- Email: [email protected]
In-hospital cardiac arrest after a rapid response team review: A matched case–control study
Resuscitation 126 (2018) 98–103
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
In-hospital cardiac arrest after a rapid response team review: A matched case–control study☆
T
⁎
Joonas Tirkkonena, , Heini Huhtalab, Sanna Hoppuc a
Department of Intensive Care Medicine, Tampere University Hospital and Department of Anaesthesiology and Intensive Care Medicine, Seinäjoki Central Hospital, PO Box 2000, FI-33521 Tampere, Finland b Faculty of Social Sciences, University of Tampere, FI-33014, Finland c Department of Intensive Care Medicine, Tampere University Hospital, PO Box 2000, FI-33521 Tampere, Finland
A R T I C LE I N FO
A B S T R A C T
Keywords: In-hospital cardiac arrest Rapid response team National early warning score
Aim: Study the incidence and reasons behind in-hospital cardiac arrests (IHCAs) after rapid response team (RRT) reviews. Methods: We conducted a matched case–control study at Tampere University Hospital, Finland. Data on adult patients who were triaged to remain on general ward after first (index) RRT review without treatment limitations but who suffered an IHCA within the following 48 h were prospectively collected for 5.3 years. These cases were matched (age ± 3 years, sex, surgical/medical ward, admission year) at a 1:4 ratio to controls (no ICHA after RRT review). Results: Of 2653 index RRT reviews, 17 patients suffered an IHCA on general ward within the 48 h after review. Their 30-day mortality rate was 88%. The incidence was 6.3/1000 index RRT reviews or 4.6/100,000 hospital admissions. Patients who suffered an IHCA within 48 h after RRT review were more likely to have a preceding ICU admission, and their median national early warning scores (NEWSs) at the end of the index RRT reviews (=last NEWSs) were higher than those of the controls. Higher last NEWS was the only factor associated with ICHA after RRT review (OR 1.22, 95% CI 1.00–1.49, p = 0.048) in a conditional multivariable regression model. Conclusions: IHCA within 48 h after an index RRT review on general ward is a rare event with poor prognosis. It is independently associated with higher NEWS at the end of the index RRT review. Careful consideration is stressed, when patients with high NEWS are left on ward after RRT reviews.
Introduction The incidence of adult in-hospital cardiac arrest (IHCA) is approximately 1.6 cases per 1000 hospital admissions [1]. Most IHCAs are preceded by derangements in vital signs, and the probability of survival to hospital discharge decreases as the number and severity of vital dysfunctions before the event increase [2,3]. If vital dysfunctions occur, rapid response teams (RRTs) review general ward patients to prevent IHCAs via minor interventions and timely transfers to intensive care [4]. Most RRT patients are triaged to remain on general ward without do not attempt resuscitation (DNAR) −orders [5,6]. One study that investigated the short-term outcomes among this RRT population found that 0.4% of no-DNAR patients triaged to remain on ward died during the following 24 h [6]. Is IHCA after an RRT review truly such a rare event, and is there something that the RRT misses during these reviews? In this matched
case–control study, we aimed 1) to report the incidence and prognosis of an IHCA occurring within 48 h after RRT triages a patient to remain on ward without a DNAR order and 2) to investigate thoroughly whether these cases (IHCA after RRT review) differed from controls (no IHCA after RRT review) at the time of their index RRT review. Methods Ethics The Ethics Committee of the Tampere University Hospital (Tays) approved the study protocol (Approval no: R10111, 2010). Patient consent was waived, as this was an observational study.
☆ ⁎
A Spanish translated version of the abstract of this article appears as Appendix in the final online version at https://doi.org/10.1016/j.resuscitation.2018.03.007. Corresponding author. E-mail addresses: [email protected].fi (J. Tirkkonen), Heini.Huhtala@staff.uta.fi (H. Huhtala), sanna.hoppu@pshp.fi (S. Hoppu).
https://doi.org/10.1016/j.resuscitation.2018.03.007 Received 10 October 2017; Received in revised form 4 February 2018; Accepted 5 March 2018 0300-9572/ © 2018 Elsevier B.V. All rights reserved.
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
Fig. 1. Study cohort. RRT, rapid response team; IHCA, in-hospital cardiac arrest; DNAR, do not attempt resuscitation; ICU, intensive care unit; OR, operation room; CT, computed tomography. a One medical patient under 50 years of age could not be matched for admission year (only six patients with similar age, sex and ward among the cohort of 1579 potential control patients).
(heart rate, systolic blood pressure, respiratory rate, peripheral blood oxygen saturation, body temperature, level of consciousness and the need for supplementary oxygen) is scored from 0 to 3 depending on how abnormal the measured value is. The total score is added up, and a score of ≥7 is considered a status that requires immediate RRT review [8].
Hospital Tays is one of five tertiary referral centres in Finland, with approximately 70,000 somatic admissions annually. The hospital provides the most advanced care for a catchment population of 1.1 million citizens. The intensive care unit (ICU) currently has 21 beds, with approximately 2100 admissions per year. There is a separate ICU with seven beds for surgical cardiothoracic patients.
Exclusion criteria RRT activations for paediatric patients (< 18 years) and index RRT reviews including an IHCA were excluded. RRT reviews for operating rooms, ICUs and the emergency room were further excluded, as were RRT reviews for outpatients. Finally, RRT reviews of patients with preexisting or new DNAR orders and RRT reviews resulting in ICU/operating room transfer were excluded.
Rapid response team The hospital’s rapid response system was implemented in January 2009. The ICU-operated RRT is physician led and the general ward staff utilise dichotomised RRT trigger criteria (heart rate < 40/min or > 140/min, systolic blood pressure < 90 mmHg, peripheral arteriolar oxygen saturation < 90%, respiratory rate < 5/min or > 24/min, decrease in state of consciousness and the ‘worried’ criterion). The RRT attends both medical emergencies and cardiac arrests 24/7. Although the RRT is physician-led, the RRT nurses may attend a patient by themselves (in collaboration with the parent unit physician) and only consult with the RRT physician by phone.
Data collection Prospective data collection began in 1 January 2010 as a part of an ongoing project involving RRT data gathering per the Utstein Style [9]. No previous study had documented the incidence of IHCAs review, so data for precise sample size calculations were not available. We aimed to collect data within a six-year period, as a longer study period would inevitably have affected the internal validity of the results. However, data collection was terminated eight months ahead of time, on 31 April 2015. Because only one case had occurred in the previous 12 months, our research resources were directed to another study. The initial data collection consisted of Utstein-style RRT templates [9], which were completed by the RRT nurses at scene. The RRT nurses recorded patient’s name and social security number, event location, RRT review reason with a short verbal comment on the preceding events, vital signs (which were measured multiple times during the reviews), all conducted examinations and interventions, possible
Definitions ‘Index RRT review’ refers to a patient’s first RRT review during his/ her hospitalisation. ‘Delayed RRT activation’ refers to a RRT activation that is preceded by recorded objective RRT activation criteria 20–360 min before the RRT review is triggered. Patients’ cumulative comorbidity was assessed with the Charlson comorbidity index (CCI) [7]. National early warning score (NEWS) is an early warning scoring system that has been implemented in all public hospitals in the United Kingdom [8]. NEWS provides a standardized method for the assessment and response to acute illness [8]. Each of the seven included vital signs 99
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
treatment limitations and follow-up plan/possible transfer to higher level of care. Further patient data and outcome data were retrieved from the electronic patient records. Controls were selected at 1:4 ratio (after being matched to cases by age ± 3 years, sex, surgical/medical ward and admission year), as is appropriate when there is no shortage of controls [10]. Controls were obtained from a sub population of RRT patients (without following IHCA) who did not have a DNAR order, did not receive a DNAR order, were not transferred to the ICU/operation room and did not die during the review (Fig. 1). When more than four matching controls were available, the first four controls were included in chronological order (according to the date of the RRT review).
