Characteristics and Outcomes of Patients With Acute Coronary Syndrome Who Received Percutaneous Coronary Intervention During Snowy Days

Background. Acute coronary syndrome (ACS) is affected by several weather conditions. Studies from different geographical locations have yielded mixed results regarding the outcomes of patients presenting with ACS during snowy days, and we aim to report the Cleveland Clinic experience. Methods. Patients who presented with an ACS and underwent percutaneous coronary intervention (PCI) from July 1, 2009 to September 30, 2017 were divided into ST-segment elevation myocardial infarction (STEMI), and non-ST segment elevation ACS (NSTE-ACS). According to snowy day arrival, we compared in-hospital mortality, culprit lesion anatomy, and door-to-balloon (DTB) time (in STEMI patients). Findings were confirmed in propensity-score matched cohorts. Results. A total of 6878 patients were included: 1608 patients with STEMI (139 snowy-day vs 1469 non-snowy day PCIs) and 5270 NSTE-ACS (419 snowy-day vs 4851 non-snowy day PCIs). Right coronary artery territories accounted for most of the stented culprit lesions in all STEMI and NSTE-ACS snowy-day PCIs. While left anterior descending artery lesions were predominant in NSTE-ACS non-snowy day PCIs. There was no difference in in-hospital mortality between the snowy-day vs non-snowy day groups (4.3% vs 4.5% in the STEMI group [P=.92] and 1.2% vs 1.7% in the NSTE-ACS group [P=41]). In STEMI patients, mean DTB times were similar (43 ± 55.1 minutes vs 46.7 ± 59.6 minutes; P=.61), which remained true after hours, during weekends and holidays. Outcomes were similar in propensity-score matched cohorts. Conclusion. At our institution, snowy days do not seem to affect in-patient mortality. In STEMI patients, DTB times were similar in those who underwent PCI regardless of the snowfall.

J INVASIVE CARDIOL 2021;33(10):E791-E796.

Key words: acute coronary syndrome, outcomes, snow, weather conditions

Several weather conditions can impact the occurrence of myocardial infarction (MI). Pressure gradient, wind activity, and ambient pressure can all increase the incidence of MI.¹ Extremes of atmospheric pressure have been associated with higher MIs and worse outcomes.² MI frequently occurs in the winter season and has been associated with cold weather and sometimes humidity.³ On snowy days, MI is believed to be due to shoveling.⁴ An increase in emergency room presentations for heart failure and cardiac arrest have also been observed during heavy snow days, and might also be attributed to shoveling.⁵ It has been reported that the incidence of MI during the winter season might vary day by day, with higher prevalence on workdays.⁶ Conflicting data exist regarding the outcomes, including in-hospital mortality and door-to-balloon (DTB) time, of MI during blizzards, and varies by geographic location. We aim to report the characteristics and outcomes of patients admitted with acute MI during snowy days at the Cleveland Clinic in Cleveland, Ohio.

All adult patients who underwent percutaneous coronary intervention (PCI) from July 1, 2009 to September 30, 2017 at the Cleveland Clinic were included in a retrospective, observational study that was approved by our institutional review board. We excluded patients who underwent elective procedures. We divided the included patients into ST-segment elevation MI (STEMI) and non-ST segment elevation acute coronary syndrome (NSTE-ACS) groups. Baseline patient characteristics (including age, body mass index, sex, race, comorbidities), coronary artery disease presentation, coronary artery anatomy, DTB time, and in-hospital mortality were provided. Furthermore, the occurrence of snow was obtained from the Climate Data Online in the National Centers for Environmental Information’s website, and was reported as “snow, snow pellets, snow grains, or ice crystals.”⁷

Statistical analysis. Continuous variables are expressed as mean ± standard deviation and were compared using Student’s t-test. Categorical variables are expressed as counts with percentages and were compared using the Chi-square test or Fisher's exact test, as appropriate. A P-value of <.05 was considered to indicate statistical significance. in-hospital mortality and culprit-lesion location and severity were compared between snowy and non-snowy days in both STEMI and NSTE-ACS groups. In STEMI patients, comparison between DTB time in snowy days vs non-snowy days was also performed at all times, and separately in regular working hours and after hours/holidays. To further explore the validity of our findings, and after dividing the data into STEMI and NSTE-ACS patient groups, we performed propensity-score matching analysis between snowy and non-snowy day arrivals. A propensity score was generated for each patient in the standard fashion by performing a logistic regression for snowy-day vs non-snowy day admission using the following variables: age, sex, race, history of smoking, MI, and prior PCI. Patients were matched 1:1 to the nearest neighbor with a propensity caliper of 0.1. Outcomes were then compared in the same fashion between snowy and non-snowy days in both STEMI and NSTE-ACS propensity-score matched patient groups. Statistical analyses were performed using SPSS, version 25.0 (IBM).

