from CA, survived at least 24 h, and remained comatose (Glasgow Coma score, GCS < 9). The local Ethical Committee approved the study, but waived the need for informed consent because the study was purely observational. 2.2. Post-resuscitation care All comatose patients were treated with TH (temperature set at 32–34 ◦ C) for 24 h, according to a standardized institutional protocol. Cooling is started immediately after ROSC has occurred (for in-hospital CA) or after hospital admission (for out-of-hospital CA [OHCA]) for patients who remain comatose. TH is induced with a cold fluid bolus (of either saline or a balanced solution given as a bolus of 20–30 ml kg−1 over 30 min) and maintained for 24 h using a circulating water blanket device (Medi-Therm II, Gamar, USA). Body temperature is initially measured using a rectal temperature probe, replaced by blood temperature monitoring as soon as an invasive haemodynamic monitoring device (PiCCO, Pulsion, Munich, Germany) is placed. Midazolam is administered for sedation (continuous infusion 0.03–0.1 mg kg−1 h−1 ) and morphine for analgesia (0.1–0.3 mg kg−1 h−1 ) during TH. Neuromuscular blocking agents (NMBAs) are administered to control shivering in the induction phase (cisatracurium as a bolus of 0.15 mg kg−1 ) and, if needed, as a continuous infusion thereafter (1–3 mcg kg−1 min−1 ). Rewarming (0.5 ◦ C h−1 ) is obtained passively after interruption of cooling, and sedation, analgesia and NMBAs are discontinued when body temperature is >37 ◦ C. Patients are kept in a 30◦ semi-recumbent position; ventilation is set to target PaCO2 between 35 and 45 mmHg and SpO2 >94%. Trans-oesophageal echocardiography is performed within the first 8–12 h after ICU admission. Blood glucose is kept between 110 and 150 mg dL−1 . Mean arterial pressure is maintained >65–70 mm Hg using volume resuscitation, dobutamine and/or norepinephrine, whenever needed. Intra-aortic balloon counterpulsation (IABP) or extracorporeal membrane oxygenation (ECMO) is used in cases of severe cardiogenic shock. After rewarming the patient and stopping sedative agents, repeated neurologic examination and standard or continuous EEG is performed. Withdrawal of care is an interdisciplinary decision based on bilateral absence of the N20 response to sensory evoked potentials (SEPs), persisting coma, or presence of status myoclonus or refractory status epilepticus.
2.3. Data collection Demographics, co-morbidities, data on cardiopulmonary resuscitation (CPR: first rhythm, bystander CPR, time to ROSC) were collected in all patients from hospital (if in-hospital) or ambulance (if OHCA) records. CRP levels were routinely measured daily using an immunological test (Cobas, Roche Diagnostics, Mannheim, Germany). Values were collected from ICU admission until day 3. The highest (peak) CRP level during this time period was recorded for each patient. Changes in CRP levels over the first day were measured as the difference in levels on admission and on the first day of ICU stay (“delta CRP”). The presence of infection was defined as the combination of a clinical or radiological sign of an infectious process associated with a positive culture from a normally sterile site (e.g., blood culture, broncho-alveolar lavage [BAL] or endotracheal aspirate) and the need for antibiotic administration. The site of infection was diagnosed according to the CDC/NHSN criteria.20 These data were reviewed by two authors (MB and FST), who were blinded to the CRP levels. In case of disagreement, a final decision was resolved by consensus (MB, AMD and FST). Only infections occurring during the first 3 days following ICU admission were considered. Neurological evaluation at hospital discharge and at 3 months, assessed using the cerebral performance category score (CPC; 1 = no neurological disability, 2 = mild neurological disability, 3 = severe neurological impairment, 4 = vegetative state, 5 = death), was recorded. A favourable outcome was defined as a CPC of 1 or 2. 2.4. Statistical analysis Statistical analyses were performed using IBM SPSS Statistics 21 for Windows. Descriptive statistics were computed for all study variables. A Kolmogorov–Smirnov test was used, and histograms and normal-quartile plots were examined to verify the normality of distribution of continuous variables. Data are presented as count (percentage) or median [25th–75th percentiles]. Differences between groups were assessed using a Fisher’s exact test for categorical variables and a Wilcoxon rank test for continuous variables. Data from repeated measures were analyzed using a two-way Friedman ANOVA because the data were not all normally distributed, and differences between groups for each time point were explored by a Mann–Whitney U test for non-parametric data. Correlation between continuous variables was evaluated using the Spearman rank correlation, with 95% confidence intervals. Multivariate binomial logistic regression was used to assess whether admission CRP levels or peak CRP levels were independently associated with a favourable neurological outcome at 3 months. The discriminative ability of variables identified by multivariate analysis as being associated with neurological outcome was evaluated using receiver operating characteristic (ROC) curves with the corresponding area under the curve (AUC) and related sensitivity and specificity. A p < 0.05 was considered as statistically significant. 3. Results 3.1. Patient characteristics and CRP measurement During the study period, 166 patients were admitted to the ICU after successfully resuscitated CA; 36 patients were excluded from the study because they survived less than 24 h. Of the 130 patients included in the final cohort, 40 had in-hospital and 90 out-of-hospital CA (Table 1). In the 76 patients with infection, the median time to antibiotic initiation was 1 [0–2] day. CRP samples were collected from all 130 patients on admission and on day 1, and
Please cite this article in press as: Dell’anna AM, et al. C-reactive protein levels after cardiac arrest in patients treated with therapeutic hypothermia. Resuscitation (2014), http://dx.doi.org/10.1016/j.resuscitation.2014.04.003
G Model RESUS-5962; No. of Pages 7
ARTICLE IN PRESS A.M. Dell’anna et al. / Resuscitation xxx (2014) xxx–xxx
3
Table 1 Characteristics of patients on ICU admission. Data are expressed as median [interquartile range] and counts (percentage). Favourable neurological outcome was defined as a CPC value of 1 or 2, unfavourable a CPC of 3–5.
Age, years Male gender, n (%) OHCA, n (%) Witnessed, n (%) Bystander CPR, n (%) VF/VT, n (%) Asystole/PEA, n (%) Infection, n (%) Time to ROSC, min Admission CRP, mg L−1 Peak CRP, mg L−1
All patients (n = 130)
Survivors (n = 57)
Non-survivors (n = 73)
p-Value
Favourable neurological outcome (n = 46)
Unfavourable neurological outcome (n = 84)
p-Value
62 [51–75] 102 (78.5%) 90 (69%) 99 (76%) 63 (48.5%) 56 (43%) 74 (56%) 76 (58%) 16.50 [10–25] 6.30 [1.3–42.5] 150 [98.5–210]
54 [48.5–68.5] 51 (89%) 40 (70%) 51 (89%) 36 (63%) 38(67%) 19 (33%) 40 (70%) 15 [9–25] 3.4 [0.9–21.5] 170 [120–215]
67 [54–76] 51 (70%) 50 (68%) 48 (66%) 26 (36%) 18 (25%) 55 (75%) 36 (49%) 20 [10–25] 10.2 [1.5–49.5] 120 [61.5–200]
0.005 0.009 0.85 0.002 0.002
