Heart Attacks and Winter - do the headlines match the data?

This week's news reports announce, "Heart attacks more deadly in winter, but it's not the cold", and "Winter heart attacks not sparked by cold" and "Cold weather doesn't increase the risk of a heart attack". But is that what the study found?

The study was presented at the American Heart Association Scientific Sessions in Los Angeles and you can read the abstract here.  The authors selected 6 states and the city of Los Angeles to examine total and cardiac mortality for the years 2005 through 2008.  They picked warm and cold climates and plotted the normalized cardiac death rates by month.  To their surprise they found similar curves with higher death rates in the winter months and lower rates in the summer months
From: Seasonal Variation in Cardiac Death Rates is Uniform across Different Climates presented November 7 at AHA Scientific Session. Click on image to access source.

From: Seasonal Variation in Cardiac Death Rates is Uniform across Different Climates presented November 7 at AHA Scientific Session. Click on image to access source.

The authors were surprised to find a consistent 26% to 36% increase in cardiac deaths in winter compared with summer and early fall in all of the sites, regardless of whether or not they were warm or cold climates. They hypothesized various other factors that may be related, such as winter flu, holiday stress, depression.  Note that these were presented as hypotheses since the data is not sufficient for evaluating causal relationships.

We were curious about the relationship to weather, since that's what the media picked up on in their reports of the study. Does this study really demonstrate that cold weather doesn't increase cardiac mortality?  So we decided to use some readily available data to plot temperature variations in the locations selected by the authors and for the same years. The National Climatic Data Center of the National Oceanic and Atmospheric Administration (NOAA) gives temperature statistics here. 

We used the site to get monthly temperature readings for the same states (we omitted Los Angeles, since it is not a state) and averaged the temperatures recorded from 2005-2008 each month. We then plotted the averages for the same months as the authors - January, April, July, October, and December as absolute difference from that state's annual average, also averaged over the years 2005-2008.  Take a look at what we found.

Transient

Surprising, no? The relative change in temperature from season to season is very similar for all states, though the absolute temperatures differ.  

So are the media headlines right? Can we be certain that colder weather is unrelated to heart attack mortality? What do you think?