This is the first prospective cohort study to compare 30-day and 6-month mortality of suspected cardiac chest pain patients admitted to EDs between two teaching hospitals located in two different cities in southern China with similar ethnicity but differenthealthcare systems [11]. The chest pain patients in GZ had a higher 30-day and 6-month mortalitythan in HK. The findings suggest that the different outcomes were associated with the varying clinical characteristics of patients and different distributions of diagnoses under two differenthealthcare systems. GZ has integrated many features of healthcare system associated with market economies, while its overall economy is largely centrally planned. In contrast,Hong Kong adopt thehealthcare system are come from England and organizational health systems that are commonly seen in centrally planned economies, while its economy functions as a highly capitalisticenterprise [10, 13].
Different distributions of chest pain patients between two hospitals
Over 50% of patients were diagnosed as non-cardiac chest pain in HK, a much higher rate than in GZ (37.1%). This may reflect the different healthcare systems. In Guangzhou, chest pain patients at low risk often present to clinics instead of EDs during office hours. The ED is the only choice for patients with chest pain in HK due to lower medical costs. So, the patients with urgent and severe chest pain/symptoms would likely to present to EDs in GZ. Compared to other studies, the proportion of non-cardiac chest pain patientsin this study was low. The proportions in GZ and HK were 50% and 37.1% respectively compared to 67.2%–70% in other studies [14,15,16]. A lower proportion of non-cardiac chest pain patients may allow earlier discharge from the ED and place less stress on ED and hospital services.
There were no significant differences in 30-day and 6-month mortality for non-ischemic cardiac chest pain and non-cardiac chest pain between the two hospitals. However, there was significantly higher 30-day and 6-month mortality rates for ischemic cardiac chest pain patients in GZ. ACS is a high-risk disease with significant mortality [9, 17]. The current study excluded STEMI patients but NSTEMI and UA patients were recruited. The combined proportion of NSTEMI and UA in GZ was 25.8%, which was higher than thatin HK (17.5%). The proportions of NSTEMI and UA were also higher than those in Cullen’s study (8.7%) whilst non ACS cardiovascular-related conditions were 20.8% [14].
Different clinical characteristics and outcomes of patients between two hospitals
Since the patients with severe chest pain/symptoms would likely to present to EDs in GZ while those patients with mild chest pains/symptoms would visit clinics, so the baseline clinical characteristicsof patients in GZ were worse than those in HK. The patients were older, had lower proportion of NHYA class I and higher creatinine level, higherproportion of ST-segment depression, lower proportion of negative TnT and higher HEART scoreat ED presentation. Those factors have also been shown to be associated with 30-day and 6-month mortality. The same parameters have been verified to predict short- and long-term risks of death from ACS [18,19,20,21,22,23,24]. Compared to HK, lower proportions of patients with risk factors were seen in GZ, whilst more patients with worse NHYA class were observed. This suggests that many chest pain patients in GZ did not have any knowledge of cardiac risk factors and were unaware of the presence of those risk factors, therefore the impact of those risk factors might be underestimated. Early detection and good control of risk factors for suspected cardiac chest pain patients may reduce mortality.
Furthermore, elevated serum creatinine level was an independent predictor for 30-day and 6-month mortality in our study.It has also been reported to be associated with worse outcomes for ACS and critically ill patients [18, 23]. Elevated serum creatinine can reflect vascular damage, renal impairment, endothelial dysfunction and impaired myocardial blood flow [24,25,26,27]. A multinational registry study by Tang et al. showed that SBP could predict the 6-month mortality of all subsets of ACS. As SBP increases by 10 mmHg, the hazard ratio would decrease by 0.95 [27].
Several studies have reported that the HEART score is not only related to the severity of chest pain in EDs but also predicts the occurrence of clinical endpoints [28,29,30,31,32]. Our study also demonstrated that the HEART score was positively associated with 30-day and 6-month mortality of suspected cardiac chest pain patients in the ED setting.
The dissimilarity of healthcare systems in the two cities is another reason for the different ED management. Limited capacity of hospital emergency care services, high out-of-pocket expenses with the need for up-front payment, prolonged discussions with the patient and families for both obtaining consent and pooling funds are likely to be major contributing factors for the long pre-hospital, in-hosptial delay and refused some expensive treatments [13, 32].
In ED chest pain pathways, guidelines recommend that an ECG should be recorded within 10 min and myocardial injury biomarkers should be measured as soon as possible in all suspected cardiac chest pain patients [4, 9]. The ED-ECG time and the proportion of patients having myocardial injury biomarkers measured in both hospitals met these guidelines.Myocardial injury biomarkers, such as troponin, CK and CK-MB are used to identify myocardial injury [33]. Many studies have shown that troponinis superior to CK and CK-MB as a biomarker for detection of myocardial injury [19, 20]. Therefore, troponin testing is recommended in most guidelines for its high sensitivity and specificity [17]. However in GZ, CK and CK-MB are still mainly used for the diagnosis of AMI and ACS due to their lower costs. High expenses with the need for up-front payment of troponin testing are likely to be a major contributing factor for the lower rate of use of troponin testing in GZ. The lower sensitivities and specificities of CK and CK-MB might contribute to the delayed diagnoses of AMI/ACS and later treatment of patients, possibly contributing to the higher early mortality in GZ. Except the mortaliy, readmissions after ACS are associated with higher long-term all-cause mortality [34].However, our study did not show the significant difference of 30-day and 6-month readmission rate in in HK and GZ. The reason was that the patients we enrolled were suspected cardiac chest pain patients instead of ACS.
Strengths and limitations
The strengths of this study include the fact it is the first comparison of outcomes in different healthcare systems in a similarly populated region of southern China, which effectively eliminates many confounding factors in any comparison. The current study also has some limitations, including the sample size of this study which met our precalculated size but may still be considered small by international standards. Secondly, patient risk factors (i.e. the past medical histories) may have been underestimated due to variations in history taking between the two centres. Thirdly, as a prospective observational study, propensity score with matching would attempt to reduce the bias due to confounding variables. However, this method requires large sample size. Also, the procedure only controls for observed variables, so any hidden bias due to latent variables may remain after matching.Logistic regression and Cox regression can meet the need of this study.