Self-Testing for Macular Degeneration Falls Short

Neil Osterweil

September 16, 2014

LONDON — Consumers who want to know their risk of developing age-related macular degeneration can shell out between $99 and $1100 for a commercial DNA test, but at least four tests marketed directly to consumers promise more than they can deliver.

That's the opinion of investigators from the Netherlands, who found that direct-to-consumer genetic tests lack sufficient genetic variants in their predictions and do not account for other, nongenetic variables in their risk prediction methods.

"We know from the literature that in some of these tests, the overall macular degeneration risk can exceed 100%, and that cannot be true," said Caroline Klaver, MD, an ophthalmologist at Erasmus University Medical Center in Rotterdam, the Netherlands.

When it comes to direct-to-consumer tests for macular degeneration, "there is considerable room for improvement," Dr. Klaver said here at the 14th EURETINA Congress.

To determine whether direct-to-consumer genetic tests were capable of predicting risk accurately, Dr. Klaver, a middle-aged male colleague, and one of Dr. Klaver's female PhD students put their own DNA to the test with four products commercially available to European consumers and compared the results against a laboratory-based single-nucleotide polymorphism array, with risk calculation based on a recently published, population-based risk prediction model. They looked at both the risk per genotype and overall risk for macular degeneration for each individual.

One of the tests looked for polymorphisms in 5 genes implicated in degeneration, one looked at 4, and the remaining tests looked only at a single gene associated with risk, CFH.

Dr. Klaver said that when she obtained results from one test (23andMe; Mountain View, California), "I was very happy, because despite my positive family history, they said I had a decreased risk compared to an average population risk of age-related macular degeneration. Unfortunately, when I got the results from the other companies, they said I had an increased risk."

In contrast, for her colleague, a 51-year-old man who had previously smoked but who had no family history of macular degeneration, two tests predicted decreased risk, and two showed increased risk.

Table Direct-to-Consumer Tests for Macular Degeneration

Test Cost DNA Source Genetic Variants
23andMe (USA) $99 Saliva CFH, C2, ARMS2 C3, TIMP3
deCODEme (Iceland) $1100 Buccal smear CFH haplotype, C2(2x), ARMS2, C3
easyDNA (UK) $510 (395 euros) Blood CFH
Genetic Testing Laboratories (USA) $285 Blood CFH


There are currently 34 genetic loci associated with risk for disease, Dr. Klaver said. Using the population-based risk prediction model and plugging in 26 risk variants with the variables of age and sex will yield a positive predictive value of 0.82 for late age-related macular degeneration, and when disease stage, smoking status, and body mass index are added to the mix, the predictive value rises to 0.88.

But by testing only a minimal set of genetic variants, as is done in one of the tests, the predictive value for late degeneration decreased to 0.73, which is lower than that of a nongenetic model incorporating age, sex, stage, smoking, and body mass index (predictive value, 0.78).

In the single-nucleotide polymorphism analysis, the authors saw that the population-based prediction model showed not only the presence of polymorphisms that are deleterious but also those that are protective against disease, whereas the consumer tests looked at fewer protective polymorphisms and overestimated the risk.

Other possible reasons for the inaccuracy of the tests include potentially invalid methodology, but the role is unclear, because some of the companies fail to disclose their modus operandi, Dr. Klaver noted.

The Cloudy Crystal Ball

"We also know that inaccuracy can come from choosing the wrong reference population," she said. "The reference population in the model may not be representative of the individual. Also, these models are all based on case-control studies comparing the extreme cases to extreme controls, and leaving out the middle of the population, which has an intermediate risk."

Dr. Klaver noted that although she is of northern European ancestry, the tests presumed that she was of Tuscan, southern European lineage. "We know that the genetic makeup of Tuscans is different from the genetic makeup of northern Europeans," she added.

An ophthalmologist who was not involved in the study told Medscape Medical News that it is premature to consider using genetic testing for predicting macular degeneration risk. "It's much better to look at the fundus appearance, and if that's okay, then there's no reason to do genetic testing," said Michael Larsen, MD, professor of clinical ophthalmology at the University of Copenhagen, Denmark.

Dr. Larsen said that people who are worried about their risk for macular degeneration can take positive steps to reduce that risk, without the need for expensive and potentially inaccurate tests.

"The rules that apply for health counseling are universal ― everyone should not smoke, everyone should eat leafy vegetables, stay slim, and exercise," he said.

Dr. Larsen moderated the session in which the data were presented. His comoderator, José-Carlos Pastor-Jimeno, MD, director of the Institute of Applied Ophthalmology at the University of Valldolid, Spain, told Medscape Medical News that direct-to-consumer genetic testing in his country prompts patients to put pressure on the national health system for potentially unneeded tests and treatment.

This study was supported by grants from Erasmus Medical Center and both governmental and not-for-profit sources in the Netherlands. Dr. Klaver, Dr. Larsen, and Dr. Pastor-Jimeno reported no relevant financial relatioships.

14th EURETINA Congress. Presented September 11, 2014.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.