We identified 10,642 patients who had received a PCI. We excluded 3764 patients who underwent an elective PCI. A total of 6878 patients were included in the final analysis (Figure 1). The mean age was 65.1 ± 12.3 years, 68.2% were men, and 8.1% (n = 558) patients presented on a snowy day (139 patients [8.6%] presented with STEMI, while 419 patients [8.0%] presented with NSTE-ACS; P=.37). Among STEMI patients, those with a snowy-day arrival had similar age (61.7 years vs 61.2 years; P=.77) and similar number of men (71.9% vs 67.2%; P=.25), were more likely to have a prior MI (45.3% vs 26.7%; P<.001), and a less-frequent history of PCI (7.9% vs 15.8%; P=.01) (Table 1). Meanwhile, among patients who presented with NSTE-ACS, those who presented on a snowy day had similar age (65.6 years vs 66.4 years; P=.10) and number of men (66.8% vs 68.5%; P=.47), and were more likely to be smokers (29.7% vs 22.8%; P<.01), had a prior MI (50.4% vs 41.3%; P<.001), and had a prior PCI (39.4% vs 33.9%; P=.03). The majority of patients had Canadian Cardiovascular Society angina class III in both snowy-day and non-snowy day arrivals across the STEMI and NSTE-ACS groups.

In the STEMI group, snowy-day arrival had the same in-hospital mortality (4.3% vs 4.5%; P=.92), culprit-lesion stenosis severity (97.2 ± 5.5% vs 69.8 ± 6.9%; P=.60), and single-vessel PCI (92.1% vs 91.4%; P=.88) as non-snow day arrival, respectively (Table 2). Right coronary artery territories accounted for the majority of culprit lesions (46.8% snowy-day arrival vs 43.2% non-snowy day arrival). Finally, snow did not affect the DTB time (43 ± 55.1 minutes vs 46.7 ± 59.6 minutes; P=.61), regardless of whether it occurred during working days/hours (50 ± 84.7 minutes vs 48 ± 67.1 minutes; P=.42) or during holidays/weekends/after hours (39 ± 26.3 minutes vs 45.9 ± 53.8 minutes; P>.99) (Figure 2). Similarly, among NSTE-ACS patients, those who presented on a snowy day had the same in-hospital mortality (1.2% vs 1.7%; P=.41) and single-vessel PCI (81.4% vs 81%; P=.83) compared with the non-snowy day NSTE-ACS patients. In the snowy-day group, right coronary artery and left anterior descending coronary artery territories accounted for 34.1% and 33.9% of the culprit lesions, respectively. In the non-snowy day group, left anterior descending artery territories accounted for the majority of culprit lesions (36.1%) followed by right coronary artery territories (33.7%).

To confirm outcomes, propensity-score matching generated 138 pairs of STEMI patients and 419 pairs of NSTE-ACS patients. We then compared the pairs according to snowy-day vs non-snowy day arrival. The two groups had similar baseline characteristics (Supplemental Table S1); standardized differences of covariables between snowy-day and non-snowy day arrival groups before and after matching are shown in Supplemental Figure S1. Both groups had similar in-hospital mortality on snowy vs non-snowy days (4.3% vs 3.6% in the STEMI group [P>.99] and 1.2% vs 1.4% in the NSTE-ACS group [P>.99]), the severity of mean culprit artery stenosis (97.2 ± 5.6% vs 97 ± 6.9% in the STEMI group [P=.30] and 88.7 ± 8.9% vs 88.8 ± 10% in the NSTE-ACS group [P=.50]), or in the percentage of patients who had stents deployed (84.8% vs 87% for single-vessel PCI in the STEMI group [P=.60] and 71.6% vs 76.1% for single-vessel PCI in the NSTE-ACS group [P=.14]).

Furthermore, in STEMI patients, right coronary artery territories accounted for the majority (44%) of the stented culprit lesions, and in 44.9% of snowy-day and 42.0% of non-snowy day arrivals. Meanwhile, in NSTE-ACS patients, the left anterior descending coronary artery territories accounted for 34% of the stented lesions (33.7% in both snowy-day and non-snowy day arrivals), while the right coronary artery territories accounted for 33% of the stented lesions (32.9% in snowy-day and 33.9% in non-snowy day arrivals). Finally, in STEMI patients, the mean DTB time does not seem to be influenced by snow (42.9 ± 55.4 minutes vs 50.6 ± 63.4 minutes; P=.72), regardless of whether it occurred during a workday or during a holiday, weekend, or after-work hours (Supplemental Table S2).

