Incidences and outcomes of intra-operative vs. postoperative paediatric cardiac arrest

Abstract

BACKGROUND

The reported incidence of paediatric perioperative cardiac arrest (PPOCA) in most developing countries ranges from 2.7 to 22.9 per 10 000 anaesthetics, resulting in mortality rates of 2.0 to 10.7 per 10 000 anaesthetics. The definitions of ‘peri-operative’ cardiac arrest often include the intra-operative period and extends from 60 min to 48 h after anaesthesia completion. However, the characteristics of cardiac arrests, care settings, and resuscitation quality may differ between intra-operative and early postoperative cardiac arrests.

OBJECTIVE

To compare the mortality rates between intraoperative and early postoperative cardiac arrests (<24 h) following anaesthesia for paediatric noncardiac surgery.

DESIGN

A retrospective cohort study.

SETTING

In a tertiary care centre in Thailand during 2014 to 2019, the peri-operative period was defined as from the beginning of anaesthesia care until 24 h after anaesthesia completion.

PATIENTS

Paediatric patients aged 0 to 17 years who underwent anaesthesia for noncardiac surgery.

MAIN OUTCOME MEASURES

Mortality rates.

RESULTS

A total of 42 776 anaesthetics were identified, with 63 PPOCAs and 23 deaths (36.5%). The incidence (95% confidence interval) of PPOCAs and mortality were 14.7 (11.5 to 18.8) and 5.4 (3.6 to 8.1) per 10 000 anaesthetics, respectively. Among 63 PPOCAs, 41 (65%) and 22 (35%) occurred during the intra-operative and postoperative periods, respectively. The median [min to max] time of postoperative cardiac arrest was 3.84 [0.05 to 19.47] h after anaesthesia completion. Mortalities (mortality rate) of postoperative cardiac arrest were significantly higher than that of intra-operative cardiac arrest at 14 (63.6%) vs. 9 (22.0%, P = 0.001). Multivariate analysis of risk factors for mortality included emergency status and duration of cardiopulmonary resuscitation with adjusted odds ratio 5.388 (95% confidence interval (1.031 to 28.160) and 1.067 (1.016 to 1.120).

CONCLUSIONS

Postoperative cardiac arrest resulted in a higher mortality rate than intra-operative cardiac arrest. A high level of care should be provided for at least 24 h after the completion of anaesthesia.

TRIAL REGISTRATION

None.

CLINICAL TRIAL NUMBER AND REGISTRY URL

NA.

KEY POINTS

• Postoperative cardiac arrest resulted in significantly longer cardiopulmonary resuscitation duration and higher mortality rates than intra-operative cardiac arrest.

• Fifty-nine percent of postoperative cardiac arrests were airway and respiration-related. Cardiac arrest is often the last event after progressive deterioration, and may be preventable in some cases by a higher level of care.

• A high level of care should be extended up to 24 h after anaesthesia completion.

Introduction

Peri-operative cardiac arrest is a rare, critical event that results in significant morbidity and mortality. Incidence, causes, and outcomes of paediatric peri-operative cardiac arrest (PPOCA) vary significantly among facilities in different settings. The reported incidence of PPOCA in developed countries ranges from 2.9 to 13.4 per 10 000 anaesthetics.^1–7 The previously reported incidence in developing countries differed between 2.7 and 22.9 per 10 000 anesthetics^8–13 and could be as high as 156 per 10 000 anaesthetics.¹⁴ Mortality rates also varied greatly from 0.29 to 13.1 per 10 000 anaesthetics in developed countries^3–6,15 to 2.0 to 97.0 per 10 000 anaesthetics in developing countries.^10–14 Highest risk factors include: infancy, American Society of Anesthesiologists physical status (ASA PS) III–V, emergency status, and out-of-hours. These have all been reported as risk factors for mortality.^1,11,15,16

The time-based definition of PPOCA differs among studies, ranging from cardiac arrests occurring in the postanaesthesia care unit (PACU)^{3,5,10,12,16–19} or 60 min after anesthesia^2,6 until 24^1,7,8,15 to 48 h after completion of anaesthesia.¹³ However, settings for patient care and monitoring during anaesthesia and in the postoperative period such as the PACU, intensive care unit (ICU), and ward, are different in different locations. For example, in the operating room, continuous monitoring and the presence of anaesthesia providers at all times are considered the standard of care. In contrast, intense monitoring might not be feasible in ward settings. Even in the ICU or PACU, the nurse-to-patient ratios and the intensity of care might not match the operating room settings.

We hypothesised that differences in cardiac arrest characteristics and resuscitation care during PPOCA in different locations might affect mortality and patient outcomes. Therefore, we conducted this observational study to compare the incidence and mortality of PPOCA occurring intra-operatively and postoperatively. The primary objective was to compare the mortality rates between intra-operative and postoperative cardiac arrests. We also explored the differences in the patient characteristics of PPOCA and outcomes between intra-operative and postoperative cardiac arrest.

