Tuesday, May 12, 2009

The Coronary Heart Disease Epidemic

Few people alive today are old enough to remember the beginning of the coronary heart disease (CHD) epidemic in the 1920s and 1930s, when physicians in the U.S. and U.K. began sounding alarm bells that an uncommon disease was rapidly becoming the leading cause of death. By the 1950s, their predictions had come true. A decade later, a new generation of physicians replaced their predecessors and began to doubt that heart attacks had ever been rare. Gradually, the idea that the disease was once uncommon faded from the public consciousness, and heart attacks were seen as an eternal plague of humankind, avoided only by dying of something else first.

According to U.S. National Vital Statistics records beginning in 1900, CHD was rarely given as the cause of death by physicians until after 1930. The following graph is from The Great Cholesterol Con, by Anthony Colpo, which I highly recommend.


The relevant line for CHD deaths begins in the lower left-hand part of the graph. Other types of heart disease, such as heart failure due to cardiomyopathy, were fairly common and well recognized at the time. These data are highly susceptible to bias because they depend on the physician's perception of the cause of death, and are not adjusted for the mean age of the population. In other words, if a diagnosis of CHD wasn't "popular" in 1920, its prevalence could have been underestimated. The invention of new technologies such as the electrocardiogram facilitated diagnosis. Changes in diagnostic criteria also affected the data; you can see them as discontinuities in 1948, 1968 and 1979. For these reasons, the trend above isn't a serious challenge to the idea that CHD has always been a common cause of death in humans who reach a certain age.

This idea was weakened in 1951 with the publication of a paper in the Lancet medical journal titled "Recent History of Coronary Disease", by Dr. Jerry N. Morris. Dr. Morris sifted through the autopsy records of London Hospital and recorded the frequency of coronary thrombosis (artery blockage in the heart) and myocardial infarction (MI; loss of oxygen to the heart muscle) over the period 1907-1949. MI is the technical term for a heart attack, and it can be caused by coronary thrombosis. Europe has a long history of autopsy study, and London Hospital had a long-standing policy of routine autopsies during which they kept detailed records of the state of the heart and coronary arteries. Here's what he found:

The dashed line is the relevant one. This is a massive increase in the prevalence of CHD death that cannot be explained by changes in average lifespan. Although the average lifespan increased considerably over that time period, most of the increase was due to reduced infant mortality. The graph only includes autopsies performed on people 35-70 years old. Life expectancy at age 35 changed by less than 10 years over the same time period. The other possible source of bias is in the diagnosis. Physicians may have been less likely to search for signs of MI when the diagnosis was not "popular". Morris addresses this in the paper:
The first possibility, of course, is that the increase is not real but merely reflects better post-mortem diagnosis. This is an unlikely explanation. There is abundant evidence throughout the forty years that the department was fully aware of the relation of infarction to thrombosis, of myocardial fibrosis to gradual occlusion, and of the topical pathology of ostial stenosis and infarction from embolism, as indeed were many pathologists last century... But what makes figures like these important is that, unlike other series of this kind, they are based on the routine examination at necropsy of the myocardium and of the coronary arteries over the whole period. Moreover Prof. H. M. Turnbull, director of the department, was making a special case of atheroma and arterial disease in general during 1907-1914 (Turnbull 1915). The possibility that cases were overlooked is therefore small, and the earlier material is as likely to be reliable as the later.
Dr. Morris's study was followed by another similar one published in 1985 in the journal Medical History, titled "Ischaemic Heart Disease, Aortic Aneurysms, and Atherosclerosis in the City of London, 1868-1982", conducted by Dr. Robert Finlayson. This study, in my opinion, is the coup de grace. Finlayson systematically scrutinized autopsy reports from St. Bartholemew's hospital, which had conducted routine and detailed cardiac autopsies since 1868, and applied modern diagnostic criteria to the records. He also compared the records from St. Bartholemew's to those from the city mortuary. Here's what he found:

The solid line is MI mortality. Striking, isn't it? The other lines are tobacco and cigarette consumption. These data are not age-adjusted, but if you look at the raw data tables provided in the paper, some of which are grouped by age, it's clear that average lifespan doesn't explain much of the change. Heart attacks are largely an occurrence of the last 80 years, and were almost totally unknown before the turn of the 20th century.

What caused the epidemic? Both Drs. Morris and Finlayson also collected data on the prevalence of atherosclerosis (plaques in the arteries) over the same time period. Dr. Morris concluded that the prevalence of severe atherosclerosis had decreased by about 50% (although mild atherosclerosis such as fatty streaks had increased), while Dr. Finlayson found that it had remained approximately the same:


He found the same trend in females. This casts doubt on the idea that coronary atherosclerosis is sufficient in and of itself to cause heart attacks, although modern studies have found a strong association between advanced atherosclerosis and the risk of heart attack on an individual level. Heart attacks are caused by several factors, one of which is atherosclerosis.  Atherosclerosis can be caused by infectious disease, so this may explain Dr. Morris's finding that it has decreased since the beginning of the 20th century.

What changes in diet and lifestyle associated with the explosion of MI in the U.K. and U.S. after 1920? Dr. Finlayson has given us a hint in the graph above: cigarette consumption increased dramatically over the same time period, and closely paralleled MI mortality. Smoking cigarettes is very strongly associated with heart attacks in observational studies. Animal studies also support the theory. While I believe cigarettes are an aggravating factor, I do not believe they are the main cause of the MI epidemic. Dr. Finlayson touched on a few other factors in the text of the paper, and of course I have my own two cents to add. I'll discuss that next time.