Table 2 Univariable analysis of the characteristics of the cases (IHCA within 48 h after patient’s index RRT review on general ward, n = 17) compared with controls (no IHCA after index RRT review, n = 68). Cases %
OR
95% CI
p–value
76 (65, 81)
1.08
0.80–1.47
0.61
59 47 77 2.0 (1.0, 3.0) 37 13 8.8 8.8 15 28 25 15 4.4
– – 1.40 0.90
– – 0.37–5.26 0.69–1.17
–a –a 0.62 0.42
0.93 1.83 1.36 0.65 1.24 0.33 0.32 2.03 1.33
0.28–3.12 0.54–6.14 0.28–7.20 0.07–5.72 0.30–5.17 0.07–1.65 0.08–1.92 0.46–8.90 0.14–12.8
0.90 0.33 0.72 0.70 0.73 0.18 0.25 0.35 0.80
82 10
Ref. 2.74
0.65–11.5
0.17
5.9
7.4
1.17
0.10–13.3
0.90
65 12 24
84 4.4 12
Ref. 3.21 2.46
0.52–20.0 0.65–9.30
0.21 0.18
24 12 35 5.9
18 18 15 5.9
0.0
4.4
–
–
e
0.0 18 0.0 0.0 5.9
1.5 22 10 5.9 0.0
Age (median, Q1, Q3)
77 (63, 82) Sex (male) 59 Medical patient 47 Acute hospital admission 82 CCI (median, Q1, Q3) 1.0 (1.0, 2.0) Coronary artery disease 35 Congestive heart failure 24 Peripheral artery disease 12 Cerebrovascular disease 5.9 COPD 18 Diabetes 12 Malignancy b 12 Chronic renal failure 24 Dementia 5.9 Preceding performance in daily living Independent 71 Requires some assistance 24
Statistical analysis Data are presented as percentages, and continuous variables as medians (Q1, Q3). Conditional univariable logistic regression was used for statistical comparisons between cases and controls. Conditional multivariable logistic regression was used to investigate independent associations with cases. As is generally considered appropriate, the introduction of one variable into the multivariable model required at least five cases (outcomes) [11]. Tests were two-sided; p < 0.05 was considered significant and 95% confidence intervals are reported. Stata® version 13.0 (StataCorp LLP, Texas, USA) was used for the analyses, all of which were conducted in collaboration with a biostatistician.
c
Severely limited d Admitted to hospital from Home Nursing home Other hospital/Primary care ward Hospital admission diagnosis Infection Malignancy Cardiovascular Cerebrovascular (nonsurgical) Neurosurgical (other than tumors) Respiratory Gastrosurgical Trauma Other medical Other surgical
Results Cohort description During the 5.3-year study period, 17 IHCAs occurred within 48 h after RRT triaged these no-DNAR patients to remain on ward after their index RRT review (Fig. 1). Twelve (71%) of these events occurred within the first 24 h. During the study period, the incidence of IHCA within 48 h after index RRT review was 6.3 per 1000 index RRT reviews or 4.6 per 100,000 hospital admissions, and these IHCAs accounted for 3.6% of all IHCA patients. Table 1 presents the data related to the 17 cardiac arrest events, while Tables 2 and 3 present background data related to patient characteristics and the index RRT reviews of these cases. Two-thirds of the case patients had previously lived at home with normal performance in daily living. The median age of the patients was 77 years; 59% were male, and 47% had a medical reason for admission at the time of the IHCA. The cumulative comorbidity among the 17 IHCA patients was low, with a median Charlson comorbidity index score of one. The IHCAs occurred at a median of 12 h after the index RRT review, and primary
IHCA, in–hospital cardiac arrest; RRT, rapid response team; OR, Odds ratio; CI, Confidence interval; CCI, Charlson comorbidity index; COPD, Chronic obstructive pulmonary disease; Ref, Reference. a Unable to calculate (perfect match). b Malignant solid tumor or hematologic malignancy. c Lives alone/with spouse, daily visits by home nurse/relatives to assist in major home duties. d Lives in nursing home, requires almost continuous assistanc. e Data too sparse, unable to conduct conditional logistic regression.
Table 1 The 17 IHCAs within 48 h of the RRT review. Hours between RRT review and IHCA (median, Q1, Q3) Witnessed arrest Primary rhythm VT/VF PEA ASY Early termination of resuscitation a Hospital mortality Discharged alive to another hospital CPC at discharge (class) 30-day mortality
Controls %
rhythm was shockable in only one case. In six cases, advanced cardiac life support was terminated within 10 min because the attempt was deemed futile. A review of these six patients’ records did not suggest that the RRT should clearly have issued DNAR orders during the preceding reviews. Two patients survived to be discharged to another hospital, both with cerebral performance category one, and were alive 30 at days after the event. The 68 control patients were selected from a RRT sub population comparable to the cases (Fig. 1) as described in the methods-section. One medical patient under 50 years of age could not be matched for admission year because there were so few control cases with similar age, sex and ward background.
12 (5, 31) 16 1 9 7 6 15 2 1 (both) 15
Continuous variable is given as median (Q1, Q3). Other data are presented as counts, except the CPC which represents a neurological performance class. IHCA, in–hospital cardiac arrest; RRT, rapid response team; VT, ventricular tachycardia; VF, ventricular fibrillation; PEA, pulseless electrical activity; ASY, asystole; CPC, cerebral performance category. a Advanced life support is deemed futile and terminated within the first 10 min from the call.
Background of the cases vs. controls Table 2 compares the cases and controls according to 14 different background characteristics. There were no statistically significant 100
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
Table 3 Univariable analysis of the RRT review characteristics of the cases (n = 17, IHCA within 48 h after patient’s index RRT review on general ward) compared with controls (n = 68, no IHCA after index RRT review).
