Objective Studies reporting an association between treatment delay and outcome for patients with ST segment elevation myocardial infarction (STEMI) have generally not included patients treated by a primary percutaneous coronary intervention (PPCI) service that systematically delivers reperfusion therapy to all eligible patients. We set out to determine the association of call-to-balloon (CTB) time with 30-day mortality after PPCI in a contemporary series of patients treated within a national reperfusion service.
Methods We analysed data on 16 907 consecutive patients with STEMI treated by PPCI in England and Wales in 2011 with CTB time of ≤6 hours.
Results The median CTB and door-to-balloon times were 111 and 41 min, respectively, with 80.9% of patients treated within 150 min of the call for help. An out-of-hours call time (58.2% of patients) was associated with a 10 min increase in CTB time, whereas inter-hospital transfer for PPCI (18.5% of patients) was associated with a 49 min increase in CTB time. CTB time was independently associated with 30-day mortality (p<0.0001) with a HR of 1.95 (95% CI 1.54 to 2.47) for a CTB time of >180–240 min compared with ≤90 min. The relationship between CTB time and 30-day mortality was influenced by patient risk profile with a greater absolute impact of increasing CTB time on mortality in high-risk patients.
Conclusion CTB time is a useful metric to assess the overall performance of a PPCI service. Delays to reperfusion remain important even in the era of organised national PPCI services with rapid treatment times and efforts should continue to minimise treatment delays.
Statistics from Altmetric.com
National and international guidelines recommend that reperfusion therapy (fibrinolysis or primary percutaneous coronary intervention (PPCI)) for patients with acute ST segment elevation myocardial infarction (STEMI) is delivered as quickly as possible to restore coronary artery blood flow and limit myocardial necrosis.1–3 In the UK, the National Infarct Angioplasty Project (NIAP) recommended PPCI as the treatment of choice for patients with acute STEMI, provided it can be delivered in a timely fashion.4 Following publication of the NIAP report in 2008, there has been a rapid roll-out of PPCI services across the UK and in 2011 over 95% of STEMI cases treated by reperfusion therapy underwent PPCI.5
The time interval between onset of symptoms due to STEMI and restoration of coronary blood flow comprises the ‘patient delay’ from onset of symptoms to the call for help and the ‘system delay’ from the call for help to the delivery of reperfusion therapy. Several studies suggest that door-to-balloon (DTB) time (time from arrival in hospital to therapeutic instrumentation of the culprit coronary artery) correlates with in-hospital6–9 and longer-term mortality8 ,10 ,11 but this has not been confirmed in all studies.12–17 By contrast, call-to-balloon (CTB) time (time from the call for help to the therapeutic instrumentation of the culprit coronary artery) may be influenced by the modifiable elements of the ‘system delay’.18 CTB time includes the time from the call for help to the arrival of emergency services through to delivery of PPCI in hospital. Hence, CTB time may provide a more reliable assessment of the overall performance of a reperfusion service.
The British Cardiovascular Intervention Society (BCIS) database records data from every percutaneous coronary intervention (PCI) procedure in the UK, including patient demographics, clinical characteristics, procedural details and outcome data.19 Definitions for these variables are published in the BCIS PCI dataset.20
We analysed the BCIS database to determine factors associated with CTB time and 30-day mortality in patients with acute STEMI in England and Wales treated by PPCI during 2011, 3 years after introduction of the national PPCI programme.
The CTB time was derived from the time of the call for help to the balloon time, which are recorded for each patient. The time of the call for help was defined as the time of the call for an ambulance for patients presenting in the community or as the time of arrival at hospital of patients presenting directly to a hospital emergency department. For patients initially admitted to a hospital without PPCI capability, the call for help was defined by the first medical contact (FMC) so that the CTB time included the time for transfer to a PPCI-capable hospital. The balloon time was defined as the time of use of the first device in the coronary artery.
Patients were stratified according to their admission route: STEMI in the community admitted directly to a PPCI hospital (direct), STEMI in the community admitted to a non-PPCI hospital and then transferred to a PPCI hospital (transfer), STEMI while an inpatient in a non-PPCI hospital and then transferred to a PPCI hospital (non-PPCI IP) and STEMI while already an inpatient in a PPCI hospital (PPCI IP). Patients were also stratified into groups presenting ‘in-hours’ (weekday 08:00–18:00 hours) or ‘out of hours’ (weekday 18:00–08:00 hours and weekends).
The National Health Service (NHS) number was used to facilitate mortality tracking via the Office of National Statistics (ONS) in England and Wales. All-cause mortality at 30 days was determined from deaths reported to ONS, a statutory requirement in England and Wales.
