Acute Myocardial Infarction after Laboratory-Confirmed Influenza Infection

Kwong, JC et al. NEJM 2018. [paper]

Why I chose this article

In IM and EM, we are on the front lines of acute epidemics.  This year’s flu season has been a bad one and we are seeing a lot of generic flu but also complications and deaths from the flu. Typical complications that we readily identify are pneumonia, sepsis, ARDS, and viral myocarditis. One complication that I was not aware of is acute myocardial infarction (AMI). As an ED resident this article is helpful for us to keep AMI on the radar when seeing patients who have classic influenza symptoms. As primary care docs it gives more compelling evidence to recommend the flu shot.


Since the 1930’s we have noticed an association between the flu season and an increase in cardiovascular events. Prior studies trying to link AMI and flu have had many limitations. They have used clinical diagnosis of the flu instead of laboratory evidence, used case-control study designs, and they have been underpowered. Before this study the literature showed inconsistent findings.

Research Question

Is there an association between laboratory-confirmed influenza and acute MI?

Study Design

Self-controlled case series. If, like me, you found yourself asking, “What the heck is that?” here’s what it is. This design is perfect for a transient exposure leading to an adverse event. It was developed in the mid 90s to study adverse reactions to vaccines!

  • Compares a control with a risk period
  • Is there a difference in relative incidence of an adverse event during the risk vs. the control period?
  • An individual acts as their own control (so you implicitly control for confounding factors like genetics, location, SES, etc.)
  • Good for very high (everyone gets it) or very low (no one gets it) exposures so that there is a control group


All Ontario residents that met the following criteria:

  • Age ≥ 35 years old
  • Tested for influenza (data obtained from 11 Public Health Ontario labs)
  • And hospitalized for an acute MI (database of all admissions to acute care hospitals)


Study participants were excluded if they had:

  • Tested positive for the flu in the preceding 14 days
  • No AMI hospitalization during observation period
  • An AMI hospitalization within 30 days of a previous AMI hospitalization
  • Flu testing occurred during an AMI hospitalization


They looked at admissions for AMI from 5/1/2008 to 5/31/2015 and testing for respiratory viruses from 5/1/2009 to 5/31/2014:

  • The 7 days after testing positive for a respiratory virus was called the “risk interval”
  • The 52 weeks before and 51 weeks after testing positive was called the “control interval”
  • All this combined time was called the “observation period”

They used a fixed-effects conditional Poisson regression model which accounts for multiple exposures to flu or AMIs during the “observation period”.

They also looked at multiple potential “risk intervals,” (days 1-3; days 4-7; days 8-14; days 15-28), none of which showed an association better than days 1-7.


During the study period 148,307 tests were done for influenza. 19,729 tested flu positive and of those 364 were hospitalized for AMI (332 unique patients). About half of the study population were women. No surprises that a large majority had cardiovascular risk factors including diabetes, hyperlipidemia, and hypertension. Only 31% had received the flu vaccine. 82% were Influenza A positive and 18% were Influenza B positive.

In the risk interval, 7 days after testing flu positive, incidence of admission for AMI was 6 times higher than during the control interval. Influenza B had a 10-fold increase in risk.


Just this week in Urgent Care, I had a 70 yo with new onset exertional angina associated with dyspnea and diaphoresis. Incidentally, he was recovering from the flu which was diagnosed about 1 week ago. An EKG  showed anterolateral T wave inversions (new since 2014). He was admitted directly from Urgent Care and got cardiac enzymes concurrently. Cardiology took him to the cath lab immediately and found undiagnosed multi-vessel coronary artery disease. Although the coronary artery disease had been brewing for years, there was a clear temporal association between his flu and his new onset angina. After his hospitalization he came back to Urgent Care to give me a hug. Which made my life.

It isn’t super surprising that an inflammatory condition such as the flu would cause an increase in cardiac events. This paper provides decent evidence to support our assumptions. It also reminds us to keep that high index of suspicion for AMI as we wade through the multitude of patients with influenza-like illnesses and associated chest pain. We also should be having very specific conversations about what it means to get the flu with our patients who refuse flu vaccination. Especially those with cardiovascular disease or cardiac risk factors. This stuff matters.


Alyssa Kwok is a PGY1 EM/IM resident