Methods

This retrospective cohort study was conducted at the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand. Ethical approval for this study (certificate of approval number Si587/2019) was provided by the Siriraj Institutional Review Board, Bangkok, Thailand on 27 August 2019. We included paediatric patients aged 0 to 17 years who underwent anaesthesia for noncardiac surgery, including procedures performed remotely from the operating room that required anaesthesia services between January 1, 2014, and December 31, 2019, and experienced peri-operative cardiac arrest. The requirement for written informed consent was waived by the institutional review board owing to the retrospective nature of the study. This manuscript adheres to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Institutional characteristics and patient care

Siriraj Hospital is a tertiary care referral centre for adult and paediatric patients. An anaesthesiologist (board certified by the Royal College of Anaesthesiologists of Thailand [RCAT]) along with certified registered nurse anaesthetists or anaesthesia residents/fellows, delivered paediatric anaesthesia care at the institution. In the operating room, patients received continuous electrocardiography and pulse oximetry monitoring, with blood pressure measured at least every 5 min, as recommended by the American Society of Anesthesiologists. The anaesthesia provider-to-patient ratio was 1:1. In the PACU, patients received continuous pulse oximetry monitoring, although electrocardiography was optional. Blood pressure was measured every 15 min. The PACU nurse-to-patient ratio was 1:2. In the ICU, patients received continuous electrocardiography and pulse oximetry monitoring, similar to the operating room; however, blood pressure was measured every 1 h unless an arterial line was present. During the study period, the nursing staff-to-patient ratio was 1:1 or 1:2. In the ward, vital signs were measured every hour in the first four hours and then every four hours. The nursing staff-to-patient ratio ranged from 1:5 to 1:8. During the transfer, the patients received the standard of care as in the operating room; however, any transfer resuscitation may not have access to the same level of equipment. An anaesthesia team managed the resuscitation of cardiac arrest in the operating room, PACU, and during transfer. Resuscitation of cardiac arrest in the ICU and the ward was generally managed by a primary physician (surgeons, paediatricians, or paediatric intensivists).

Data collection

Cardiac arrest was defined as the cessation of cardiac mechanical activity or circulatory arrest as confirmed by the absence of signs of circulation that require chest compression or that resulted in death.^{1,3,5,6,16–19} For this study, we defined PPOCA as cardiac arrest occurring after patient arrival in the operating room until 24 h after completion of anaesthesia. Participants were identified from the hospital database using the International Classification of Diseases (ICD-9 and ICD-10). The ICD-9 codes were 3965 extracorporeal membrane oxygenation (ECMO), 9960 cardiopulmonary resuscitations, not otherwise specified, 9962 other electric countershock of heart, 9963 closed chest cardiac massage; and the ICD-10 codes were I46 cardiac arrest, I460 cardiac arrest with successful resuscitation, I461 sudden cardiac death, so described, I469 cardiac arrest unspecified. Participants who received intra-operative care for organ donation were excluded from the study. We further categorised PPOCA into intra-operative and postoperative cardiac arrest. Intra-operative cardiac arrest was defined as cardiac arrest occurring in the operating room (during induction, maintenance, or emergence). Postoperative cardiac arrest was defined as cardiac arrest occurring after completion of anaesthesia until 24 h postsurgery (during transfer, PACU, ICU, or regular ward).

The data were screened, rechecked, and extracted by one investigator (SS). Demographic data such as age, sex, American Society of Anaesthesiologists physical status (ASA PS) classification, emergency or urgent status, and department of service were included. Characteristics of each anaesthetic, such as diagnosis, surgical procedure, operative date and time, anaesthesia technique, anaesthesiologist's speciality, and experience, were also included. Regarding each cardiac arrest, timing, location, cardiopulmonary resuscitation (CPR) duration, ECMO initiation and cardiac arrest outcomes were collected. Two paediatric anaesthesiologists (SR and TK) independently reviewed each patient's medical record and identified the causes of cardiac arrests according to the Paediatric Peri-operative Cardiac Arrest Registry and the Wake Up Safe organisation.^1,16,17,20 Any disagreements between the two paediatric anaesthesiologists were resolved by discussion. After discussion, if a dispute remained, a third anaesthesiologist (NC) was consulted. The consensus was determined when at least two of the three members agreed on the causes of cardiac arrest. If there were multiple possible causes of cardiac arrest, the primary cause was identified as the main cause of cardiac arrest by an agreement between the two paediatric anaesthesiologists. Possible causes of cardiac arrest were divided into five categories: cardiovascular, medication, airway and respiratory, metabolic, and other (Table 1, Supplemental Digital Content).^1,16–18 Patient outcomes were assessed at the last clinical evaluation before hospital discharge for temporary harm, permanent harm, or death.

The causes of cardiac arrest were divided according to the Paediatric Peri-operative Cardiac Arrest Registry definition.^1,16–18 Cardiac arrests were categorised as anaesthesia-related cardiac arrest (ARCA) and non-anaesthesia-related cardiac arrest (NARCA). If the anaesthetic process or personnel could have played at least some role in the genesis of cardiac arrest, it was determined as ARCA. Meanwhile, if the anaesthetic process did not contribute to even a minor degree, the cardiac arrest was classified as NARCA. For instance, cardiac arrest due to haemorrhage was defined as ARCA if it had been prevented by appropriate fluid replacement and blood transfusion. Cardiac arrest without an identifiable cause was classified as ARCA. The severity of surgical procedures was categorised into low, moderate, or high levels (low = superficial or diagnostic procedure with minimal blood loss, diagnostic cardiac catheterisation, moderate = invasive, anticipated moderated blood loss, therapeutic cardiac catheterisation, airway procedure, high = major procedure with a significant effect on haemodynamics).¹⁹

Statistical analysis

Descriptive statistics were used to describe the demographic data and causes of cardiac arrest. Incidence of PPOCA, ARCA, NARCA, and in-hospital mortality were reported as relative frequencies per 10 000 anaesthetics with a 95% confidence interval (CI). The CI was calculated using the Wilson Score method. If the details of cardiac arrest were missing, the patient who experienced PPOCA was included in the incidence report but not in the final analysis. The characteristics of intra-operative and postoperative cardiac arrest were compared using χ² or Fisher's exact tests, as appropriate. Non-normally distributed continuous variables were analysed using the Mann–Whitney U test. Risk factors for mortality were analysed using logistic regression analysis. Factors with P-values <0.100 in the univariate analysis were included and adjusted for in the multiple logistic regression analysis. Statistical significance was set at P < 0.05. Data were analysed using PASW Statistics for Windows (version 18.0; SPSS Inc., Chicago, IL, USA).

