Symptom onset-to-balloon time and mortality in the first seven years after STEMI treated with primary percutaneous coronary intervention
- Daniela Rollando1,
- Enrico Puggioni1,
- Stefano Robotti1,
- Angelo De Lisi1,
- Maura Ferrari Bravo2,
- Adriana Vardanega3,
- Ivo Pattaro1,
- Federica De Benedetti1,
- Michele Brignole1
- 1Department of Cardiology, Ospedali del Tigullio, Lavagna, Italy
- 2Department of Prevention, ASL4 “Chiavarese”, Chiavari, Italy
- 3Department of Information Technology, ASL 4 “Chiavarese”, Chiavari, Italy
- Correspondence to Professor Michele Brignole, Department of Cardiology, Ospedali del Tigullio, Lavagna 16033, Italy;
- Accepted 3 September 2012
- Published Online First 28 September 2012
Objective To evaluate the consequence of treatment delay of primary percutaneous coronary intervention (PPCI) on long-term survival.
Background Network organisation based on early recognition, shortening prehospital time delays and procedural delays is the cornerstone of optimal clinical results in the acute phase of ST-segment elevation myocardial infarction (STEMI). Nevertheless, the evidence of a relationship between symptom onset-to-balloon time and mortality is weak, and few long-term data are available.
Setting and measures In this single-centre observational follow-up study, we evaluated the long-term survival of 790 consecutive STEMI patients (mean age 68±13 years; 73% males) undergoing PPCI≤12 h from symptom onset, or 12–36 h in the case of persistence of symptoms or hemodynamic instability.
Results The median (IQR) treatment delay, defined as the time from symptom onset to reperfusion, was 180 min (120;310), fairly balanced between patient delay (80 min (40;140)) and system delay (80 min (60–114)). Patients with a treatment delay <180 min displayed lower mortality at 1, 3, 5 and 7 years (12%, 17%, 22% and 26%, respectively) than those with a treatment delay >180 min (15%, 24%, 28% and 37%, respectively). The HR was 0.7 (95% CI 0.5 to 0.9). On univariate and stepwise multiple regression analysis, field triage and transportation (p=0.0001), shorter distance from hospital (p=0.02) and male gender (p=0.02), but not clinical variables, were independent predictors of shorter treatment delay.
Conclusions Shorter symptom onset-to-balloon time predicts long-term lower mortality in STEMI patients treated with PPCI. Our findings emphasise the need to minimise any component of treatment delay.
ST-segment elevation myocardial infarction (STEMI) is a major medical emergency, and early mortality is mainly due to the extent of ischaemic injury. In order to reduce immediate mortality, either fibrinolysis or primary percutaneous coronary intervention (PPCI) is performed so as to achieve rapid and complete reperfusion of the infarct-related artery, thereby limiting the size of the infarction. In recent decades, two large systematic reviews have shown that thrombolytic therapy reduces 35-day mortality by 18% in comparison with control subjects (absolute decrease from 11.5% to 9.6%),1 and PPCI determines a further 22% reduction in 4–6-week mortality in comparison with thrombolytic therapy (absolute decrease from 9% to 7%).2 Primary PCI is also associated with a reduction in non-fatal reinfarction (3% vs 7%) and stroke (1% vs 2%) in comparison with thrombolytic therapy during 1 year of follow-up.2
It is well known that the immediate survival benefit of PPCI depends on the treatment delay3 ,4; indeed, network organisation based on early recognition, shortening prehospital time delays and procedural delays is the cornerstone of optimal clinical results in the acute phase of STEMI. Accordingly, it is essential to implement a system aimed at minimising all time delays. The recent guidelines of the European Society of Cardiology (ESC)5 recommend that every effort be made to minimise total delay especially within the first 2 h after onset of symptoms. Nevertheless, the literature evidence of a relationship between symptom onset-to-balloon time and mortality is weak; moreover, registry data show that this time goal is extremely difficult to achieve in clinical practice.6–9 Finally, while the scientific literature mainly focuses on the impact of reperfusion delay on immediate or short-term mortality due to STEMI, few data10 ,11 are available on its long-term effect.
The aim of the present study was to evaluate the effect of total treatment delay on long-term mortality.
We performed a single-centre cross-sectional analysis of consecutive STEMI patients who had undergone PPCI from January 2004 to June 2011. Total mortality data were obtained in November 2011 and were correlated with total treatment delay (see below). We included patients with STEMI (defined as ST segment elevation ≥0.1 mV in at least two contiguous leads, or ≥0.2 mV V1-V3, or presumed new-onset left bundle branch block) who had had a symptom duration of ≤12 h or <36 h in the case of persistence of symptoms or hemodynamic instability. The first index STEMI during the study period was included for further analyses. We excluded STEMI patients in whom PPCI was not attempted, and those in whom the final diagnosis on discharge did not confirm the initial diagnosis. We also excluded patients who had undergone an early invasive procedure but did not meet the criteria for STEMI.
