Major interventions are associated with survival of out of hospital cardiac arrest patients-a population based survey

Background. The overall survival rate of out-of-hospital cardiac arrest (OHCA) in Taiwan or even in the whole of Asia is relatively low. Major interventions, such as target temperature management (TTM), coronary artery angiography, and extracorporeal membrane oxygenation (ECMO), have been associated with better patient outcome. However, studies in Taiwan revealing evidence of the benefits of these interventions are limited. Methods. A population-based study used an 8-year database to analyze overall survival and risk factors  ̋among OHCA patients. All adult non-trauma OHCA patients were identified through diagnostic and procedure codes. Hospital survival and return of spontaneous circulation (ROSC) were primary and secondary outcomes. Logistic regression and Cox regression analyses were conducted. Results. There was a relationship between major interventions (including TTM, coronary artery angiography, and ECMO) and better hospital survival. Age, income, major interventions, and acute myocardial infarction history were associated with hospital survival. The adjusted hazard ratios (HRs) were 0.406 (95% CI, 0.295 to 0.558), 1.109 (95% CI, 1.027 to 1.197), 1.075 (95% CI, 1.002 to 1.154), 1.097 (95% CI, 1.02 to 1.181) and 0.799(95% CI, 0.677 to 0.942) for patients with major interventions, age≥50, medium low and low income, middle income, and acute myocardial infarction history, respectively. Conclusion. This population-based study in Taiwan revealed that older age (≥50), medium low and low income were associated with a lower rate of survival. Major interventions, including TTM, coronary angiography, and ECMO, were related to better survival.


INTRODUCTION
Out-of-hospital cardiac arrest (OHCA) remains a leading cause of mortality and a substantial issue for public health worldwide.According to a statistical update from the American Heart Association, the incidence of OHCA in the United States is approximately 359,400 each year with an overall survival rate of 9.5%.(1) Overall survival to hospital discharge rates are lower in the Asian population with rates ranging from 0.5% to 8.5%.(2)(3)(4) The overall survival to hospital discharge in Taiwan ranges from 1.4% to 6%. (5)(6)(7) To overcome the low survival rate, the concept of an out-of-hospital chain of survival has been established, including 1) early recognition of cardiac arrest and activation of the emergency medical service (EMS) system, 2) early cardiopulmonary resuscitation (CPR) with highlight on high quality chest compressions, 3) early defibrillation, 4) basic and advanced EMS, 5) advanced life support (ALS) and post-cardiac arrest care.(8) Several interventions, such as extracorporeal membrane oxygenation (ECMO), (9,10) target temperature management (TTM), (11,12) and percutaneous coronary intervention, (13,14) have been associated with favorable patient outcomes.In Taiwan, there is limited data on major interventions (TTM, coronary artery angiography, ECMO) and their influence on patient outcomes.The aim of our study was to assess the impact of patient demographic features, hospital characteristics, and major interventions on prognoses of OHCA.

Data Sources
The National Health Insurance (NHI) program was conducted in Taiwan in 1995 and provides compulsory universal health insurance.It enrolls about 99% of the Taiwanese population and contracts with 97% of all the country's health care services.(15) For research purposes, a random sample of 1,000,000 people who received health services from the NHI program was selected based on calendar-year 2005 reimbursement data.According to the Taiwan National Health Research Institute (NHRI), the random sample group is considered representative of the entire population since the group did not diff er statistically from the whole cohort in age, sex, or health care costs.Th is random sample of 1,000,000 people was used as our study cohort.Information about admissions and outpatient visits, both of which included information on patient characteristics (sex, date of birth, date of admission, date of discharge, dates of visits) and up to fi ve discharge diagnoses or three out-patient visit diagnoses (based on the International Classifi cation of Diseases, Ninth Revision [ICD-9]), was collected.( 16) We also gained information on patient prescriptions, including the names of prescribed drugs, dosage, duration, and total expenditure.Previously used for epidemiologic research, the NHRI database provided high quality information on prescription use, diagnoses, and hospitalizations (17) and this database had been previously used for OHCA research.(18,19) Th e NHRI strictly followed confi dentiality guidelines in accordance with personal electronic data protection regulations, making the NHI reimbursement data suitable for research.(15) In addition, this study was approved by Chang Gung Medical Foundation Institutional Review Board.