Results

Demographic data

During the 6 years, 42 780 anaesthetics were administered for noncardiac surgery in patients aged 0 to 17 years at our institution. Four anaesthetics were excluded because of intra-operative organ donation, resulting in 42 776 anaesthetics for the final analysis. Infants (aged <12 months) accounted for 12.5% of all anaesthetics. Most of the anaesthetics delivered were for patients with ASA physical status I–II (79.0%) and elective cases (95.8%) (Table 1). All participants were followed-up until hospital discharge. No participant was excluded because of incomplete data.

Table 1.

Demographic data of all anaesthetics, incidence of peri-operative cardiac arrest, and mortality between 2014 and 2019

	n (%)	Cardiac arrest	Cardiac arrest per 10 000 anaesthetics (95% CI)	In-hospital mortality	Mortality per 10 000 anaesthetics (95% CI)
Overall	42 776	63	14.7 (11.5 to 18.8)	23	5.4 (3.6 to 8.1)
Age
0–11 months	5365 (12.5)	36	67.1 (48.5 to 92.8)	12	22.4 (12.8 to 39.1)
1–3 years	13 472 (31.5)	12	8.9 (5.1 to 15.6)	4	3.0 (1.2 to 7.6)
4–8 years	11 314 (26.4)	8	7.1 (3.6 to 13.9)	3	2.7 (0.9 to 7.8)
9–12 years	5819 (13.6)	4	6.9 (2.7 to 17.7)	2	3.4 (0.9 to 12.5)
13–17 years	6806 (15.9)	3	4.4 (1.5 to 13.0)	2	2.9 (0.8 to 10.7)
Male sex	25 336 (59.2)	32	N/A	12	N/A
ASA classification
I	16 063 (37.5)	1	0.6 (0.1 to 3.5)	0	0.0 (0.0 to 2.4)
II	17 746 (41.5)	6	3.4 (1.5 to 7.4)	2	1.1 (0.3 to 4.1)
III	8438 (19.7)	41	48.6 (35.8 to 65.8)	13	15.4 (9.0 to 26.3)
IV	502 (1.2)	12	239.0 (137.3 to 413.1)	6	119.5 (54.9 to 258.3)
V	27 (0.1)	3	1111.1 (385.2 to 2805.8)	2	740.7 (205.5 to 2337.0)
Elective surgery	40 969 (95.8)	40	9.8 (7.2 to 13.3)	10	2.4 (1.3 to 4.5)
Emergency/urgent surgery	1807 (4.2)	23	127.3 (85.0 to 190.3)	13	71.9 (42.1 to 122.7)
Department of service
Thoracic	494 (1.2)	2	40.5 (11.1 to 146.4)	1	20.2 (3.6 to 113.8)
General	19 200 (44.9)	13	6.8 (4.0 to 11.6)	7	3.6 (1.8 to 7.5)
Ear-Nose-Throat	5647 (13.2)	5	8.9 (3.8 to 20.7)	2	3.5 (1.0 to 12.9)
Orthopaedics	3579 (8.4)	0	0 (0 to 10.7)	0	0.0 (0.0 to 10.7)
Neurosurgery	994 (2.3)	3	30.2 (10.3 to 88.4)	2	20.1 (5.5 to 73.1)
Trauma	861 (2.0)	1	11.6 (2.1 to 65.5)	1	11.6 (2.1 to 65.5)
Non-OR anaesthesia	9584 (22.4)	4	4.2 (1.6 to 10.7)	2	2.1 (0.6 to 7.6)
Cardiac catheterization	2417 (5.7)	35	144.8 (104.3 to 200.7)	8	33.1 (16.8 to 65.2)

CI, confidence interval; OR, operating room.

Incidence of cardiac arrests

During the six years, 63 PPOCAs were identified within 24 h of anaesthesia. Overall, the incidence of PPOCA after noncardiac surgery was 14.7 (95% CI 11.5 to 18.8) per 10 000 anaesthetics, and each patient had only one PPOCA. The incidence of PPOCA was higher in infants, physical status III–V, emergency/urgent surgeries, and cardiac catheterisation procedures (Table 1). The intraoperative and postoperative cardiac arrest incidence was 9.6 (95% CI 7.1 to 13.0) and 5.1 (95% CI 3.4 to 7.8) per 10 000 anaesthetics, respectively. The incidence of ARCA and NARCA was 7.9 (95% CI 5.7 to 11.1) and 6.8 (95% CI 4.7 to 9.7) per 10 000 anaesthetics, respectively (Table 2).

Table 2.