Univariable associations of baseline, procedural and timing variables with CTB time were assessed using linear regression. Variables independently associated with CTB time were identified using a manual forward stepwise approach with a p value of 0.01 as the criterion for inclusion in the final multivariable model. A manual approach was used rather than an automated stepwise procedure to allow for clinical judgement. The linear relationship of quantitative variables was considered and variables were grouped into appropriate categories as necessary.
To investigate the impact of CTB times and other factors on 30-day mortality, univariable and multivariable Cox models were fitted using a manual forward stepwise approach, always including gender and CTB time. CTB time was not assumed to be linearly related to 30-day mortality across all times and hence was categorised into time intervals to reflect clinically relevant groups while ensuring sufficient numbers of patients in each group.
To estimate the impact of admission route, calls made ‘out of hours’, centre procedural volume and the components of symptom to balloon time (symptom to call, call to door and DTB time), the Cox model was extended to include these variables, excluding CTB time as appropriate.
Data were missing for up to 14% of any one variable included in the models. In order to adjust for missing data, multiple imputation techniques with chained equations were used, assuming missingness is at random, with 10 imputations using all covariates in the model.21
To examine the relationship between CTB time and 30-day mortality for patients at different underlying risk, a multivariable model was created fitting CTB time as a continuous variable. The coefficients for this model were used to predict a patient's 30-day mortality risk assuming a CTB time of 60 min and patients at the 25th, 50th and 75th centiles of risk were identified. By applying the coefficient for CTB time in the model to these risks, the predicted 30-day mortality at each of these centiles could be determined for any CTB time up to 360 min. The predicted 30-day mortality was plotted against CTB time to illustrate the association between CTB time and mortality risk at different levels of underlying risk.
All analyses were carried out with STATA (V.13.1; StataCorp, College Station, Texas, USA).
In 2011, the BCIS database recorded 79 433 PCI procedures in England and Wales including 19 534 procedures for patients with acute STEMI. We excluded patients with STEMI who were treated by fibrinolysis and who subsequently underwent ‘rescue’ PCI, duplicate or repeat PPCI, missing call or balloon times and whose CTB times are >6 hours. Our study cohort comprised 16 907 patients with STEMI treated by PPCI (figure 1).
The baseline demographic, clinical, procedural and timing characteristics are shown in table 1. The mean CTB time was 121 min (median 111 min, IQR 90–139 min) and 80.9% of patients were treated within 150 min of the call for help (figure 2).
The mean DTB time for patients in the ‘direct’ cohort was 51 min (median 42 min, IQR 29–62 min). For patients in the ‘transfer’ cohort, the mean DTB time from first hospital admission was 133 min (median 123 min, IQR 95–161 min), but the mean DTB time from arrival at the PPCI centre was 39 min (median 34 min, IQR 25–47 min).
At 30 days 898 patients (5.3%) had died, with 331 (2.0%) and 611 (3.6%) deaths at 24 hours and 7 days, respectively. Mortality rates at 30 days for patients with CTB times ≤90 min, >90–150 min and >150 min were 3.5%, 4.8% and 9.4%, respectively (figure 3).
Factors associated with CTB time
CTB time was independently associated with age and was 13 min longer for patients aged over 80 than for those aged under 55 (table 2). CTB time was 4 min longer in female patients and 5 min in those with diabetes. A requirement for mechanical ventilation before PPCI was associated with the greatest absolute effect on CTB with an increase of 33 min.
CTB time was 10 min longer for patients presenting ‘out of hours’ than for patients presenting during normal working hours (‘in-hours’). Patients who were initially admitted to a hospital without PPCI capability and who were then transferred to a PPCI centre had 49 min longer mean CTB times than patients admitted directly from the community to a PPCI service.
CTB time was 9 min shorter in centres doing fewer than 100 PPCI procedures compared with those doing 200 or more. Symptom to call time was not significantly associated with CTB time.
Factors associated with 30-day mortality
In a univariable analysis, each 30 min increase in CTB time increased the hazard of 30-day mortality by around one-fifth (table 3). In multivariable analysis, there was strong evidence of an independent association between CTB time and 30-day mortality (trend test p<0.0001).
Advancing age was independently associated with a progressive increase in 30-day mortality such that patients aged over 80 were at over threefold higher risk of 30-day mortality than patients under the age of 55. In a univariable analysis, female gender was associated with higher 30-day mortality but this is explained by women presenting at older ages than men (average age at presentation 69 vs 62 years, respectively) and, after multivariable analysis, gender was not associated with 30-day mortality.
