Recently I was doing research on xanthelasmas. Xanthelasmas are soft, yellow plaques on the eyelids that are made up of cholesterol deposits. An example is shown in the photo above. They are fairly common and completely benign. Other than being cosmetically bothersome, they do not cause any harm.
Some studies have shown they are more common in people with high cholesterol and, more importantly, that they may be an INDEPENDENT marker for heart disease risk. In other words: if two people have the same high level of cholesterol but one of them has xanthelasma and the other doesn’t, the one who has the xanthelasma may be at higher risk for a heart attack than the one who doesn’t. However, the studies on this question are conflicting, with some research showing no particular link between xanthelasma and cardiovascular risk.
I thought I would discuss this study today, not because most of you are affected by xanthelasmas, but rather because it is a good example of how medical literature can take its eye off the proverbial ball of what truly matters, and how crucially important details are often buried deep inside a paper.
In the study, researchers reviewed a database of hundreds of thousands of patients who had come to an ophthalmology clinic for treatment of Presbyopia (far sightedness that tends to come on during middle age for most people). Why choose this database to study? Because it provided a huge number of people (Presbyopia is an extremely common condition) who have been examined by an ophthalmologist, a specialist who is virtually 100% certain to notice during an exam that a patient has xanthelasma and to note that in the patient’s chart.
The researchers then selected out about 18,000 patients who were diagnosed with xanthelasma, and matched them against similar patients (in terms of age, gender, ethnicity, etc.) who had not been diagnosed with xanthelasma during their exam. They excluded people who had already had any heart disease or strokes, and matched the two groups for known risk factors such as diabetes, high blood pressure, and smoking.
So far so good. We now have two groups who do not have any known cardiovascular disease and who – best we can tell – are very alike in every way we can think of except for the fact that one group has xanthelasmas, and the other group does not.
Next the researchers queried the database to find out what happened to all of these patients over the coming one year, five years, and ten years, marking out who had experienced a heart attack, a stroke, or a TIA (that’s Transient Ischemic Attack – basically a “pre-stroke” that causes stroke symptoms but resolves on its own within an hour or so and leaves behind no permanent damage).
Before getting into the results, it’s worth pointing out that this type of study is called a retrospective cohort study, meaning a cohort (group) of patients were analyzed retrospectively (looking backwards after diagnoses and events had already occurred). While these kinds of studies have their place, they are considered to be a weaker form of evidence than a prospective cohort study (in which we start with a group of people and follow them over time, looking for pre-defined parameters) or even better a randomized controlled trial (in which we conduct an actual experiment by randomly splitting people into one group that we do something to and another group to whom we give some kind of placebo, and then watching to see what happens to each group), because retrospective cohort studies are more prone to biases and errors. Ergo, while results of a retrospective cohort study should not be dismissed out of hand, they should always be taken with a grain of salt.
Having said that, what did the scientists find? Alarmingly for me and for other xanthelasma sufferers, they found significantly higher rates of disease in those with xanthelasma. For example, here is a graph from the paper that shows the results at the five year mark. The orange lines are the patients without xanthelasma, and the blue lines the patients with xanthelasma. Each graph shows how likely patients were to NOT experience the event of concern, meaning that the line descending more rapidly over time is concerning because it implies more people in that group have suffered heart attacks, strokes, etc, whereas the higher/flatter line means fewer of these events:
As you can see, the blue line (xanthelasma) patients did worse than the controls (orange line). And indeed, at the end of the paper, the authors wrote:
“Conclusions: Xanthelasma is associated with higher short- and long-term incidence of MACEs [Major Adverse Cardiac Events] up to 10 years, highlighting its value as a clinical marker for cardiocerebrovascular risk stratification and preventive management.”
If you merely read the abstract for the paper on Pubmed, or the cover sheet for the study that’s probably all you’d get from it. And that sounds pretty bad. But there is just one little catch.
Buried deep in the paper, and never further commented on by the authors, was an important detail: the people with xanthelasma had a lower ALL-CAUSE mortality rate than those who did not. In other words, in this paper having xanthelasma led to a higher rate of experiencing cardiovascular disease, but an overall slightlylowerrate of death itself.
Now to be clear, this doesn’t mean that if you have xanthelasma you should be reassured of a long life expectancy. As pointed out earlier, this is a weak type of study, and prone to reaching erroneous conclusions. Moreover, the improvement in all cause mortality was only a tiny one that didn’t quite reach statistical significance. And finally, there may well be biases that account for this difference (an obvious question: were the people with xanthelasma more aggressively treated for cardiovascular risk factors than those without, since doctors tend to associate xanthelasma with increased risk for heart disease, and this in turn led to them having smaller and more survivable heart attacks and strokes)? So, I would not take the mortality data in itself as reassuring.
On the other hand, many studies have shown that those with high cholesterol may have slightly below average rates of death from non-cardiovascular causes like infection and cancer, such that while lowering cholesterol can be heart protective, it’s somewhat of a wash in terms of how much it might actually help you to live longer. It’s possible that we are seeing that same tendency raise its head here, which of course would be a very relevant piece of information for readers to know, but which the authors never comment upon, simply putting the statement in a footnote that “All-cause mortality did not differ between both cohorts at short- and long-term follow-up”.
The data point that probably most of us care about above all others – all-cause mortality, e.g. will this thing help or hurt me in the quest to live a longer time – is frequently buried in papers like this which are more myopically focused on a specific outcome such as cardiac disease. This is something to think about anytime you read a headline stating that this thing or that can reduce your chance of developing disease x. One question you should reflexively ask in response is “yes, but what impact does it have on all-cause mortality?” Avoiding a heart attack by dying of cancer first is not exactly the win we are going for.