Days from admission to RRT review (median, Q1, Q3) Preceding ICU admission Surgery conducted within 24 h Review during on-call hours a Patient is in monitored area Previous laboratory markers (median, Q1, Q3) Hemoglobin (g/dl) Leucocytes (x 109/l) C–Reactive protein GFR (CKD-EPI) (ml/min/1.73 m2) Sodium (mmol/l) Potassium (mmol/l) Positive blood cultures Delayed RRT activation b Reason for RRT activation Respiratory Circulatory Neurologic Multiple Other RRT physician is only consulted by phone First NEWS recorded by the RRT (median, Q1, Q3) RRT diagnostics and interventions POC–ABG obtained Native radiography/Computed tomography I.v. fluids administered I.v. medications administered FiO2 level altered/CPAP issued Last NEWS recorded by the RRT (median, Q1, Q3) Difference first – last NEWS recorded by the RRT (median, Q1, Q3) Length of the RRT review (min) (median, Q1, Q3)
Cases %
Controls %
OR
95% CI
p–value
5 (1, 13) 47 5.9 71 41
2.5 (1, 8) 15 13 81 34
1.02 3.24 0.40 0.57 1.47
0.99–1.04 1.01–10.4 0.05–3.45 0.17–1.90 0.43–5.00
0.23 0.05 0.41 0.36 0.54
10.9 (9.6, 12.9) 9.8 (6.7, 12.7) 60 (19, 107) 66 (19, 93) 138 (135, 140) 4.1 (3.9, 4.4) 5.9 65
10.9 (9.4, 12.3) 10.1 (7.7, 12.9) 68 (19, 138) 63 (42, 88) 138 (135, 141) 3.8 (3.5, 4.3) 5.9 41
1.00 1.03 1.00 0.98 0.98 2.11 1.00 2.43
0.97–1.03 0.96–1.11 0.99–1.01 0.83–1.15 0.88–1.09 0.96–4.67 0.10–10.1 0.84–7.00
0.96 0.36 0.98 0.77 0.68 0.06 1.00 0.10
24 47 10 18 0.0 23 9 (5, 11.5)
46 28 12 4.4 12 48 6 (4, 9)
0.51 1.62 Ref. 4.11 – 0.33 1.16
0.08–3.12 0.28–9.38
0.47 0.59
0.45–37.6 – 0.10–1.13 0.99–1.37
53 35 35 35 41 7 (4.5, 10.5) 1 (0, 1.5) 31 (14.5, 42)
29 25 31 37 35 5.5 (4, 7) 0 (0, 2) 23 (17, 29)
2.74 1.64 1.21 0.93 1.30 1.25 0.96 1.02
0.91–8.28 0.53–5.13 0.41–3.60 0.28–3.10 0.43–3.92 1.03–1.52 0.81–1.30 0.99–1.06
0.21 c
0.08 0.07 0.07 0.39 0.74 0.90 0.65 0.03 0.81 0.24
RRT, rapid response team; IHCA, in–hospital cardiac arrest; OR, Odds ratio; CI, Confidence interval; ICU, intensive care unit; GFR, glomerular filtration rate; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; NEWS, national early warning score; Ref, Reference; POC–ABG, point-of-care arterial blood gases; I.v, intravenous; FiO2, fraction of inspired oxygen; CPAP, continuous positive airway pressure. a Other than Monday–Friday 8.00 a.m.–15.00 p.m. b Positive objective RRT activation criteria recorded 20–360 min before the RRT activation. c Unable to conduct conditional logistic regression (value in cases–cohort 0).
differences in comorbidities, cumulative comorbidity, preceding performance in daily living, admission characteristics or any other variable examined.
Table 4 Multivariable analysis of the factors associated with IHCA after index RRT review.
Preceding intensive care unit admission Charlson comorbidity index Last measured NEWS
RRT review characteristics of the cases vs. controls Table 3 presents the 24 variables examined that were related to the RRT reviews after which 17 cases suffered an IHCA and the 68 controls did not. There were no differences in preceding laboratory markers, preceding length of hospital stay or RRT trigger reasons. However, the cases were more likely to have had a preceding ICU admission than the controls (47% vs. 15%, respectively, OR 3.24, 95% CI 1.01–10.4). The review characteristics, first NEWS and conducted interventions were statistically comparable. NEWS decreased in both groups during the review (the NEWS was calculated for the first and last measured vital signs during the RRT review). However, the median NEWS at the end of the RRT review (the last NEWS) was higher among the cases than among the controls (7 vs. 5.5, respectively, OR 1.25, 95% CI 1.03–1.52).
OR
95% CI
p–value
2.95 0.94 1.22
0.83–10.4 0.72–1.23 1.00–1.49
0.09 0.65 0.048
NEWS, National early warning score; IHCA, in–hospital cardiac arrest; RRT, rapid response team; OR, Odds ratio; CI, Confidence interval.
IHCA within 48 h after the review.
Discussion Key findings After over five years of prospective data gathering and 2692 index RRT reviews in a Nordic tertiary referral centre, we found that IHCA among no-DNAR patients triaged to remain on ward is very rare within 48 h after a RRT review. The prognosis after such an adverse event is poor; indeed, one-third of these resuscitation attempts were quickly deemed futile and terminated. Thorough comparison with matched controls revealed that patients suffering an IHCA shortly after RRT review had more disturbed vital signs than the controls at the end of the review.
Factors independently associated with cases Three variables were introduced to the conditional multivariable logistic regression model: the two variables with statistical difference in the univariable analyses and the clinically relevant CCI representing the cumulative comorbidity (Table 4). Higher NEWS at the end of the RRT review was independently associated with the patients suffering an 101
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
review, although these associations were not statistically significant. It seems logical, that patients with hours of vital dysfunctions and more disturbed physiology at RRT arrival would be more likely to deteriorate to cardiac arrest. Based on our results, however, the most significant factor associated with IHCA after RRT review was having more deranged vital signs before RRT departure. The main goal of RRT is to intervene aggressively in this matter. It is debatable whether some of these 17 IHCA patients could have been salvaged with, for example, direct transfer to an ICU. In 13 cases, the RRT leader (an ICU physician) attended the patient with RRT nurses and a parent unit doctor; arterial blood gases were checked in nine patients, and NEWS score decreased by median one score during the reviews. This all suggests that patients were thoroughly checked. Moreover, the median last NEWS of the cases was seven; in our point-prevalence study in 2014 we found, that 6.5% of all general ward patients had NEWS ≥ 7 and all these patients cannot be directly admitted to the ICU for obvious reasons related to the ICU capacity [20]. Nevertheless, it is essential that RRTs pay extra attention to the last vital signs before they depart from the scene and ask themselves whether the patient is stable enough to remain on ward with the current interventions.