Diabetes, peripheral vascular disease, renal disease, Q waves on the presenting ECG, multi-vessel PCI and femoral artery access were all independently associated with an increased risk of 30-day mortality. Cardiogenic shock and ventilation before PPCI had the greatest absolute independent effect on the hazard of 30-day mortality (table 3).
The 49 min increase in CTB time associated with inter-hospital transfer was associated with an observed 18% increase in the hazard of 30-day mortality (95% CI 0.99 to 1.39) but the evidence was not strong (p=0.059) (table 4). The 10 min increase in CTB time associated with an ‘out of hours’ call time was not associated with an increase in 30-day mortality.
The relationship between CTB time and predicted 30-day mortality at different levels of individual patient risk is shown in figure 4. For a patient at low risk (25th centile of risk), an increase in CTB time from 60 to 360 min is predicted to increase 30-day mortality by <1%, whereas such a treatment delay among higher risk patients (75th centile of risk) is predicted to increase 30-day mortality by nearly 3% (figure 4).
We analysed data from a large contemporary series of patients with acute STEMI who were treated by a coordinated national PPCI service and identified several variables that were independently associated with CTB time and 30-day mortality. Some of these factors are unalterable (age, gender) but some are potentially modifiable (CTB time), suggesting the scope to further improve the outcomes of PPCI.
Previous studies have generally reported a positive association between treatment delay and mortality after PPCI for STEMI, but inconsistencies in the evidence base may reflect differences in study design, data collection periods and definitions of treatment delay and outcomes. Early studies from the USA reported median DTB times of over 100 min6 ,9 and median symptom-to-balloon (STB) times of 234 min.6 Recent studies have reported DTB times approaching 60 min but these improvements in STEMI care have not consistently been associated with improvements in outcome,15–17 possibly because of the use of population rather than individual-level treatment delays in some studies.22
DTB time may be a poor indicator of the overall impact of treatment delay on outcome, as it does not include pre-hospital care. The STB time is a measure of total ischaemic time, but symptom onset may be difficult to define accurately because of recall bias, prodromal anginal symptoms and silent or atypical presentations. By contrast, CTB time is derived from two easily measured time points (time of call to emergency services or self-presentation at hospital and time of therapeutic instrumentation of the culprit coronary artery), which are systematically recorded for all patients treated by PPCI in the UK. CTB time includes the potentially modifiable components of the system delay and may provide the best measure of the overall performance of a PPCI service.
In a previous study of 13 790 patients with STEMI, who were enrolled in the Swedish Coronary Angiography and Angioplasty Registry from 2003 to 2008, the delay from FMC to PPCI was associated with 1-year mortality and severe left ventricular systolic dysfunction at discharge from hospital.23 The time of FMC, however, was defined by the time of the first ECG, which was recorded before arrival at the hospital in only 51% of patients. Moreover, the time of PPCI was determined by the start of the coronary arteriogram, rather than therapeutic instrumentation of the coronary artery. This may explain why the median time from FMC to PPCI in this study was only 70 min. Another study reported on 6209 patients treated by PPCI from 2002 to 2008 at three centres in Western Denmark. CTB time was defined as the time from FMC with the emergency medical services to insertion of the guiding catheter before PPCI. CTB times following direct admission to the PPCI centre and following transfer from a local hospital were 97 and 139 min, respectively, and CTB time was independently associated with mortality at a median of 3.4 years.11 In our study, symptom-to-call time was not associated with 30-day mortality, whereas call-to-door and DTB times were (table 4), thus suggesting that pre-hospital and hospital-based emergency care are equally important contributors to patient outcome.
In our study, the relationship between CTB time and mortality was influenced by patient risk profile, such that an increase in CTB time was predicted to have substantially greater impact on 30-day mortality among patients at high risk relative to those at low risk. Several variables associated with CTB time were also associated with 30-day mortality, suggesting that patients most likely to experience delays to treatment may also be most likely to benefit from efforts to minimise treatment delay.
The results of PPCI outside of normal working hours have been studied extensively. A meta-analysis of data from 1.9 million patients from 36 studies reported that ‘out of hours’ PPCI was associated with a 14.8 min increase in DTB time and a 12% increase in the odds of in-hospital and 30-day mortality.24 Three large registries also reported that out-of-hours PPCI was associated with longer DTB times but these treatment delays had no impact on in-hospital mortality.25–27
In our study, an ‘out of hours’ call time was associated with an increase in CTB time of only 10 min, which did not translate into an increase in 30-day mortality. These data suggest that the increase in CTB time associated with ‘out of hours’ procedures was insufficient to impact 30-day mortality and support current models of PPCI service delivery across England and Wales.