Incidence of peri-operative cardiac arrest and mortality by intra-operative vs. postoperative period (total 42 776 anaesthetics)

	Overall (n = 63)		Intra-operative CA (n = 41)		Postoperative CA (N = 22)
	n	Per 10 000 anaesthetics (95% CI)	n	Per 10 000 anaesthetics (95% CI)	n	Per 10 000 anaesthetics (95% CI)
Cardiac arrest
Overall	63	14.7 (11.5 to 18.8)	41	9.6 (7.1 to 13.0)	22	5.1 (3.4 to 7.8)
ARCA	34	7.9 (5.7 to 11.1)	20	4.7 (3.0 to 7.2)	14	3.3 (1.9 to 5.5)
NARCA	29	6.8 (4.7 to 9.7)	21	4.9 (3.2 to 7.5)	8	1.9 (0.9 to 3.7)
Mortality
Overall	23	5.4 (3.6 to 8.1)	9	2.1 (1.1 to 4.0)	14	3.3 (1.9 to 5.5)
ARCA	14	3.3 (1.9 to 5.5)	7	1.6 (0.8 to 3.4)	7	1.6 (0.8 to 3.4)
NARCA	9	2.1 (1.1 to 4.0)	2	0.5 (0.1 to 1.7)	7	1.6 (0.8 to 3.4)

ARCA, anaesthesia-related cardiac arrest; CI, confidence interval; NARCA, nonanaesthesia-related cardiac arrest.

Characteristics of cardiac arrests

Of the 63 PPOCAs, 41 (65%) occurred intra-operatively, and 22 (35%) occurred during the postoperative period. Median [range, IQR] time for postoperative cardiac arrest was 3.84 [0.05 to 19.47, 0.60 to 10.79] h after anaesthesia completion. Overall, the leading causes of PPOCA were cardiovascular-related (58.7%), followed by airway and respiratory causes (31.7%) (Fig. 1 and Table 2, Supplemental Digital Content). The most common cause of intra-operative cardiac arrest was cardiovascular-related (80.5%). In contrast, airway and respiratory origins were the most common causes of postoperative cardiac arrest (59.1%), P < 0.001 (Table 3). Participants who experienced intraoperative cardiac arrest had a higher ASA physical status (95.1%), were mainly undergoing cardiac catheterisation (73.2%), and received shorter CPR duration (P = 0.045, <0.001, and 0.003, respectively). Details of 34 ARCAs are described in Table 4.

Fig. 1.

Causes of cardiac arrest for noncardiac surgery (n = 63 cardiac arrests).

(a) Categorised by intra-operative vs. postoperative, (b) categorised by anaesthesia-related cardiac arrest (ARCA) vs. nonanaesthesia-related cardiac arrest (NARCA).

Table 3.

Characteristics of peri-operative cardiac arrest by the timing of the cardiac arrest. Data are presented as number (%) or median [IQR]

Characteristics	Total (63)	Intraoperative CA (41)	Postoperative CA (22)	P value
Time of cardiac arrest				N/A
- During induction	4 (6.3)	4 (9.8)	0
- During maintenance	36 (57.1)	36 (87.8)	0
- During emergence	1 (1.6)	1 (2.4)	0
- In PACU	5 (7.9)	0	5 (22.7)
- During transfer	5 (7.9)	0	5 (22.7)
- In ICU	9 (14.3)	0	9 (40.9)
- In ward	3 (4.8)	0	3 (13.6)
Infant	36 (57.1)	24 (58.5)	12 (54.5)	0.760
High ASA PS	56 (88.9)	39 (95.1)	17 (77.3)	0.045
Heart disease	46 (73.0)	34 (82.9)	12 (54.5)	0.020
Emergency/urgency	23 (36.5)	15 (36.6)	8 (36.4)	0.986
Multiple operations	25 (39.7)	16 (39.0)	4 (40.9)	0.884
Out of hours	9 (14.3)	7 (17.1)	2 (9.1)	0.476
Extracorporeal CPR	4 (6.3)	2 (4.9)	2 (9.1)	0.606
Service				< 0.001
- Cardiac catheterisation	35 (55.6)	30 (73.2)	5 (22.7)
- Thoracic	2 (3.2)	2 (4.9)	0
- Ear-Nose-Throat	5 (7.9)	1 (20.0)	4 (18.2)
- Neurosurgery	3 (4.8)	3 (7.3)	0
- General	13 (20.6)	2 (4.9)	11 (50.0)
- Trauma	1 (1.6)	1 (2.4)	0
- Non-OR anaesthesia	4 (6.3)	2 (4.9)	2 (9.1)
Surgical severity (moderate – high)	39 (61.9)	27 (65.9)	12 (54.5)	0.378
Anaesthesiologist with subspecialty	39 (61.9)	26 (63.4)	13 (59.1)	0.736
Anaesthesiologist experience > 15 years	19 (30.2)	14 (34.1)	5 (22.7)	0.346
Anaesthesia-related	34 (54.0)	20 (48.8)	14 (63.6)	0.259
Cause of cardiac arrest				< 0.001
- Cardiovascular	37 (58.7)	33 (80.5)	4 (18.2)
- Airway and respiratory	20 (31.7)	7 (17.1)	13 (59.1)
- Metabolic	3 (4.8)	1 (2.4)	2 (9.1)
- Medication	2 (3.2)	0	2 (9.1)
- Other	1 (1.6)	0	1 (4.5)
CPR duration	6 [2 to 15.3]	4 [2 to 10]	11 [4.5 to 85.0]	0.003

ASA PS, American Society of Anaesthesiologists physical status; CA, cardiac arrest; CPR, cardiopulmonary resuscitation; ICU, intensive care unit; OR, operating room; PACU, postanaesthesia care unit.

Table 4.