The Italian National Health Service provides tax-supported healthcare for all inhabitants, guaranteeing access to treatment at hospitals and emergency medical service (EMS) transportation. Ospedali del Tigullio is the only hospital in its local healthcare area (called Azienda Sanitaria Locale 4 ‘Chiavarese’), which has an extension of 9272 km and serves a population of 149 892 inhabitants (in the year 2011); 94% of the population lives at a road distance of 0–22 km from the hospital with an average travelling time of 22±12 min, whereas the remaining 6% lives in a rural area at 23–51 km from the hospital with an average travelling time of 54±13 min.
STEMI patients are referred to our catheterisation laboratory from the Emergency Department (ED) of the hospital or through the local EMS field triage and direct transportation (Tigullio Soccorso 118). The Tigullio Soccorso 118 system carries out initial dispatcher triage and provides physician-manned ambulances. All ambulances have equipment for the acquisition and transmission of electrocardiographic data. The catheterisation laboratory is notified when the diagnosis of STEMI is established in the prehospital phase, and patients are admitted directly to the catheterisation laboratory.
The local ethics committee approved the study.
The end point of the study was to evaluate the extent to which the total treatment delay (ie, time from symptom onset to reperfusion) influences long-term mortality. The various delays, from symptom onset to reperfusion therapy, were estimated on the basis of the prehospital and inhospital data registered in the patient's electronic medical records at the time of admission. The ‘patient delay’ was defined as the time from the onset of symptoms to the first medical contact (the ambulance call or arrival at the emergency room). The ‘medical delay’ was defined as the time from the first medical contact to arrival at the catheterisation laboratory. The ‘procedural delay’ was defined as the time from arrival at the catheterisation laboratory to the opening of the culprit vessel. The ‘system delay’ was the sum of the medical delay and the procedural delay and, finally, the ‘total treatment delay’ was the sum of the system delay and the patient delay (figure 1).
Data on total mortality were obtained from the Italian death registry in November 2011.
Continuous data are shown as averages±SDs or medians (25th–75th percentile), as appropriate. The Shapiro-Wilks test was performed to check the skewness of distributions. Absolute and relative frequencies were used to show categorical data. The unpaired Student's t test and the non-parametric Mann-Whitney test were used to compare continuous variables as appropriate. Fisher's exact test was used to compare proportions. The primary analysis of the study was planned as a comparison of the cumulative risk of death between the treatment delay groups by means of a log-rank test. The risk of death was based on the HR obtained by means of the univariate Cox model, using the Breslow method for ties. The stepwise multiple logistic regression technique was used to analyse the causal relationship between total treatment delay and some organisational and clinical variables.
Out of 1000 consecutive patients affected by acute coronary syndrome who had undergone an early invasive procedure in our catheterisation laboratory between January 2004 and June 2011, 806 had an established diagnosis of STEMI and underwent PPCI; of these, 790 had follow-up data and were analysed. During the overall period of study, the annual incidence of PPCI for STEMI was 68 per 100 000 resident inhabitants (total 664 procedures); in addition, 126 (16%) PPCIs were performed in non-resident patients. Patient flow is summarised in figure 2. The clinical characteristics of the study population are summarised in table 1.
Treatment delays are shown in figure 1. The median total treatment delay was 180 min (IQR 120–312), and was used for analysis. In 43 patients (5.4%), PPCI was performed 12–36 h after symptom onset because of persistence of symptoms or hemodynamic instability. Approximately half the treatment delay was due to patient delay (median delay 80 min (IQR 40–140)) and half to system delay (80 min (60–114)).
By November 2011, 171 patients had died. Those who died had a longer median treatment delay than those who survived (210 min (137–360) vs 175 min (120–299)). Patients with a total treatment delay <180 min had an estimated lower product-limit death rate at 1, 3, 5 and 7 years (12%, 17%, 22% and 26%, respectively) than those with a treatment delay >180 min (15%, 24%, 28% and 37%) (p=0.02). The HR was 0.70 (95% CI 0.5 to 0.9) (figure 3). A similar difference persisted even when the 43 patients with a treatment delay >12 h were excluded (HR=0.69 (95% CI 0.5 to 0.9)).
Patients (15% of the total) who had a treatment delay ≤120 min had a lower mortality rate at 1, 3, 5 and 7 years (8%, 13%, 17% and 27%, respectively) than the others (p=0.04) (figure 4).