Incidence and prognosis of IHCA after RRT review The 17 IHCAs occurred during a study period that included over 370,000 hospital admissions, which suggests that these adverse events are rare. This observation is further supported by the fact that data collection for this study’s purposes was terminated early because only one event occurred during the final year of the study period. Some studies have reported the 24-h mortality rates after a RRT review, and these vary between 3 and 9% [12–14]. However, these mortality rates include patients with DNAR orders and those transferred to the ICU [12–14]. In a study from our own hospital, the 24-h mortality of RRT patients without limitations of medical treatment was just 2.4%; however, this cohort also included patients transferred to the ICU [15]. Therefore, only Schneider et al. provide data that can be compared with our results [6]. In their study of RRT triage performance in Melbourne, Australia, Schneider et al. found that only 6 of 3583 RRT patients (0.2%) triaged to remain on ward without a DNAR order had an unexpected IHCA within the following 24 h [6]. This ratio is substantially lower than the ratio found in our current study (12 out of 1610 RRT patients triaged to remain on ward without a DNAR order (0.7%), Fig. 1). However, in the study by Schneider et al., the RRT had been implemented for over five years when their study began, and their index call/1000 hospital admissions ratio was over 1.5 times higher during the study period, indicating that their rapid response system was more developed than the system at our institution [6]. If the general trend lies somewhere in between Schneider et al.’s results and our own, it seems safe to say that IHCA after RRT triages a patient to remain on ward without DNAR order is very unlikely. Fifteen of the 17 IHCA patients in our study died during their hospital admission, and, similarly, the 30-day mortality rate was 88%. The two surviving patients were discharged to another hospital, both with good neurological outcome. While the outcome percentages in our study should be interpreted with extreme caution, as the overall number of cases was very small, the prognosis of IHCA patients with preceding RRT review seemed even poorer than for IHCA patients in general [1]. One reason for the poor survival rate could be that these IHCAs were by definition preceded by vital dysfunctions, as the RRT had reviewed these patients, and the median NEWS scores were high both at team arrival and at departure. The recent findings of Andersen et al. support this deduction [3]: the more the vital signs are disturbed prior an IHCA, the poorer the prognosis. Six resuscitation attempts were quickly deemed futile and terminated. Were these resuscitation attempts inappropriate? Issuing DNAR orders is a complex process, especially during on-call hours, and there are several barriers to delayed DNAR order documentation [16]. However, the RRT did not consider issuing DNAR orders during their index reviews of these patients, nor did the patient records clearly suggest that DNAR orders should have been implemented. Further studies on IHCA resuscitation termination decisions are warranted; retrospectively, it is difficult to determine what information led to termination decisions among the six patients in our cohort. It is possible, however, that information from the patients’ index RRT reviews provided evidence of hours of (further) deterioration, which supported the decision to terminate these resuscitation attempts.
Limitations This study was conducted in a single centre in Finland; thus, our results must be generalised with caution. Further, our RRT’s trigger criteria and triggering frequency differ from those of other institutions. Data collection was terminated eight months early, as it was determined that waiting for more cases would not be worth the required resources, and the long study period was already inevitably affected by unmeasured secular trends. The incidence of the investigated event was very rare, so despite the suitability of the study design (a matched case–control study with four times the controls), [10] it is possible that some factors did not reach statistical significance due to the cohort size. However, the study had the power to detect an association between the clinically most relevant factor (and the only factor in which RRT can intervene during the review: patient’s physiological condition in the acute situation) and the undesired outcome. Conclusions In-hospital cardiac arrest occurring within 48 h after RRT decides that a patient remains on ward is a rare event with poor prognosis. Patients who suffer this adverse event have more disturbed vital signs at the end of their index RRT review than patients not suffering cardiac arrest after their RRT review. Conflict of interest statement This study was funded by the Competitive Research Funding of the Tampere University Hospital (Grant 9T009). The funding organization had no involvement in any aspect of the study. Acknowledgements None.
Possible risk factors for an IHCA after a RRT review
References
Previous studies have identified several risk factors for in-hospital and 30-day mortality among RRT patients through multivariable analyses, including age, comorbidities, non-elective hospital admission, medical background, delayed RRT activation, number of hospital days before RRT activation, RRT trigger reason and requiring fluid resuscitation during the review [14,17–19]. We analysed all these factors in a case–control setting, along with several other factors that could potentially affect the outcome. Indeed, delayed RRT activation and NEWS at RRT arrival seemed to be associated with IHCA after RRT
[1] Nolan J, Soar J, Smith GB, Gwinnutt C, Parrott F, Power S, et al. Incidence and outcome of in-hospital cardiac arrest in the United Kingdom National Cardiac Arrest Audit. Resuscitation 2014;85:987–92. [2] Nurmi J, Harjola VP, Nolan J, Castrén M. Observations and warning signs prior to cardiac arrest: should a medical emergency team intervene earlier. Acta Anaesthesiol Scand 2005;49:702–6. [3] Andersen LW, Kim WY, Chase M, Berg KM, Mortensen SJ, Moskowitz A, et al. The prevalence and significance of abnormal vital signs prior to in-hospital cardiac
102
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
arrest. Resuscitation 2016;98:112–7. [4] Jones D, DeVita M, Bellomo R. Rapid-response teams. N Engl J Med 2011;365:139–46. [5] Tirkkonen J, Tamminen T, Skrifvars MB. Outcome of adult patients attended by rapid response teams: a systematic review of the literature. Resuscitation 2017;112:43–52. [6] Schneider AG, Warrilow S, Robbins R, Jones DA, Bellomo R. An assessment of the triage performance of the efferent arm of the rapid response system. Resuscitation 2013;84:477–82. [7] Piscator E, Hedberg P, Göransson K, Djärv T. Survival after in-hospital cardiac arrest is highly associated with the age-combined Charlson Co-morbidity Index in a cohort study from a two-site Swedish University hospital. Resuscitation 2016;99:79–83. [8] Royal College of Physicians. National Early Warning Score (NEWS): Standardising the assessment of acute-illness severity in the NHS. Report of a working party. London: RCP; 2012. [9] Peberdy MA, Cretikos M, Abella BS, DeVita M, Goldhill D, Kloeck W, et al. Recommended guidelines for monitoring, reporting, and conducting research on medical emergency team, outreach, and rapid response systems: an Utstein-style scientific statement. Circulation 2007;116:2481–500. [10] Dos Santos Silva I. Cancer epidemiology: principles and methods. Lyon, France: International Agency for Research on Cancer; 1999. p. 189–212. [11] Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and cox regression. Am J Epidemiol 2006;165:701–18. [12] Dacey MJ, Mirza ER, Wilcox V, Doherty M, Mello J, Boyer A, et al. The effect of a
[13] [14]
[15] [16]
[17]
[18]
[19]
[20]
103
rapid response team on major clinical outcome measures in a community hospital. Crit Care Med 2007;35:2076–82. Herod R, Frost SA, Parr M, Hillman K, Aneman A. Long term trends in medical emergency team activations and outcomes. Resuscitation 2014;85:1083–7. Tirkkonen J, Ylä-Mattila J, Olkkola KT, Huhtala H, Tenhunen J, Hoppu S. Factors associated with delayed activation of medical emergency team and excess mortality: an Utstein-style analysis. Resuscitation 2013;84:173–8. Tirkkonen J, Olkkola KT, Tenhunen J, Hoppu S. Ethically justified treatment limitations in emergency situations. Eur J Emerg Med 2016;23:214–8. Micallef S, Skrifvars MB, Parr MJA. Level of agreement on resuscitation decisions among hospital specialists and barriers to documenting do not attempt resuscitation (DNAR) orders in ward patients. Resuscitation 2011;82:815–8. Boniatti MM, Azzolini N, Viana MV, Ribeiro BS, Coelho RS, Castilho RK, et al. Delayed medical emergency team calls and associated outcomes. Crit Care Med 2014;42:26–30. Calzavacca P, Licari E, Tee A, Egi M, Haase M, Haase-Fielitz A, et al. A prospective study of factors influencing the outcome of patients after a Medical Emergency Team review. Intensive Care Med 2008;34:2112–6. Calzavacca P, Licari E, Tee A, Mercer I, Haase M, Haase-Fielitz A, et al. Features and outcome of patients receiving multiple medical emergency team reviews. Resuscitation 2010;81:1509–15. Tirkkonen J, Olkkola KT, Huhtala H, Tenhunen J, Hoppu S. Medical emergency team activation: performance of conventional dichotomised criteria versus national early warning score. Acta Anaesthesiol Scand 2014;58:411–9.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
In-hospital cardiac arrest after a rapid response team review: A matched case–control study☆
T
⁎
Joonas Tirkkonena, , Heini Huhtalab, Sanna Hoppuc a
Department of Intensive Care Medicine, Tampere University Hospital and Department of Anaesthesiology and Intensive Care Medicine, Seinäjoki Central Hospital, PO Box 2000, FI-33521 Tampere, Finland b Faculty of Social Sciences, University of Tampere, FI-33014, Finland c Department of Intensive Care Medicine, Tampere University Hospital, PO Box 2000, FI-33521 Tampere, Finland
A R T I C LE I N FO
A B S T R A C T
Keywords: In-hospital cardiac arrest Rapid response team National early warning score
Aim: Study the incidence and reasons behind in-hospital cardiac arrests (IHCAs) after rapid response team (RRT) reviews. Methods: We conducted a matched case–control study at Tampere University Hospital, Finland. Data on adult patients who were triaged to remain on general ward after first (index) RRT review without treatment limitations but who suffered an IHCA within the following 48 h were prospectively collected for 5.3 years. These cases were matched (age ± 3 years, sex, surgical/medical ward, admission year) at a 1:4 ratio to controls (no ICHA after RRT review). Results: Of 2653 index RRT reviews, 17 patients suffered an IHCA on general ward within the 48 h after review. Their 30-day mortality rate was 88%. The incidence was 6.3/1000 index RRT reviews or 4.6/100,000 hospital admissions. Patients who suffered an IHCA within 48 h after RRT review were more likely to have a preceding ICU admission, and their median national early warning scores (NEWSs) at the end of the index RRT reviews (=last NEWSs) were higher than those of the controls. Higher last NEWS was the only factor associated with ICHA after RRT review (OR 1.22, 95% CI 1.00–1.49, p = 0.048) in a conditional multivariable regression model. Conclusions: IHCA within 48 h after an index RRT review on general ward is a rare event with poor prognosis. It is independently associated with higher NEWS at the end of the index RRT review. Careful consideration is stressed, when patients with high NEWS are left on ward after RRT reviews.