The shorter CTB times in ‘low volume’ centres (<100 PPCI procedures) may reflect opportunistic intervention in centres that do not provide a 24/7 service. Overall, centre volume had no impact on 30-day mortality but these data require cautious interpretation because it has limited statistical power.
In England and Wales, a minority of patients with STEMI self-present to a non-PPCI hospital or are initially taken to the nearest emergency department because of diagnostic uncertainty. These patients experience a 49 min increase in CTB time, which is associated with a 18% increase in the hazard of 30-day mortality. These data support the need for preferential transfer of patients with suspected STEMI directly to a hospital with PPCI capability to minimise delays to treatment.
In England and Wales, over 80% of patients meet the national audit standard of a CTB time of <150 min. Recent guidance from the National Institute for Health and Care Excellence (NICE) recommended that PPCI should be the preferred reperfusion strategy, provided that PPCI can be delivered within 120 min of the time at which fibrinolysis could be given.28 For most patients the CTB time includes the time from the call for help to the arrival of an ambulance, the time to make a diagnosis and the time that would have been required to set up a fibrinolytic infusion. Cumulatively these delays are likely to exceed 30 min; hence the majority of patients in this study would have been compliant with the NICE guidelines. Long CTB times may be unavoidable in patients who present in geographically remote areas, where ambulance transfer is delayed or where there is diagnostic uncertainty at the time of presentation. Future research should focus on this subgroup and if CTB times cannot be shortened pharmaco-invasive treatment may be an alternative strategy.29 ,30
We assessed the impact of patient-specific CTB times on 30-day mortality in a large cohort of patients undergoing PPCI in a single year. Substantial variation in observed and unobserved factors during this period is unlikely but residual confounding by factors not included in the multivariable analyses cannot be excluded.
Our data suggest that CTB time is a useful metric to evaluate the performance of PPCI services and has advantages over DTB time. Nevertheless, CTB time may not be a reliable surrogate for total ischaemic time because of variation in the duration and severity of symptoms before the person calls for help and because the time of the first therapeutic intervention on the occluded coronary artery may not completely reperfuse the affected myocardium.
We report outcome at 30 days and most deaths after PPCI occur within this time window but longer-term follow-up data may provide additional information. In addition, our study only included patients treated by PPCI and is not applicable to patients with STEMI who are managed with fibrinolysis or do not receive reperfusion therapy, who may have worse outcomes.5 ,31
In this contemporary study of patients treated by PPCI within an established national system of STEMI care, there was a strong independent association between CTB time and 30-day mortality. This relationship was influenced by patient risk profile and in high-risk patients an increase in CTB time of 90 min would be expected to increase 30-day mortality by around 1%.
Approximately 20% of patients have CTB times longer than 150 min and further research is required to determine the causes of treatment delay in this subgroup. Efforts to improve performance of PPCI services should ensure that all patients with STEMI are preferentially directed to a hospital with PPCI capability so that the route of entry into the healthcare system does not influence outcomes.
What is already known on this subject?
Primary percutaneous coronary intervention (PPCI) is the treatment of choice for patients presenting with acute ST segment elevation myocardial infarction (STEMI), provided it can be delivered in a timely fashion. Delays to treatment are associated with outcome and most efforts have focused on improvements in door-to-balloon time.
What might this study add?
We have demonstrated a strong independent association between call-to-balloon time (CTB) and 30-day mortality in an organised national PPCI service with short treatment delays. This relationship is influenced by the patient's baseline risk profile such that those at highest risk are likely to benefit the most from reducing treatment delays.
How might this impact on clinical practice?
CTB time may more reliably assess the overall performance of a PPCI service as it includes the potentially modifiable components of both pre-hospital and hospital emergency care.
Contributors All listed authors fulfil the four authorship criteria as specified in the guidelines of the International Committee of Medical Journal Editors 2013. No persons other than the listed authors have made substantial contributions to this manuscript. RWV and RAH devised and planned the study; TCC performed all the statistical analyses; RWV, RAH and TCC wrote the manuscript; MAdB, HHG and PFL reviewed the manuscript and the statistical analyses and made changes to the content of the manuscript; RWV is responsible for the overall content of the manuscript and acts as guarantor.
Competing interests TCC has received research funding from the Medicines Company. MAdB has received travel grants from Abbott Vascular. RAH is an Advisory Board member for Quantum Imaging. RWV, HHG and PFL have no relationships with industry to declare.
Ethics approval National registry data.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement RWV and TCC had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.