Identifi cation of Study Sample
We carried out a population-based cohort study in which we identifi ed all alive patients older than 18 years who suffered from out-of-hospital cardiac arrest (OHCA) (ICD-9 codes 427.5, 798.9, 798.1, 798.2) and were admitted to the emergency department (ED) between January 1, 2005 and December 31, 2012.Th us, we formed a study cohort with a total number of 6,815 individuals.Individuals with cardiac arrest announced death by EMS at the scene or without transportation to the hospital were not included because EMS records were not available in our NHI database.In Taiwan, prehospital emergency care is provided by emergency medical technicians (EMTs) in diff erent fi re bureaus.Most EMTs are EMT-II, but some cities have paramedics (EMT-Ps).Once a patient with cardiac arrest is identifi ed, EMTs or EMT-Ps do not terminate resuscitation even when the patient had signed a do not resuscitate (DNR) consent or has signs of obvious death (rigor, mortis, lividity, decapitation etc.) unless the patient's family understands and requests to discontinue.Complete prehospital advanced life support (ALS) is performed only in some urban areas since the number of EMT-Ps is limited.On the basis of an evaluation in urban Taipei, ALS demand accounted for approximately 9-16% of total EMS calls, and the average response time was 4.1-4.9minutes.In rural areas, the average response time was found to be longer, even exceeding 6.6 minutes.(20) Two hundred and twenty-four (224) individuals with trauma, intoxication, submersion, and accidents (ICD-9 codes 800-999 and E800-999) were excluded.Th ose who did not receive any CPR, defi brillation, or endotracheal intubation during ED stay were also excluded (number =1253).All enrolled patients were followed up until December 31, 2013 (Figure1).Th e primary outcome was hospital survival, which was defi ned as survival to hospital discharge.Th e secondary outcome was return of spontaneous circulation (ROSC), which was defi ned in this study as sustained ROSC with hospital admission.

Potential Covariates and Management
We identifi ed in a systematic way any comorbid conditions as potential confounders, defi ned by the following diagnoses recorded before January 1, 2005: angina (ICD-9 codes 413.0, 413.1,and 413.9), acute myocardial infarction (ICD-  ).Th e CCI is a scoring system that strengthens important concomitant diseases and has been validated for studies that employ ICD-9 data.(18) We also considered socio-demographic characteristics (age, sex, level of urbanization, level of hospital, and level of monthly income) in the modeling.Level of hospital is classifi ed as either medical center or non-medical center.( 21) Urbanization levels in Taiwan are divided into seven groups according to the Taiwan National Health Research Institute publications.(22) Group 1 is called highly urbanized area, group 2 is moderately urbanized area, group 3 is emerging area, group 4 is gen-eral rural area, group 5 is aging area, group 6 is agricultural area, and group 7 is remote area.(22) In our study, groups 4, 5, 6, and 7 are merged into a single group, called general rural area.Th us, urbanization levels in our report are classifi ed as highly urbanized area, moderately urbanized area, emerging area and general rural area.Individual monthly income in New Taiwan dollars (NTD) is divided into four groups: 0,1-15,840,15,841-25000 and ≥25,001.According to the Ministry of Health and Welfare of Taiwan, individual monthly income less than 150 percent of the poverty level amount is defi ned as low income.(23) Th us, we classifi ed these four groups as follows: no income (0), medium low and low income (1-15,840), middle income (15,841-25000), high income (≥25,001).For the survey of factors associated with hospital survival, major interventions, which were composed of three managements: TTM, coronary artery angiography, and ECMO, were identifi ed during the hospital stay.