Our study of 6878 patients aimed to determine the outcomes of patients admitted to the Cleveland Clinic with acute MI and had a PCI during snowy days. Our key findings are that in-hospital mortality after PCI in our institution was not significantly different in snowy vs non-snowy days. Also, there was no difference in DTB time between snowy and non-snowy days. This finding remained the same when comparing workday events with those occurring after hours, during weekends, and on holidays.

Studies on the impact of snowy days on outcomes after MI have yielded mixed results. A study done in Quebec on 128,073 patients showed an increase in MI admission and death a day after snowy days. This association was found mainly in men, depended on the quantity and duration of snow, and was attributed to snow shoveling.⁸ Similar high mortality was seen in the winter season in New South Wales, Australia.⁹ Cardiopulmonary death was reported to be higher in extremes of temperature in Northeast Asia.¹⁰ Another study showed no difference in in-hospital mortality between snowy and non-snowy days and no significant disparity in the use of invasive procedures.¹¹ Interestingly, a study done in Montreal, Canada showed lower in-hospital mortality in the winter compared with the summer (0.4% vs 1.3%, respectively; P=.04).¹² At the Cleveland Clinic, we did not find any statistically significant difference in in-hospital mortality between snowy and non-snowy days.

On the other hand, DTB time is thought to be impacted during snowstorms in colder regions due to the difficult commute. Curran et al studied 254 patients who presented with STEMI and had a PCI in Atlantic Canada. Similar to our results, DTB time was not significantly different in snowy vs non-snowy days. In snowy days, 48.1% of patients had a median DTB time of 80 minutes.¹³ Interestingly, in Montreal, the summer season has seen longer DTB times compared with the winter season in men, and was associated with a delay in emergency room presentation after symptom onset.¹²

Multiple studies reported different effects of snow on the incidence of MI. Some have reported higher rates of MI and PCIs during snowy days,¹¹ while others have observed fewer STEMI presentations during heavy snow days compared with mild snowy days.¹³ Cold temperature and shoveling have been previously deemed the cause of the increased MI rate in snowy days. Furthermore, cold temperature, regardless of snowfall, has been also associated with increased MI admissions. This has been seen in China and Norway.^14,15 In Canada, STEMIs have been reported to be more common in extreme temperatures.¹³ Cold weather has been thought to decrease the threshold to myocardial ischemia in patients with coronary artery disease. Meyer et al studied patients with coronary artery disease who underwent exercise stress tests at different ambient temperatures. They found that those who were tested at -20 °C showed ischemic changes on electrocardiography 53 seconds earlier than those who were tested at +20 °C (P<.01).¹⁶ This phenomenon can be explained by a local coronary spasm of atherosclerotic arteries in response to cold. Coronary spasm has been described in the cold pressor test, which consists of placing the patient’s hand in ice water for a minute. Spasm in response to cold has been most pronounced in patients with a history of variant angina¹⁷ and likely associated with endothelial dysfunction.¹⁸ A study in Minnesota did not find an association between cold weather and MI.¹⁹ Shoveling has been associated with an increased risk of MI 1 day after a snowfall.⁸ Multiple studies have shown that it is not only the presence of snowfall that increases the risk of MI, but also the quantity and duration of snow; MI risk increased up to 44% if the snow depth was 1 cm.^8,15 Those events have occurred as well during automated snow removal.²⁰

Study limitations. Our study has certain limitations. It was a retrospective analysis from a single center, which has biases associated with this type of research. Also, the weather database was downloaded from the National Center for Environmental Information and reflected the weather at a nearby weather observing station, which might not accurately reflect the patients' home residences. Furthermore, this registry does not include referred patients who underwent coronary artery bypass grafting without PCI. Also, only in-hospital mortality was studied, with no long-term follow-up.

Patients admitted with ACS who underwent a PCI during snowy days at our institution did not exhibit higher in-hospital mortality compared with those who presented on non-snowy days. Similarly, DTB time was not influenced by snow or work hours. Furthermore, the clinical characteristics of patients presenting with MI on snowy and non-snowy days were similar. Large, multicenter studies are needed to validate our findings.