Introduction The incidence of adult in-hospital cardiac arrest (IHCA) is approximately 1.6 cases per 1000 hospital admissions [1]. Most IHCAs are preceded by derangements in vital signs, and the probability of survival to hospital discharge decreases as the number and severity of vital dysfunctions before the event increase [2,3]. If vital dysfunctions occur, rapid response teams (RRTs) review general ward patients to prevent IHCAs via minor interventions and timely transfers to intensive care [4]. Most RRT patients are triaged to remain on general ward without do not attempt resuscitation (DNAR) −orders [5,6]. One study that investigated the short-term outcomes among this RRT population found that 0.4% of no-DNAR patients triaged to remain on ward died during the following 24 h [6]. Is IHCA after an RRT review truly such a rare event, and is there something that the RRT misses during these reviews? In this matched
case–control study, we aimed 1) to report the incidence and prognosis of an IHCA occurring within 48 h after RRT triages a patient to remain on ward without a DNAR order and 2) to investigate thoroughly whether these cases (IHCA after RRT review) differed from controls (no IHCA after RRT review) at the time of their index RRT review. Methods Ethics The Ethics Committee of the Tampere University Hospital (Tays) approved the study protocol (Approval no: R10111, 2010). Patient consent was waived, as this was an observational study.
☆ ⁎
A Spanish translated version of the abstract of this article appears as Appendix in the final online version at https://doi.org/10.1016/j.resuscitation.2018.03.007. Corresponding author. E-mail addresses: [email protected].fi (J. Tirkkonen), Heini.Huhtala@staff.uta.fi (H. Huhtala), sanna.hoppu@pshp.fi (S. Hoppu).
https://doi.org/10.1016/j.resuscitation.2018.03.007 Received 10 October 2017; Received in revised form 4 February 2018; Accepted 5 March 2018 0300-9572/ © 2018 Elsevier B.V. All rights reserved.
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
Fig. 1. Study cohort. RRT, rapid response team; IHCA, in-hospital cardiac arrest; DNAR, do not attempt resuscitation; ICU, intensive care unit; OR, operation room; CT, computed tomography. a One medical patient under 50 years of age could not be matched for admission year (only six patients with similar age, sex and ward among the cohort of 1579 potential control patients).
(heart rate, systolic blood pressure, respiratory rate, peripheral blood oxygen saturation, body temperature, level of consciousness and the need for supplementary oxygen) is scored from 0 to 3 depending on how abnormal the measured value is. The total score is added up, and a score of ≥7 is considered a status that requires immediate RRT review [8].
Hospital Tays is one of five tertiary referral centres in Finland, with approximately 70,000 somatic admissions annually. The hospital provides the most advanced care for a catchment population of 1.1 million citizens. The intensive care unit (ICU) currently has 21 beds, with approximately 2100 admissions per year. There is a separate ICU with seven beds for surgical cardiothoracic patients.
Exclusion criteria RRT activations for paediatric patients (< 18 years) and index RRT reviews including an IHCA were excluded. RRT reviews for operating rooms, ICUs and the emergency room were further excluded, as were RRT reviews for outpatients. Finally, RRT reviews of patients with preexisting or new DNAR orders and RRT reviews resulting in ICU/operating room transfer were excluded.
Rapid response team The hospital’s rapid response system was implemented in January 2009. The ICU-operated RRT is physician led and the general ward staff utilise dichotomised RRT trigger criteria (heart rate < 40/min or > 140/min, systolic blood pressure < 90 mmHg, peripheral arteriolar oxygen saturation < 90%, respiratory rate < 5/min or > 24/min, decrease in state of consciousness and the ‘worried’ criterion). The RRT attends both medical emergencies and cardiac arrests 24/7. Although the RRT is physician-led, the RRT nurses may attend a patient by themselves (in collaboration with the parent unit physician) and only consult with the RRT physician by phone.
Data collection Prospective data collection began in 1 January 2010 as a part of an ongoing project involving RRT data gathering per the Utstein Style [9]. No previous study had documented the incidence of IHCAs review, so data for precise sample size calculations were not available. We aimed to collect data within a six-year period, as a longer study period would inevitably have affected the internal validity of the results. However, data collection was terminated eight months ahead of time, on 31 April 2015. Because only one case had occurred in the previous 12 months, our research resources were directed to another study. The initial data collection consisted of Utstein-style RRT templates [9], which were completed by the RRT nurses at scene. The RRT nurses recorded patient’s name and social security number, event location, RRT review reason with a short verbal comment on the preceding events, vital signs (which were measured multiple times during the reviews), all conducted examinations and interventions, possible
Definitions ‘Index RRT review’ refers to a patient’s first RRT review during his/ her hospitalisation. ‘Delayed RRT activation’ refers to a RRT activation that is preceded by recorded objective RRT activation criteria 20–360 min before the RRT review is triggered. Patients’ cumulative comorbidity was assessed with the Charlson comorbidity index (CCI) [7]. National early warning score (NEWS) is an early warning scoring system that has been implemented in all public hospitals in the United Kingdom [8]. NEWS provides a standardized method for the assessment and response to acute illness [8]. Each of the seven included vital signs 99
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
treatment limitations and follow-up plan/possible transfer to higher level of care. Further patient data and outcome data were retrieved from the electronic patient records. Controls were selected at 1:4 ratio (after being matched to cases by age ± 3 years, sex, surgical/medical ward and admission year), as is appropriate when there is no shortage of controls [10]. Controls were obtained from a sub population of RRT patients (without following IHCA) who did not have a DNAR order, did not receive a DNAR order, were not transferred to the ICU/operation room and did not die during the review (Fig. 1). When more than four matching controls were available, the first four controls were included in chronological order (according to the date of the RRT review).