On univariate analysis, among the 10 variables listed in table 2 (EMS direct transportation to catheterisation laboratory, distance from hospital, gender, age, infarct-related artery, total occlusion of culprit vessel, cardiogenic shock on presentation, Thrombolysis in Myocardial Infarction (TIMI) flow grade III at the end of the procedure and ejection fraction), direct EMS transportation, shorter distance from hospital and male gender were predictive of shorter total treatment delay. On stepwise multiple logistic regression analysis, these three variables remained independent predictors of shorter total treatment delay (p=0.0001, p=0.02 and p=0.01, respectively).
The median total treatment delay was 165 min (120–260) for patients who underwent direct EMS transportation, and 210 min (140–375) for the others. This difference was mainly due to a shorter patient delay (60 vs 120 min, p=0.0001) rather than to a difference in system delay time, which was 80 min in both cases.
The median total treatment delay was 180 min (120–300) for 92% patients who lived in urban areas ≤22 km from the hospital, and 260 min (150–470) for 8% patients who lived in rural areas >22 Km from the hospital. Surprisingly, this difference was mainly due to a shorter patient delay (80 vs 150 min, p=0.01), rather than to a difference in system delay time, which was 80 and 85 min, respectively.
The median total treatment delay was 205 min (135–355) for females and 175 min (120–285) for males. This difference was mainly due to a longer patient delay for females (100 vs 75 min, p=0.003), and not to a difference in system delay time, which was 80 and 85 min, respectively.
During the follow-up, 21 patients underwent a second PPCI because recurrence of STEMI: 11 of these belonged to the group ≤180 min and 10 to the group >180 min.
Our data showed that shorter symptom onset-to-balloon time delay predicted long-term lower mortality in STEMI patients treated with PPCI. The treatment delay was fairly balanced between patient delay and system delay. The benefit of a short treatment delay was maintained, and even increased, up to 7 years after the acute events. On the basis of the HR, a treatment delay <180 min reduced the risk of death at 7 years by 30% compared with a treatment delay >180 min. The minority of patients who had treatment delay ≤120 min had the highest long-term survival benefit. Referral to the ED of the hospital, instead of direct EMS transportation, living in a rural area and female gender, but not clinical variables, were independent predictors of longer treatment delay. Our findings emphasise the need to minimise any component of treatment delay.
Several studies have addressed the time to PPCI and its impact on mortality, but only a few have shown a survival benefit of a short treatment delay (ie, patient plus system delay). The methodology of delay calculation in the present study was similar to that utilised in a population-based study performed in western Denmark.11 In that study, patients were either field-triaged directly to the PCI centre (n=2183) or admitted to a local hospital and then transferred to the PCI centre (n=4026). The median delays of patients field-triaged directly to the PCI centre were similar to those observed in the present study: total treatment delay of 172 min, patient delay of 74 min and system delay of 97 min. In the Danish study, longer times were observed for patients transported from local hospitals (240, 106 and 139 min, respectively). Over a median follow-up of 3.4 years, treatment delay was associated with mortality. In a single-centre registry of 1791 patients,12 the symptom onset-to-balloon time (similar to total treatment delay in the present study) was linearly associated with mortality, and a time >4 h was identified as an independent predictor of 1-year mortality. In a pooled analysis of 22 trials (total 6763 patients), 30-day mortality increased as presentation delay (defined as the time from symptom onset to randomisation) increased, although the difference became robust only after 6 h.4 Other studies found lower inhospital and 30-day mortality rates for shorter door-to-balloon time delay, but not for symptom onset-to-balloon time.8 ,13–15 Finally, some studies failed to find an association between either symptom onset-to-balloon time or door-to-balloon time and mortality after 1 or 6 months of follow-up.9 ,16 In general, while shorter door-to-balloon and system delays are frequently associated with better survival, this association is more difficult to prove when patient delay is considered.
The survival benefit obtained by earlier initiation of reperfusion therapy is difficult to assess in observational studies, as confounding and selection biases may hamper such analyses. Moreover, patient delay and treatment delay depend on the time from the onset of symptoms; these measurements are subject to substantial error, since patients have to recall symptom onset. In addition, the biologically relevant point is the time of onset of infarction, which may not be identical to the time of the first symptoms. In the present study, we were able to show the survival benefit of a short treatment delay. In this study, follow-up was longer than in the literature, in which most studies evaluated short-term mortality (inhospital up to few months). We showed that the benefit of a short treatment delay becomes more evident some years after the PCI procedure (figure 3). This is an original and important clinical finding. The likely explanation is that the immediate revascularisation achieved by PPCI prevents the late development of heart failure, coronary ischaemia and life-threatening arrhythmias which are the most common causes of late mortality in patients who had suffered from acute myocardial infarction.