Statistical Analyses
We used the Kaplan-Meier method to analyze the survival of patients with and without major interventions, and the log-rank test was performed to examine the diff erences in mortality.Cox proportional hazard models were used to compute the hazard ratios (HRs) for mortality of OHCA patients and logistic regression model was applied for risk analysis of ROSC, accompanying 95% confi dence interval (CI) aft er adjustment for the variables mentioned.Two-tailed P<0.05 was considered significant.All of these analyses were conducted using SAS statistical soft ware (version 9.4; SAS Institute, Cary, NC).

RESULTS
Figure 1 illustrates the selection process of study participants.A total of 5,338 OHCA patients met inclusion and exclusion criteria for this study.Of these patients, 12.2% (652) were admitted to hospital and 2.64% (141) survived to hospital discharge.

DISCUSSION
There were two main findings in this study.First, older age (≥50), medium low and low income were associated with a lower rate of hospital survival.Second, major interventions (including TTM, coronary artery angiography, and ECMO) were associated with a higher rate of in-hospital survival.
Consistent with other reports, (24,25) the incidence of OHCA in men was higher than in women, and men were prone to have a lower rate of ROSC than women in our investigation.The possible reasons might be the effect of estrogen and progesterone on reduction of cardiac arrest risk in women and higher prevalence of lifestyle and cardiovascular risk factors in men.(26) However, the effect of gender on hospital survival of OHCA was not significant, as previous studies have described.(24,25) According to a recent retrospective survey in Sweden, there was no association between urbanization and outcomes after OHCA, consistent with our discovery.(36,37) or with no ST elevation on ECG.(38,39) The benefit of ECMO on survival and neurological function in a selected patient group following OHCA had also been documented.(9,10) Though there was a relatively small number of major interventions in our study, patients with TTM, or coronary artery angiography, or ECMO had higher hospital survival, in agreement with the results of many reports.Limitations There were several potential limitations of this study.First, the data was obtained from an administrative database, which has been shown in the past to have inaccuracies when used for cardiac arrest research.(40) To overcome this, procedures and drugs administrated during resuscitation were another key element in data selection in addition to ICD-9 diagnosis code.Only those with attempts in resuscitation were enrolled in the analysis.Second, several unmeasured confounders, which were related to pre-hospital factors and the emergency medical service(EMS) system, including place of occurrence, witnessed collapse, bystander-initiated CPR, EMS response time, initial arrest rhythm, prehospital defibrillation, airway management, and drug administration, were not available in our database.Third, we were unable to document non-sustained ROSC but merely those sustained ROCS with survival to hospital admission or survival to outpatient follow up.Fourth, neurologic functional status was not available in our database; hence, the effect of major interventions on neurologic outcome could not be evaluated.Another concern of our report was that the numbers of patients treated with TTM and angiography were very low.One probable explanation is the treating physician's selection bias for who would receive the treatments.Another probable explanation is that some materials for TTM and angiography were not covered by NHI and were relatively expensive, decreasing patients' willingness to receive the interventions.

CONCLUSION
This population-based study in Taiwan revealed that older age (≥50), medium low and low income were associated with a lower rate of survival.Major interventions, including TTM, coronary angiography, and ECMO were related to better survival.

Figure 1 .
Figure 1.Flow chart of identifi cation of study sample and follow up.CPR, cardiopulmonary resuscitation; ED, emergency department; EMS, emergency medical system; ICD-9, International Classifi cation of Diseases, Ninth Revision; OHCA, out of hospital cardiac arres.

Table 2 (table 3 )
Table1lists the demographic characteristics of all OHCA patients.Twenty-three (0.4%) .Figure2illustrates the results of the Kaplan-Meier method for the cohort.

Table 1 .
Demographic characteristics of OHCA patients

Table 2 .
Cox regression analysis for hospital survival of all OHCA patients

Table 3 .
Logistic regression analysis for return of spontaneous circulation (ROSC)