Table 2 Univariable analysis of the characteristics of the cases (IHCA within 48 h after patient’s index RRT review on general ward, n = 17) compared with controls (no IHCA after index RRT review, n = 68). Cases %
OR
95% CI
p–value
76 (65, 81)
1.08
0.80–1.47
0.61
59 47 77 2.0 (1.0, 3.0) 37 13 8.8 8.8 15 28 25 15 4.4
– – 1.40 0.90
– – 0.37–5.26 0.69–1.17
–a –a 0.62 0.42
0.93 1.83 1.36 0.65 1.24 0.33 0.32 2.03 1.33
0.28–3.12 0.54–6.14 0.28–7.20 0.07–5.72 0.30–5.17 0.07–1.65 0.08–1.92 0.46–8.90 0.14–12.8
0.90 0.33 0.72 0.70 0.73 0.18 0.25 0.35 0.80
82 10
Ref. 2.74
0.65–11.5
0.17
5.9
7.4
1.17
0.10–13.3
0.90
65 12 24
84 4.4 12
Ref. 3.21 2.46
0.52–20.0 0.65–9.30
0.21 0.18
24 12 35 5.9
18 18 15 5.9
0.0
4.4
–
–
e
0.0 18 0.0 0.0 5.9
1.5 22 10 5.9 0.0
Age (median, Q1, Q3)
77 (63, 82) Sex (male) 59 Medical patient 47 Acute hospital admission 82 CCI (median, Q1, Q3) 1.0 (1.0, 2.0) Coronary artery disease 35 Congestive heart failure 24 Peripheral artery disease 12 Cerebrovascular disease 5.9 COPD 18 Diabetes 12 Malignancy b 12 Chronic renal failure 24 Dementia 5.9 Preceding performance in daily living Independent 71 Requires some assistance 24
Statistical analysis Data are presented as percentages, and continuous variables as medians (Q1, Q3). Conditional univariable logistic regression was used for statistical comparisons between cases and controls. Conditional multivariable logistic regression was used to investigate independent associations with cases. As is generally considered appropriate, the introduction of one variable into the multivariable model required at least five cases (outcomes) [11]. Tests were two-sided; p < 0.05 was considered significant and 95% confidence intervals are reported. Stata® version 13.0 (StataCorp LLP, Texas, USA) was used for the analyses, all of which were conducted in collaboration with a biostatistician.
c
Severely limited d Admitted to hospital from Home Nursing home Other hospital/Primary care ward Hospital admission diagnosis Infection Malignancy Cardiovascular Cerebrovascular (nonsurgical) Neurosurgical (other than tumors) Respiratory Gastrosurgical Trauma Other medical Other surgical
Results Cohort description During the 5.3-year study period, 17 IHCAs occurred within 48 h after RRT triaged these no-DNAR patients to remain on ward after their index RRT review (Fig. 1). Twelve (71%) of these events occurred within the first 24 h. During the study period, the incidence of IHCA within 48 h after index RRT review was 6.3 per 1000 index RRT reviews or 4.6 per 100,000 hospital admissions, and these IHCAs accounted for 3.6% of all IHCA patients. Table 1 presents the data related to the 17 cardiac arrest events, while Tables 2 and 3 present background data related to patient characteristics and the index RRT reviews of these cases. Two-thirds of the case patients had previously lived at home with normal performance in daily living. The median age of the patients was 77 years; 59% were male, and 47% had a medical reason for admission at the time of the IHCA. The cumulative comorbidity among the 17 IHCA patients was low, with a median Charlson comorbidity index score of one. The IHCAs occurred at a median of 12 h after the index RRT review, and primary
IHCA, in–hospital cardiac arrest; RRT, rapid response team; OR, Odds ratio; CI, Confidence interval; CCI, Charlson comorbidity index; COPD, Chronic obstructive pulmonary disease; Ref, Reference. a Unable to calculate (perfect match). b Malignant solid tumor or hematologic malignancy. c Lives alone/with spouse, daily visits by home nurse/relatives to assist in major home duties. d Lives in nursing home, requires almost continuous assistanc. e Data too sparse, unable to conduct conditional logistic regression.
Table 1 The 17 IHCAs within 48 h of the RRT review. Hours between RRT review and IHCA (median, Q1, Q3) Witnessed arrest Primary rhythm VT/VF PEA ASY Early termination of resuscitation a Hospital mortality Discharged alive to another hospital CPC at discharge (class) 30-day mortality
Controls %
rhythm was shockable in only one case. In six cases, advanced cardiac life support was terminated within 10 min because the attempt was deemed futile. A review of these six patients’ records did not suggest that the RRT should clearly have issued DNAR orders during the preceding reviews. Two patients survived to be discharged to another hospital, both with cerebral performance category one, and were alive 30 at days after the event. The 68 control patients were selected from a RRT sub population comparable to the cases (Fig. 1) as described in the methods-section. One medical patient under 50 years of age could not be matched for admission year because there were so few control cases with similar age, sex and ward background.
12 (5, 31) 16 1 9 7 6 15 2 1 (both) 15
Continuous variable is given as median (Q1, Q3). Other data are presented as counts, except the CPC which represents a neurological performance class. IHCA, in–hospital cardiac arrest; RRT, rapid response team; VT, ventricular tachycardia; VF, ventricular fibrillation; PEA, pulseless electrical activity; ASY, asystole; CPC, cerebral performance category. a Advanced life support is deemed futile and terminated within the first 10 min from the call.
Background of the cases vs. controls Table 2 compares the cases and controls according to 14 different background characteristics. There were no statistically significant 100
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
Table 3 Univariable analysis of the RRT review characteristics of the cases (n = 17, IHCA within 48 h after patient’s index RRT review on general ward) compared with controls (n = 68, no IHCA after index RRT review).