We identified three factors associated with a longer treatment delay: referral to the ED of the hospital instead of direct EMS transportation, living in a rural area >22 km from hospital and female gender. Indeed, patients admitted to the ED had a significantly longer treatment delay, which was mainly due to a longer time lapse between symptom onset and the first medical contact. Theoretically, even if the longer distance from hospital of patients living in rural areas should prolong the transportation time of about 30 min (see method), this was not observed in our patients who had a similar system delay of those living nearer to the hospital. Surprisingly, the patients living in rural areas showed a longer patient delay (150 vs 80 min). We suppose that in both situations, the patients may have been less aware of their disease or more reluctant to call the EMS. We took advantage from direct access to hospital and short distances (≤22 km in 92% of patients). In the literature, patients living at a longer distance from the PPCI centre and admitted to a local hospital prior to transportation to the PCI centre are less likely to receive PCI,17 and have longer treatment delays and increased mortality.11
The reason why females have a longer delay than males is more uncertain. This finding is in agreement with the recent literature, even though the precise reason for this disparity remains unknown.18 One meta-analysis19 found that women with STEMI had lower odds and a lower rate of presenting with chest pain than men (OR 0.63; risk ratio 0.93). Women were significantly more likely than men to present with fatigue, neck pain, syncope, nausea, right arm pain, dizziness and jaw pain. We can hypothesise that the reason for the longer delay in females is under-recognition of their symptoms. We observed a 30 min longer delay in women than in men. It is well known that mortality due to acute coronary syndrome is higher in women than in men.20–22 Our data suggest that a shorter delay might reduce mortality in women to the same rate as in men.
This was a retrospective study which analysed a long time frame. Although the operative protocol did not change during the study period, it is possible that clinical practice may have changed in some way during the 7 years of observation. For example, the use of thrombus aspiration catheters increased from 8% of patients at the beginning to 23% at the end of the study period. In this study, there were many uncontrolled clinical variables (eg, renal dysfunction or diabetes). Observed similarities or differences in outcomes may therefore be related to differences in baseline characteristics rather than to treatment effects. While the use of aspiration catheter, or the presence of renal dysfunction or diabetes may play a role on long-term survival, it is likely that these variables were fairly distributed in the study cohorts of treatment delay, and therefore, did not impact substantially on our results.
Increasing age is the strong predictor of mortality which has been calculated to increase 0.75% with every year of age.11 The median age of the population of the present study was 5 years higher than that of the western Denmark study11 which was 64 years. Our mortality rate of 21.6% matched well with the 17.3% mortality rate observed in the western Denmark study.
Conclusions and perspectives
Shorter treatment delay predicts long-term lower mortality in STEMI patients treated with PPCI. How can this delay be shortened? Half the delay was due to patient reluctance to call the EMS (figure 1). Patient delay was particularly long in women and in patients living in rural areas. We found a median system delay of 80 min (55 min from the first medical contact to arrival at the catheterisation laboratory, plus 25 min procedural time), which is fairly consistent with the quality performance goals recommended by ESC guidelines5 (ie, <60 min from the first medical contact to the start of PPCI) and by American guidelines23 (ie, <90 min from the first medical contact to opening the culprit vessel). While a further significant reduction in system delay is difficult to achieve, it seems that a considerable reduction in patient delay is more feasible. The patients who died had a treatment delay 35 min longer than those who survived. We calculated that halving the patient delay would have resulted in achieving the goal of a total treatment delay of 120 min in 41% of patients, as against the 15% actually observed. Furthermore, as discussed above, our data suggest that a more generalised use of the EMS organisation would be useful in order to achieve a shorter treatment delay and lower long-term mortality. Thus, it seems that a strategy for increasing the population's awareness of the need to call the EMS promptly after symptom onset, and the utilisation of direct transportation is warranted. Preliminary, uncontrolled, single-community investigations have suggested that educational programmes can reduce the duration of out-of-hospital delay following the onset of symptoms of acute coronary syndromes.24 ,25 Health campaigns on the manifestation of STEMI should continue to promote chest pain as the cardinal symptom. However, they should also encompass a wider spectrum of possible symptoms and highlight potential differences in symptom presentation between men and women.26
Contributors Study design: DR, EP, SR, ADL, MFB, MB. Data collection: DR, EP, SR, ADL, IP, FDB. Statistical analysis: MFB, MB, AV, DR. Writing text: MB, DR. Text discussion and approval: all authors.
Competing interests None.
Ethics approval Ethical committee of ASL4 ‘Chiavarese’, via GB Ghio 8, 16043 Chiavari, Italy.
Provenance and peer review Not commissioned; externally peer reviewed.