Details of 34 peri-operative anaesthesia-related cardiac arrests (ARCAs) in a noncardiac surgery population

Case no.	Age	BW (kg)	ASA	Diagnosis	Procedure	Cause of cardiac arrest	Event leading to cardiac arrest	CPR duration (min)	Outcome
1.	8 years	33.0	2	PVT/SVT	EPS with RFA	Arrhythmia	Developed VF/VT after sedation, before procedure start	2	Survive
2.	4 months	5.0	4E	Acute subdural haematoma	Craniectomy	Haemorrhage	Blood loss 400 ml	5	Death POD7
3.	17 years	70.0	5E	Liver and IVC injury, traumatic brain injury	Exploratory laparotomy	Haemorrhage	Blood loss 5000 ml	38	Death
4.	9 years	25.0	4E	Acute subdural haematoma with liver injury	Craniectomy	Haemorrhage	Blood loss 500 ml	7	Brain death, Death POD11
5.	4 weeks	3.1	4	Brain tumour	Craniotomy with tumour removal	Haemorrhage	Bleeding 500 ml, massive transfusion lead to hyperkalaemia	8	Survive
6.	8 days	2.2	3E	HLHS with truncus arteriosus with PDA	Cardiac catheterisation with atrial septostomy	Haemorrhage	Blood loss 10 ml	1	Survive
7.	9 days	2.9	3E	Pulmonic stenosis	Cardiac catheterisation with PBPV	Haemorrhage	Blood loss 30 ml	1	Survive
8.	1 month	3.3	3	Hypoplastic aortic arch	Cardiac catheterisation with angioplasty	Haemorrhage	Blood loss 30 ml	3	Survive
9.	20 days	3.0	3	Dilated cardiomyopathy	Cardiac catheterisation	Haemorrhage	Blood loss 40 ml, Haematocrit 19%	2	Survive
10.	10 months	7.5	3	PA with VSD s/p RMBTS	Cardiac catheterisation	Hypovolaemia	Hypoxia from shunt thrombosis after procedure due to polycythaemia	60 then ECMO	Survive with HIE
11.	3 years	11.7	3	PA, VSD s/p RMBTS	Cardiac catheterisation	Pulmonary hypertension	2 hr postop, cry and hypercyanotic spell on ward (work up found sepsis)	90	Death
12.	8 months	5.8	3	UVH s/p RMBTS	Cardiac catheterisation	Pulmonary hypertension	3 hr postop, hypercyanotic spell at regular ward	60	Death
13.	6 months	6.0	5	Complete tracheal ring with TOF	Sliding trabeculectomy	Unclear cardiovascular	Frequent hypoxic spell with history of failed intubation, developed bradycardia during induction while successful ventilation via LMA.	16	Survive
14.	12 years	20.0	4	Severe Ebstein anomalies	Cardiac catheterisation	Unclear cardiovascular	Ebstein anomaly with severe hypoxia, cardiac arrest for 8 episodes	N/A	Death
15.	6 months	5.6	3	Hypoplastic left heart syndrome	Cardiac catheterisation	Unclear cardiovascular	30 min after PACU arrival, developed bradycardia & hypotension.	120	Death
16.	12 months	8.5	1	Bilateral complete cleft lip	Cheiloplasty	Narcotic	Crying at PACU, fentanyl 10 μg was given and found cyanosis 10 min later	2	Survive with HIE
17.	1 year 5 months	9.5	4E	UVH with partial thrombosed LMBTS with MAPCA	Computed tomographic angiogram (CTA)	Unclear medication event	Cardiac arrest on PACU arrival after 35 min of i.v. sedation with midazolam and propofol	5	Survive
18.	4 years	10.5	3	ASD, VSD s/p repair with AV block	Permanent pacemaker	Aspiration	Gross pulmonary aspiration during face mask ventilation after induction. Cannot maintain oxygenation for 4 h then developed cardiac arrest.	10 then ECMO	Brain death, Death POD27
19.	23 days	3.5	3E	Congenital diaphragmatic hernia with lung sequestration	Thoracotomy	Oesophageal intubation	Oesophageal intubation in neonate with lung pathology	3	Survive
20.	1 day	2.6	3E	Esophageal atresia with anorectal malformation	Thoracotomy & colostomy	Oesophageal intubation	Loss of end-tidal CO2 after successful intubation due to ETT displacement into tracheoesophageal fistula	4	Survive
21.	10 months	7.0	3	ASD VSD PS	Cardiac catheterisation + PBPV + ASD device	Inability to intubate	ETT No. 3.0 was required, suspected subglottic stenosis due to intubation in neonatal period	11	Survive
22.	2 years	14.1	2E	Subglottic stenosis s/p tracheostomy	DL+FOL	Inability to ventilate	Cannot ventilate during changing tracheostomy tube	2	Survive
23.	4 years	13.0	3	Severe PS	Cardiac catheterisation with PBPV	Oxygenation failure	20 min after PACU arrival, severe pulmonary edema	80 then ECMO	Death POD7
24.	17 years	112.0	3E	Systemic lupus erythoematosus with thrombocytopenia and intracranial haemorrhage	Splenectomy	Oxygenation failure	BMI 41.6 kg m–2, cardiac arrest during transfer due to changes in ventilator setting	30	Death
25.	3 months	4.6	4E	TAPVR with pneumonia	Cardiac catheterisation	Oxygenation failure	Baseline SpO2 82% but SpO2 was only 66 to 76% in the OR	48	Death POD21
26.	6 weeks	2.3	2E	Retinopathy of prematurity	Laser surgery	Pneumothorax	Bilateral pneumothorax on the ambulance during transfer back to primary hospital	∗	Survive
27.	2 months	5.8	2	Bilateral inguinal hernia	Bilateral herniotomy	Premature extubation	ETT dislodge during transfer	2	Survive
28.	3 months	4.5	3	Bilateral inguinal hernia	Bilateral herniotomy	Premature extubation	ETT dislodge during transfer	2	Survive
29.	4 months	4.2	3	Anterior glottic web, multiple anomalies	DL with CO2 laser	Premature extubation	ETT dislodge during transfer	5	Death 8 months later due to respiratory failure
30.	1 year 2 months	9.2	3	Biliary atresia with gastrointestinal bleeding	EGD with sclerosing agent injection	Premature extubation	ETT dislodge during oesophagogastroscopy	1	Death POD6 due to massive haemorrhage
31.	2 years	9.5	3	TOF S/P RMBTS	Cardiac catheterisation	Unclear respiratory cause	On PACU arrival, developed hypoxia then bradycardia.	4	Survive with HIE
32.	1 year 10 months	9.0	4	Tracheomalacia with adenoid hypertrophy with TOF	DL with adenoidectomy, auditory steady-state response	Upper airway obstruction	Extubation at 19 hr postop and developed upper airway obstruction, work up found vascular ring.	4	Survive
33.	7 years	15.0	3	End-stage renal disease	Kidney transplantation	Electrolyte abnormalities	Hyperkalaemia prior to reperfusion, potassium 5.6 mmol l–1	1	Survive
34.	9 months	5.2	3	Intestinal malabsorption	Central venous line insertion	Could not be determined	7 hr postop, regular ward, found cyanosis	131	Death