Days from admission to RRT review (median, Q1, Q3) Preceding ICU admission Surgery conducted within 24 h Review during on-call hours a Patient is in monitored area Previous laboratory markers (median, Q1, Q3) Hemoglobin (g/dl) Leucocytes (x 109/l) C–Reactive protein GFR (CKD-EPI) (ml/min/1.73 m2) Sodium (mmol/l) Potassium (mmol/l) Positive blood cultures Delayed RRT activation b Reason for RRT activation Respiratory Circulatory Neurologic Multiple Other RRT physician is only consulted by phone First NEWS recorded by the RRT (median, Q1, Q3) RRT diagnostics and interventions POC–ABG obtained Native radiography/Computed tomography I.v. fluids administered I.v. medications administered FiO2 level altered/CPAP issued Last NEWS recorded by the RRT (median, Q1, Q3) Difference first – last NEWS recorded by the RRT (median, Q1, Q3) Length of the RRT review (min) (median, Q1, Q3)
Cases %
Controls %
OR
95% CI
p–value
5 (1, 13) 47 5.9 71 41
2.5 (1, 8) 15 13 81 34
1.02 3.24 0.40 0.57 1.47
0.99–1.04 1.01–10.4 0.05–3.45 0.17–1.90 0.43–5.00
0.23 0.05 0.41 0.36 0.54
10.9 (9.6, 12.9) 9.8 (6.7, 12.7) 60 (19, 107) 66 (19, 93) 138 (135, 140) 4.1 (3.9, 4.4) 5.9 65
10.9 (9.4, 12.3) 10.1 (7.7, 12.9) 68 (19, 138) 63 (42, 88) 138 (135, 141) 3.8 (3.5, 4.3) 5.9 41
1.00 1.03 1.00 0.98 0.98 2.11 1.00 2.43
0.97–1.03 0.96–1.11 0.99–1.01 0.83–1.15 0.88–1.09 0.96–4.67 0.10–10.1 0.84–7.00
0.96 0.36 0.98 0.77 0.68 0.06 1.00 0.10
24 47 10 18 0.0 23 9 (5, 11.5)
46 28 12 4.4 12 48 6 (4, 9)
0.51 1.62 Ref. 4.11 – 0.33 1.16
0.08–3.12 0.28–9.38
0.47 0.59
0.45–37.6 – 0.10–1.13 0.99–1.37
53 35 35 35 41 7 (4.5, 10.5) 1 (0, 1.5) 31 (14.5, 42)
29 25 31 37 35 5.5 (4, 7) 0 (0, 2) 23 (17, 29)
2.74 1.64 1.21 0.93 1.30 1.25 0.96 1.02
0.91–8.28 0.53–5.13 0.41–3.60 0.28–3.10 0.43–3.92 1.03–1.52 0.81–1.30 0.99–1.06
0.21 c
0.08 0.07 0.07 0.39 0.74 0.90 0.65 0.03 0.81 0.24
RRT, rapid response team; IHCA, in–hospital cardiac arrest; OR, Odds ratio; CI, Confidence interval; ICU, intensive care unit; GFR, glomerular filtration rate; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; NEWS, national early warning score; Ref, Reference; POC–ABG, point-of-care arterial blood gases; I.v, intravenous; FiO2, fraction of inspired oxygen; CPAP, continuous positive airway pressure. a Other than Monday–Friday 8.00 a.m.–15.00 p.m. b Positive objective RRT activation criteria recorded 20–360 min before the RRT activation. c Unable to conduct conditional logistic regression (value in cases–cohort 0).
differences in comorbidities, cumulative comorbidity, preceding performance in daily living, admission characteristics or any other variable examined.
Table 4 Multivariable analysis of the factors associated with IHCA after index RRT review.
Preceding intensive care unit admission Charlson comorbidity index Last measured NEWS
RRT review characteristics of the cases vs. controls Table 3 presents the 24 variables examined that were related to the RRT reviews after which 17 cases suffered an IHCA and the 68 controls did not. There were no differences in preceding laboratory markers, preceding length of hospital stay or RRT trigger reasons. However, the cases were more likely to have had a preceding ICU admission than the controls (47% vs. 15%, respectively, OR 3.24, 95% CI 1.01–10.4). The review characteristics, first NEWS and conducted interventions were statistically comparable. NEWS decreased in both groups during the review (the NEWS was calculated for the first and last measured vital signs during the RRT review). However, the median NEWS at the end of the RRT review (the last NEWS) was higher among the cases than among the controls (7 vs. 5.5, respectively, OR 1.25, 95% CI 1.03–1.52).
OR
95% CI
p–value
2.95 0.94 1.22
0.83–10.4 0.72–1.23 1.00–1.49
0.09 0.65 0.048
NEWS, National early warning score; IHCA, in–hospital cardiac arrest; RRT, rapid response team; OR, Odds ratio; CI, Confidence interval.
IHCA within 48 h after the review.
Discussion Key findings After over five years of prospective data gathering and 2692 index RRT reviews in a Nordic tertiary referral centre, we found that IHCA among no-DNAR patients triaged to remain on ward is very rare within 48 h after a RRT review. The prognosis after such an adverse event is poor; indeed, one-third of these resuscitation attempts were quickly deemed futile and terminated. Thorough comparison with matched controls revealed that patients suffering an IHCA shortly after RRT review had more disturbed vital signs than the controls at the end of the review.
Factors independently associated with cases Three variables were introduced to the conditional multivariable logistic regression model: the two variables with statistical difference in the univariable analyses and the clinically relevant CCI representing the cumulative comorbidity (Table 4). Higher NEWS at the end of the RRT review was independently associated with the patients suffering an 101
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
review, although these associations were not statistically significant. It seems logical, that patients with hours of vital dysfunctions and more disturbed physiology at RRT arrival would be more likely to deteriorate to cardiac arrest. Based on our results, however, the most significant factor associated with IHCA after RRT review was having more deranged vital signs before RRT departure. The main goal of RRT is to intervene aggressively in this matter. It is debatable whether some of these 17 IHCA patients could have been salvaged with, for example, direct transfer to an ICU. In 13 cases, the RRT leader (an ICU physician) attended the patient with RRT nurses and a parent unit doctor; arterial blood gases were checked in nine patients, and NEWS score decreased by median one score during the reviews. This all suggests that patients were thoroughly checked. Moreover, the median last NEWS of the cases was seven; in our point-prevalence study in 2014 we found, that 6.5% of all general ward patients had NEWS ≥ 7 and all these patients cannot be directly admitted to the ICU for obvious reasons related to the ICU capacity [20]. Nevertheless, it is essential that RRTs pay extra attention to the last vital signs before they depart from the scene and ask themselves whether the patient is stable enough to remain on ward with the current interventions.
Incidence and prognosis of IHCA after RRT review The 17 IHCAs occurred during a study period that included over 370,000 hospital admissions, which suggests that these adverse events are rare. This observation is further supported by the fact that data collection for this study’s purposes was terminated early because only one event occurred during the final year of the study period. Some studies have reported the 24-h mortality rates after a RRT review, and these vary between 3 and 9% [12–14]. However, these mortality rates include patients with DNAR orders and those transferred to the ICU [12–14]. In a study from our own hospital, the 24-h mortality of RRT patients without limitations of medical treatment was just 2.4%; however, this cohort also included patients transferred to the ICU [15]. Therefore, only Schneider et al. provide data that can be compared with our results [6]. In their study of RRT triage performance in Melbourne, Australia, Schneider et al. found that only 6 of 3583 RRT patients (0.2%) triaged to remain on ward without a DNAR order had an unexpected IHCA within the following 24 h [6]. This ratio is substantially lower than the ratio found in our current study (12 out of 1610 RRT patients triaged to remain on ward without a DNAR order (0.7%), Fig. 1). However, in the study by Schneider et al., the RRT had been implemented for over five years when their study began, and their index call/1000 hospital admissions ratio was over 1.5 times higher during the study period, indicating that their rapid response system was more developed than the system at our institution [6]. If the general trend lies somewhere in between Schneider et al.’s results and our own, it seems safe to say that IHCA after RRT triages a patient to remain on ward without DNAR order is very unlikely. Fifteen of the 17 IHCA patients in our study died during their hospital admission, and, similarly, the 30-day mortality rate was 88%. The two surviving patients were discharged to another hospital, both with good neurological outcome. While the outcome percentages in our study should be interpreted with extreme caution, as the overall number of cases was very small, the prognosis of IHCA patients with preceding RRT review seemed even poorer than for IHCA patients in general [1]. One reason for the poor survival rate could be that these IHCAs were by definition preceded by vital dysfunctions, as the RRT had reviewed these patients, and the median NEWS scores were high both at team arrival and at departure. The recent findings of Andersen et al. support this deduction [3]: the more the vital signs are disturbed prior an IHCA, the poorer the prognosis. Six resuscitation attempts were quickly deemed futile and terminated. Were these resuscitation attempts inappropriate? Issuing DNAR orders is a complex process, especially during on-call hours, and there are several barriers to delayed DNAR order documentation [16]. However, the RRT did not consider issuing DNAR orders during their index reviews of these patients, nor did the patient records clearly suggest that DNAR orders should have been implemented. Further studies on IHCA resuscitation termination decisions are warranted; retrospectively, it is difficult to determine what information led to termination decisions among the six patients in our cohort. It is possible, however, that information from the patients’ index RRT reviews provided evidence of hours of (further) deterioration, which supported the decision to terminate these resuscitation attempts.
