Is eye disease linked to stroke?

Age-related macular degeneration (ARMD), the leading cause of sight loss in the UK, more than doubles the risk of dying from a heart attack or stroke, The Daily Express reports. As their condition worsens, “people become five times more likely to suffer a fatal heart attack and 10 times more at risk of a deadly stroke”, the newspaper says. Some of the sources say that this study raises the possibility that drugs to treat the condition may be to blame, although experts dispute this.

The reports are based on an 11-year study that looked at the links between ARMD and deaths from cardiovascular disease or stroke. Overall, the study found that there was no link. There are important limitations with this study: it did not take into account all known risk factors for cardiovascular disease and stroke during the analysis. On the back of this study alone, people with ARMD should not be concerned that their condition increases the risk of cardiovascular disease. Similarly, the study did not investigate if the treatments for ARMD are linked to risk.

Where did the story come from?

Dr Jennifer Tan and colleagues from the Centre for Vision Research at the University of Sydney, Australia, carried out this research. The study was funded by the Australian National Health and Medical Research Council. It was published in the British Journal of Ophthalmology , a peer-reviewed medical journal.

What kind of scientific study was this?

ARMD is the late stage of age-related maculopathy (ARM), a disorder that affects the macula, which is a small but vital part of the retina at the back of the eye. Cells in the macula deteriorate and die. This affects vision and has severe effects on the quality of life. There are two stages of the disease: early ARMD is diagnosed when there are large, indistinct lesions seen in the macula; late ARMD is classed as either ‘wet’, where new blood vessels leak blood into the retina affecting vision, or ‘dry’, where there are areas of dead cells in the retina. ‘Dry’ ARMD is far more common. It progresses slowly and there is no treatment. ‘Wet’ ARMD progresses more quickly, and may be treatable with drugs such as ranibizumab (Lucentis®), a type of drug called an anti-VEGF.

This study was a prospective cohort study. It looked at any link between the extent of ARMD at the start of the study and cardiovascular or stroke-related deaths over 11 years of follow up. It was carried out as part of the Blue Mountains Eye Study (a study of common eye diseases in Australians aged over 49 years). Between 1992 and 1994, 3,654 people were enrolled. After five years, 2,335 (75%) of them were re-assessed; after 10 years, 1,952 (77%) were re-assessed. At the start of the study and at each re-assessment, photographs were taken of the retina of both eyes to determine the extent of ARMD and whether it could be classed as ‘early’ or ‘late’. The researchers obtained death information from the Australian National Death Index.

In their analysis, the researchers divided people into two groups according to their age: under-75s and over-75s. They also took into account other factors that may affect the risk of dying of cardiovascular disease, i.e., high blood pressure, diabetes, smoking and body mass index (BMI). After excluding people who had a history of heart attack, angina or stroke at the start of the study and those who didn’t have the necessary retinal photographs or who had no data about death available, 2,853 participants were available for analysis.

What were the results of the study?

During the 11 years of follow up, 183 out of the 2,853 (6.4%) people died from cardiovascular causes and 99 out of 2,853 (3.4%) died from stroke. Overall (i.e. in all age groups) there was no link between ARMD at baseline and death from cardiovascular causes or stroke-related deaths (though this was only assessable for early ARMD).

However, analysis by age group showed that under-75s with early ARMD at the start of the study were twice as likely as those without ARMD to die from cardiovascular causes. This result included an adjustment for some other factors that might increase death risk. However, the researchers report that when they adjusted further for serum lipids, white-cell count and fibrinogen levels, the result became non-significant.

Under-75s with late ARMD at the start of the study were about five times more likely to die from cardiovascular causes. However, the confidence intervals were wide (1.35–22.99) and this result did not take into account the other main factors that may have increased risk of death from cardiovascular causes (e.g. weight, smoking, etc.).

As no under-75s with early ARMD had a stroke, this relationship could not be assessed. However, late ARMD at the start of the study was associated with a 10 times greater risk of death due to stroke. Again, however, the confidence intervals were wide (2.39–43.60) and the result did not take into account the other main factors that could be linked to death from stroke.

There was no link among over-75s between baseline ARMD (early or late) and cardiovascular or stroke mortality.

What interpretations did the researchers draw from these results?

The researchers conclude that ARMD predicts stroke and cardiovascular events over the long term in people aged between 49 and 75 years. They say that “this may have potential implications for new intravitreal anti-VEGF AMD therapies”. This suggests that people receiving treatments for ARMD (i.e. drugs known collectively as anti-VEGF for the ‘wet’ kind of ARMD) who are also at high risk of cardiovascular disease may need to be monitored closely. However, the researchers say very clearly that more studies are needed to confirm a link between ARMD and cardiovascular events before such recommendations could be made.

What does the NHS Knowledge Service make of this study?

This study gives very preliminary results and the researchers themselves acknowledge that it does not prove a link between ARMD and cardiovascular events. They call for more research to confirm this. There are some important limitations here:

• Some of the results of the study are not adjusted for other factors that could explain the increase in cardiovascular death risk. Age and lipid levels are particularly strong predictors of this risk and it is not clear how dividing the groups at age 75 years and then adjusting for age but not for lipid levels may have affected the results. In such cases, it is not possible to conclude that there is a link between ARMD and deaths from heart disease or stroke. 
• The results that indicated the largest increase death risk, i.e., 10 times greater risk of death by stroke in under-75s with late ARMD at the start of the study, and five times increased risk of death by cardiovascular causes in the same group, were not adjusted for factors that are strongly linked to these types of death, e.g. smoking, BMI, hypertension and diabetes. As such, they do not offer convincing evidence of any link. The researchers explain that they did not make these adjustments because there were so few people with late ARMD in their sample who had died from these causes (nine who died of stroke and nine who died of cardiovascular causes). A larger study that takes these factors into account is needed.
• The researchers raise the possibility that it is anti-VEGF treatments for AMD that could “increase stroke risk”, but this an extrapolation and is not supported by these results. The researchers did not determine whether people were being treated for their ARMD nor did they distinguish between ‘wet’ and ‘dry’ ARMD. Only ‘wet’ ARMD is treated with anti-VEGF drugs and this is by far the less common of the two types of the disorder. It is known that 90% of people have ‘dry’ (i.e. untreatable) ARMD and 10% have ‘wet’ ARMD (i.e. treatable disease). As only nine out of 2,347 of the under-75s had late ARMD, we might estimate that only one of these had a treatable form of the disease. One person, or even nine, is not enough of a sample on which to base conclusions such as this.
• The researchers are right to call for “more studies confirming a link between ARMD and cardiovascular events”. Until then, people suffering from ARMD should not be alarmed about their cardiovascular risk. Importantly, those receiving treatment should not conclude on the basis of this study that their treatment is responsible for any “increased” risk.

Studies that look at heart disease and stroke deaths need to consider adequately the recognised risk factors for these diseases. As the authors of this study acknowledge, it is possible that differences in the levels of lipids in the blood of these patients at least partly accounts for the increases in risk seen in this study.

Sir Muir Gray adds...

If you analyse data in enough detail you will always find some association, but association does not mean causation. If you are getting treatment for AMRD this study offers no reason to stop.