ASD, atrial septum defect; BMI, body mass index; BW, body weight; CPR, cardiopulmonary resuscitation; DL, direct laryngoscope; ECMO, extracorporeal membrane oxygenation; EGD, oesophagogastroduodenoscopy; EPS, electrophysiology study; ETT, endotracheal tube; FOL, fiberoptic laryngoscopy; HIE, hypoxic-ischaemic encephalopathy; HLHS, hypoplastic left heart syndrome; LMA, laryngeal mask airway; LMBTS, left modified Blalock-Taussig shunt; PA, pulmonary atresia; PACU, postanaesthesia care unit; PBPV, percutaneous balloon pulmonary valvuloplasty; PDA, patent ductus arteriosus; POD, postoperative day; PS, pulmonary stenosis; PVT, paroxysmal ventricular tachycardia; RFA, radiofrequency ablation; RMBTS, right modified Blalock-Taussig shunt; SVT, supraventricular tachycardia; TOF, Tetralogy of Fallot; UVH, univentricular heart; VF, ventricular fibrillation; VSD, ventricular septal defect; VT, ventricular tachycardia.

Outcomes of peri-operative cardiac arrests

In-hospital mortality after PPOCA was 5.4 (95% CI 3.6 to 8.1) per 10 000 anaesthetics. The incidence of intra-operative and postoperative mortality was 2.1 (95% CI 1.1 to 4.0) and 3.3 (95% CI 1.9 to 5.5) per 10 000 anaesthetics, respectively (Table 2). Sixty-three PPOCAs resulted in 23 (36.5%) in-hospital mortalities (Table 5). Twelve (19.0%) patients died within 24 h after PPOCA, and 11 (17.5%) died after 24 h. The mortality (mortality rate) of postoperative cardiac arrest was significantly higher than that of intra-operative cardiac arrest, 14 (63.6%) vs. 9 (22.0% P = 0.001). The incidence of PPOCA was the highest for anaesthesia for cardiac catheterisation (144.8 per 10 000 anaesthetics). Post hoc analysis showed that the mortality rate after PPOCA in cardiac catheterisation was significantly higher if cardiac arrest occurred during the postoperative period (80%) compared to intra-operative cardiac arrest (13.3%), P = 0.006. There was no difference in mortality between postoperative and intra-operative cardiac arrest in anaesthesia services for other procedures (P = 0.488). The mortality rates categorised by the causes of cardiac arrest are provided in Table 3, Supplemental Digital Content.

Table 5.

Outcomes of cardiac arrest. The data are presented as n (%)

Outcomes	Overall (63)	Intra-operative CA (41)	Postoperative CA (22)	P value
Mortality	23 (36.5)	9 (22.0)	14 (63.6)	0.001
- Temporary harm	37 (58.7)	31 (75.6)	6 (27.3)
- Permanent harm	3 (4.8)	1 (2.4)	2 (9.1)
- Death within 24 h	12 (19.0)	2 (4.9)	10 (45.5)
- Death after 24 h	11 (17.5)	7 (17.1)	4 (18.2)
Subgroup by type of service
Cardiac catheterisation	n = 35	n = 30	n = 5
Survive	27 (77.1)	26 (86.7)	1 (20.0)	0.006
Death	8 (22.9)	4 (13.3)	4 (80.0)
Operating room and others	n = 28	n = 11	n = 17
Survive	13 (46.4)	6 (54.5)	7 (41.2)	0.488
Death	15 (53.6)	5 (45.5)	10 (58.8)

CA, cardiac arrest.

Multivariate analysis of the risk factors for mortality was adjusted for postoperative cardiac arrest, cardiac catheterisation, heart disease, emergency/urgency, cardiovascular cause, and CPR duration. Only emergency/urgency status and CPR duration were predictors of mortality (Table 6). The adjusted odds ratio of emergency/urgency was 5.388 (95% CI 1.031 to 28.160), P = 0.046. The adjusted odds ratio for CPR duration (min) was 1.067 (95% CI 1.016 to 1.120), P = 0.010.

Table 6.