Limitations This study was conducted in a single centre in Finland; thus, our results must be generalised with caution. Further, our RRT’s trigger criteria and triggering frequency differ from those of other institutions. Data collection was terminated eight months early, as it was determined that waiting for more cases would not be worth the required resources, and the long study period was already inevitably affected by unmeasured secular trends. The incidence of the investigated event was very rare, so despite the suitability of the study design (a matched case–control study with four times the controls), [10] it is possible that some factors did not reach statistical significance due to the cohort size. However, the study had the power to detect an association between the clinically most relevant factor (and the only factor in which RRT can intervene during the review: patient’s physiological condition in the acute situation) and the undesired outcome. Conclusions In-hospital cardiac arrest occurring within 48 h after RRT decides that a patient remains on ward is a rare event with poor prognosis. Patients who suffer this adverse event have more disturbed vital signs at the end of their index RRT review than patients not suffering cardiac arrest after their RRT review. Conflict of interest statement This study was funded by the Competitive Research Funding of the Tampere University Hospital (Grant 9T009). The funding organization had no involvement in any aspect of the study. Acknowledgements None.
Possible risk factors for an IHCA after a RRT review
References
Previous studies have identified several risk factors for in-hospital and 30-day mortality among RRT patients through multivariable analyses, including age, comorbidities, non-elective hospital admission, medical background, delayed RRT activation, number of hospital days before RRT activation, RRT trigger reason and requiring fluid resuscitation during the review [14,17–19]. We analysed all these factors in a case–control setting, along with several other factors that could potentially affect the outcome. Indeed, delayed RRT activation and NEWS at RRT arrival seemed to be associated with IHCA after RRT
[1] Nolan J, Soar J, Smith GB, Gwinnutt C, Parrott F, Power S, et al. Incidence and outcome of in-hospital cardiac arrest in the United Kingdom National Cardiac Arrest Audit. Resuscitation 2014;85:987–92. [2] Nurmi J, Harjola VP, Nolan J, Castrén M. Observations and warning signs prior to cardiac arrest: should a medical emergency team intervene earlier. Acta Anaesthesiol Scand 2005;49:702–6. [3] Andersen LW, Kim WY, Chase M, Berg KM, Mortensen SJ, Moskowitz A, et al. The prevalence and significance of abnormal vital signs prior to in-hospital cardiac
102
Resuscitation 126 (2018) 98–103
J. Tirkkonen et al.
arrest. Resuscitation 2016;98:112–7. [4] Jones D, DeVita M, Bellomo R. Rapid-response teams. N Engl J Med 2011;365:139–46. [5] Tirkkonen J, Tamminen T, Skrifvars MB. Outcome of adult patients attended by rapid response teams: a systematic review of the literature. Resuscitation 2017;112:43–52. [6] Schneider AG, Warrilow S, Robbins R, Jones DA, Bellomo R. An assessment of the triage performance of the efferent arm of the rapid response system. Resuscitation 2013;84:477–82. [7] Piscator E, Hedberg P, Göransson K, Djärv T. Survival after in-hospital cardiac arrest is highly associated with the age-combined Charlson Co-morbidity Index in a cohort study from a two-site Swedish University hospital. Resuscitation 2016;99:79–83. [8] Royal College of Physicians. National Early Warning Score (NEWS): Standardising the assessment of acute-illness severity in the NHS. Report of a working party. London: RCP; 2012. [9] Peberdy MA, Cretikos M, Abella BS, DeVita M, Goldhill D, Kloeck W, et al. Recommended guidelines for monitoring, reporting, and conducting research on medical emergency team, outreach, and rapid response systems: an Utstein-style scientific statement. Circulation 2007;116:2481–500. [10] Dos Santos Silva I. Cancer epidemiology: principles and methods. Lyon, France: International Agency for Research on Cancer; 1999. p. 189–212. [11] Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and cox regression. Am J Epidemiol 2006;165:701–18. [12] Dacey MJ, Mirza ER, Wilcox V, Doherty M, Mello J, Boyer A, et al. The effect of a
[13] [14]
[15] [16]
[17]
[18]
[19]
[20]
103
rapid response team on major clinical outcome measures in a community hospital. Crit Care Med 2007;35:2076–82. Herod R, Frost SA, Parr M, Hillman K, Aneman A. Long term trends in medical emergency team activations and outcomes. Resuscitation 2014;85:1083–7. Tirkkonen J, Ylä-Mattila J, Olkkola KT, Huhtala H, Tenhunen J, Hoppu S. Factors associated with delayed activation of medical emergency team and excess mortality: an Utstein-style analysis. Resuscitation 2013;84:173–8. Tirkkonen J, Olkkola KT, Tenhunen J, Hoppu S. Ethically justified treatment limitations in emergency situations. Eur J Emerg Med 2016;23:214–8. Micallef S, Skrifvars MB, Parr MJA. Level of agreement on resuscitation decisions among hospital specialists and barriers to documenting do not attempt resuscitation (DNAR) orders in ward patients. Resuscitation 2011;82:815–8. Boniatti MM, Azzolini N, Viana MV, Ribeiro BS, Coelho RS, Castilho RK, et al. Delayed medical emergency team calls and associated outcomes. Crit Care Med 2014;42:26–30. Calzavacca P, Licari E, Tee A, Egi M, Haase M, Haase-Fielitz A, et al. A prospective study of factors influencing the outcome of patients after a Medical Emergency Team review. Intensive Care Med 2008;34:2112–6. Calzavacca P, Licari E, Tee A, Mercer I, Haase M, Haase-Fielitz A, et al. Features and outcome of patients receiving multiple medical emergency team reviews. Resuscitation 2010;81:1509–15. Tirkkonen J, Olkkola KT, Huhtala H, Tenhunen J, Hoppu S. Medical emergency team activation: performance of conventional dichotomised criteria versus national early warning score. Acta Anaesthesiol Scand 2014;58:411–9.