Univariate and multivariate analyses of risk factors for mortality

	Univariate analysis			Multivariate analysis
	Odds ratio	95% CI	P value	Adjusted odds ratio	95% CI	P value
Postoperative cardiac arrest	6.222	1.988 to 19.471	0.002	3.340	0.395 to 28.265	0.268
Cardiac catheterisation	0.257	0.087 to 0.759	0.014	0.733	0.056 to 9.614	0.813
Infant	0.727	0.259 to 2.046	0.546	–	–	–
High ASA PS	1.500	0.267 to 8.434	0.645	–	–	–
Heart disease	0.193	0.058 to 0.635	0.007	0.183	0.024 to 1.408	0.103
Emergency/urgency	3.900	1.309 to 11.620	0.015	5.388	1.031 to 28.160	0.046
Multiple operations	1.282	0.451 to 3.641	0.641	–	–	–
Cardiovascular cause (y/n)	0.370	0.129 to 1.066	0.066	1.387	0.145 to 13.315	0.777
Respiratory cause (y/n)	2.308	0.775 to 6.875	0.133	–	–	–
CPR duration (minute)	1.072	1.018 to 1.130	0.009	1.067	1.016 to 1.120	0.010

ASA PS, American Society of Anaesthesiologists physical status; CPR, cardiopulmonary resuscitation.

Discussion

This study reported the incidence of PPOCA after noncardiac surgery and in-hospital mortality after cardiac arrest in a single-centre tertiary-care university hospital in Thailand. The incidence of PPOCA and ARCA was 14.7 and 7.9 per 10 000 anaesthetics. Compared with the THAI study conducted nationwide from 2003 to 2004, a previous study reported the incidence of PPOCA and ARCA as 19.9 and 5.1 per 10 000 anaesthetics.⁸ The incidence of PPOCA was similar to that in other developing countries (2.7 to 22.9 per 10 000 anaesthetics)^8–13 while the overall incidence of PPOCA reported in developed countries was lower (2.9 to 7.9 per 10 000 anaesthetics).^1–3,5–7 However, the incidence of ARCA was similar in developing and developed countries (1.4 to 7.4 per 10 000 anaesthetics),^5,8,10,11,16 suggesting that the difference in PPOCA may be due to other factors such as surgical or patient-related factors. The mortality rates differed among studies, possibly due to the different follow-up time points.^{3–6,10–15} The studies reporting the incidence of PPOCA in different settings are summarised in Table 7.

Table 7.

Incidence of paediatric peri-operative cardiac arrest (PPOCA), anaesthesia-related cardiac arrest (ARCA), and mortality in different settings. The incidence was reported per 10 000 anaesthetics

	Author, year	Country	Definition of the peri-operative period	Setting	Incidence of PPOCA	Incidence of ARCA	Incidence of mortality	Time of mortality evaluation
Developed Countries	Flick et al.3, 2007	USA	PACU	Noncardiac	2.9	–	1.6	In-hospital
	Habre et al.2, 2017	33 European countries	60 min after anaesthesia	Mix	2.9a	–	–	–
	Zgleszewski et al.5, 2016	USA	PACU or transfer to ICU	Noncardiac, excluding cardiac catheterisation	5.1	2.6	0.3	In-hospital
	Christensen et al.1, 2018	USA	24 h after induction	Mix	5.3	3.3	–	–
	Hohn et al.7, 2019	Germany	24 h after anaesthesia	Mix	6.9	3.3	–	–
	Jansen et al.6, 2021	Germany	60 min after anaesthesia	Noncardiac	7.9	4.0	3.5	30-day
	Morray et al.15, 2000	USA	PACU	Mix	–	1.4	–	–
	van der Griend et al.4, 2011	Australia	N/A	Noncardiac	–	–	8.2	24-h
	De Bruin et al.14, 2015	Netherlands	N/A	Noncardiac	–	–	33.6	30-day
Developing countries	Khoso et al.13, 2021	Pakistan	48 h after anaesthesia	Mix	2.7	–	2.0	48-h
	Lee et al.9, 2016	Korea	Not Defined	Mix	8.5	–	–	–
	Bunchungmongkol et al.8, 2009	Thailand	24 h after anaesthesia	Mix	19.9	5.1	–	–
	Gonzalez et al.10, 2014	Brazil	PACU	Mix	20.7	2.8	10.3	PACU
	Bharti et al.11, 2009	India	PACU	Noncardiac, except in remote locations	22.2	7.4	10.7	In-hospital
	Gobbo Braz et al.12, 2006	Brazil	PACU	Mix	22.9	4.6	9.8	Not Defined
	Zoumenou et al.13, 2010	Benin	Not Defined	Mix	156.0	–	97.0	Not Defined

ICU, intensive care unit; PACU, postanaesthesia care unit; USA, United States of America.

^aData per 10 000 patients.

Differences in the characteristics between patients who experienced intra-operative and postoperative cardiac arrest were also demonstrated. Most intra-operative cardiac arrests occurred in patients with high ASA PS ≥ III (95.1%) and during cardiac catheterisation services (73.2%). Patients with high ASA PS are known to have individual risks for increased peri-operative adverse events.^{1,10,11,15,16} The incidence of cardiac arrests during cardiac catheterisation procedures has been reported to be as high as 0.3 to 3.0% in the literature.^21–26 Our cohort consisted of 35 (55.6%) catheterisation procedures in patients with heart disease; 19 cardiac arrests were related to the cardiology wire and successfully resuscitated in the operating room.

The most common cause of cardiac arrests during anaesthesia was cardiovascular (80.5%), while airway and res-piratory (59.1%) were the leading causes of postoperative cardiac arrest. This is similar to the Wake-Up Safe Registry in which cardiovascular-related causes became the most common cause of PPOCAs (cardiovascular 40%, respiratory 35%).¹ Christensen et al. also reported respiratory-related aetiology as the leading cause of cardiac arrest in the PACU (respiratory 44.4%, cardiovascular 22.2%), which was deemed preventable.²⁷ Intra-operative cardiac arrest is often associated with an acute cause and is rapidly recognised and successfully resuscitated. However, postoperative cardiac arrest is often airway-related and is associated with prolonged hypoxia. Cardiac arrest is often the last event after progressive deterioration. This may explain the poorer outcome in these patients and the requirement of prolonged resuscitation efforts. In our cohort, five cardiac arrests were identified as transfer-related, and all suffered from respiratory causes. Patient transfer often involves the transition of care between providers and changes in the level of sedation and physical positioning.^19,28 According to the Wake-Up Safe registry,²⁸ 148 transport-associated adverse events comprised 5.0% of all paediatric anaesthesia adverse events reported. They were primarily respiratory in nature and preventable.

To our knowledge, this is the first study to compare postoperative and intra-operative cardiac arrest mortality rates in paediatric patients. Postoperative cardiac arrest had significantly higher mortality than intra-operative cardiac arrest (odds ratio 6.222, 95% CI 1.988 to 19.471, P = 0.002). However, after adjusting for other factors, only emergency/urgency status and CPR duration affected the mortality. In the operating room, patients were monitored with the highest vigilance by an individual anaesthesia provider using continuous monitoring for the early detection of critical events. This environment is unique and different from the PACU or wards, where personnel and monitoring are less intensive than during the intra-operative period. Our results indicated that postoperative cardiac arrest had a longer median CPR duration than intra-operative cardiac arrest (11 min vs. 4 min, P = 0.003), and the adjusted odds ratio (95% CI) of mortality for CPR duration in minutes was 1.067 (1.016 to 1.120). The CPR duration was shown to be inversely associated with survival to hospital discharge.²⁹ Although patients in the ICU usually received continuous monitoring and intensive care from experienced providers, patients admitted to the ICU were considered critically ill relative to the patients admitted to the ward, had higher ASA PS, and hence, were at greater risk from anaesthesia. According to Jansen et al., 2.7% of children were critically ill, and anaesthesia was induced in the ICU.⁶ The incidence (95% CI) of cardiac arrests with anaesthesia induced in the ICU was 131.6 (57 to 257.6) per 10 000 anaesthetics compared to only 4.5 (2.2 to 8.3) per 10 000 anaesthetics when anaesthesia was performed in the operating room (P < 0.001).⁶ They concluded that children who had anaesthesia induction in the ICU were at high risk for PPOCA and PPOCA-associated mortality.

Emergency/urgency status was a predictive factor for mortality in this study (adjusted odds ratio 5.388, 95% CI 1.031 to 28.160) and consistent with multiple studies.^1,11,15,16 High ASA PS^{1,10,11,15,16} and infant^10,11,15 were reported as predictors of mortality in several studies. In general, cardiac arrests of respiratory origin were associated more often with no harm, whereas cardiac arrests of cardiac origin were associated with greater levels of harm.^6,19 However, we cannot demonstrate a significant effect of these factors on mortality after multivariate analysis: cardiovascular cause, respiratory cause, infant, high ASA PS, cardiac catheterisation procedure, and heart disease. These inconsistencies in the findings suggest that multiple factors, including patient, procedural, and institutional characteristics, contribute to the outcomes of PPOCA. Practitioners should focus on individual pathophysiology.⁶

Overall, 35% of peri-operative cardiac arrests occurred during the postoperative period. After excluding the cardiac catheterisation procedure and focusing only on anaesthesia services for other procedures, 17 (61%) of 28 cardiac arrests occurred postoperatively. Of the 17 cardiac arrests, 11 (64.7%) were attributed to airway and respiratory aetiologies (Table 2, Supplemental Digital Content). Patients with postoperative cardiac arrest were mostly respiratory in origin and associated with longer CPR duration and poorer prognosis. We speculate that some of these events might have been prevented, at least in part, by a higher level of care. We encouraged the implementation of postoperative observation protocols in the PACU and ward, including criteria for postoperative ICU admission in high-risk patients. Healthcare providers must extend a higher level of care for surgical patients for at least 24 h after surgery. Implementing a quality improvement project for greater standardisation of clinical practice has been proven to reduce the incidence of anaesthesia-attributable cardiac arrests⁷ and airway cardiac arrests.³⁰ Moreover, a longer CPR duration is associated with poor survival. Emphasising institutional protocols for extracorporeal cardiopulmonary resuscitation (ECPR) initiation during the early stage of CPR is another important factor for improving patient outcomes.

Our study had several limitations. First, there was a risk of underreporting in this retrospective study, including events that occurred within 24 h after discharge. Second, the present study was conducted in a single tertiary-care centre that usually admits patients of complex diseases; therefore, the incidence of cardiac arrest and mortality cannot be generalised to the national paediatric anaesthesia practice in Thailand. Third, the numbers of cardiac arrest cases in this study were influenced by the large contribution (55.6%) of cardiac catheterisation cases in only 5.7% of a noncardiac surgery population. Fourth, the outcome was reported as in-hospital mortality. There is a possibility that the cause of mortality was different from the cause of cardiac arrest. However, this study reported the number of patients who died within 24 h after cardiac arrest. Finally, the study was conducted at a training centre, and we could not evaluate the extent of involvement of different levels of anaesthesia providers.

In summary, paediatric peri-operative cardiac arrests that occurred intra-operatively and postoperatively had different characteristics and outcomes. Postoperative cardiac arrest resulted in a longer CPR duration and a higher mortality rate than intra-operative cardiac arrest. The emergency/urgency status and CPR duration were predictors of mortality after PPOCAs. A high level of care should be applied to high-risk groups of patients for at least 24 h after surgery.