Article Text


Predictive DNA testing in ophthalmology: View 2
  1. A R Webster
  1. Moorfields Eye Hospital, University Lecturer, Institute of Ophthalmology, Bath Street, London EC1V 9EL, UK; andrew.webster{at}

    Statistics from

    I would counsel strongly against a “genetic test” for this young girl, even if there were a sensitive and specific genetic test available to determine the genetic status of an individual from this family. I would similarly argue that it would be wrong to perform such a test even if there were a sensitive and specific genetic test available to determine the genetic status of an individual from this family. The problem we face in managing this family is how best to present this view to parents who, as is not unusual, have already decided that a DNA test is exactly what their daughter needs.

    Laboratory testing considerations

    Let‘s imagine for a moment that we have the technology and detective skills to offer genetic testing for this girl. A positive result means she has a 100% chance of carrying the gene causative for RP in the family, and a negative result that there is a 100% chance that she has escaped inheriting the gene causative for RP in this family. This is rarely the case given the vast genetic heterogeneity of this disease. How might this unusual state of affairs have arisen? Well, we might have found that the family is segregating a missense mutation in the rod opsin gene, which would be the first gene investigated in a laboratory for the very reason that it is the most common cause of autosomal dominant RP.1 Rarely are we able to determine the mode of inheritance in a patient presenting with RP, but in this family, given the male to male transmission and three affected individuals, autosomal dominant (AD) RP seems to be on the cards. So, we have previously approached the family members, both affected and unaffected, expounding on the great future benefits of molecular research into these conditions, though we do admit that there are none yet. After such persuasion, most of the members gladly consent to entering our research and donate their DNA.

    So what might happen in the laboratory? Let‘s fantasise for a moment and imagine we have at our disposal a laboratory with limitless molecular technology and a large number of well motivated laboratory technicians. The chief technician notes that ADRP is suspected; he considers the lab‘s strategy for ADRP and finds that screening for small mutations in the gene for rhodopsin is the first thing to do. He sets to work amplifying DNA with the polymerase chain reaction (PCR) from an affected member of the family for the five coding fragments of the gene (a gene of modest size consisting of only five exons). There are many methods that one can use to detect mutations in such fragments of DNA but, in this case, the technician prepares the high throughput wave denaturing high performance liquid chromatography machine (DHPLC for short) and analyses the products from the affected patient. One of these appears to show an abnormal pattern, suggesting a base change. He then determines the exact DNA sequence from this fragment on one of the ABI 3700 semiautomated sequencers at his disposal. Notice how many machines with large names this technician is depending on; molecular testing is indeed an expensive process. On looking at the sequence with some expensive computer software, he discovers not one, but two, DNA changes in the affected person‘s DNA sequence.

    A change in a DNA sequence from the population “norm” does not necessarily indicate one that might cause disease. So he consults his online virtual library. He notices that both changes have been reported in the literature before. He then sequences the same PCR product from the other two affected members, an unaffected spouse and a couple of unaffected individuals from the family who were happy to donate DNA for the research. One of the two changes is only found in the first person analysed and his unaffected mother. That‘s important. The sequence change does not segregate with the disease and so cannot be the change that causes the disease in the family. Instead, it‘s one of those annoyingly common polymorphisms that happen in the population from time to time just to confuse geneticists and clinicians. He then prepares to write a short report to the BJO on how this previously considered disease-causing change exists on a non-segregating chromosome and is therefore a polymorphism. Less interesting from his point of view is the fact that the second change is found in the three affected individuals, not found in either of the two unaffected individuals, and not found in the spouse. It seems to segregate perfectly with the affected members of the family. Is that proof that it causes the disease? He taps out an approximate calculation on his computer and finds the odds of such a distribution in the family being due to true linkage with the disease compared to a chance finding is 16 to 1. So, that‘s good odds that rhodopsin, at least, is responsible for this family‘s disease. But what about that specific change? He consults the literature again to find that the exact same base change has been described in at least four families with ADRP and not detected in at least 400 control chromosomes. Job done! The cause of the disease in this family has been determined. He doesn‘t even need to use the mutiplexed marker panel for other less common ADRP genes (further expense and usage of machines with long names). He reports back to the clinicians that should the change be confirmed in a diagnostic laboratory (remember we‘re in a research lab at the moment) on a second blood sample from an affected family member, then it should be possible to offer a sensitive and specific presymptomatic test on members of the family.

    Clinical considerations

    So, we go back to the clinic. The point of that preceding section was to demonstrate that molecular testing, particularly in a disease such as RP, is expensive, time consuming, and often imperfectly sensitive and specific. But, in order to consider ethical issues in this family‘s case, it is simpler to consider a test that can be offered with 100% sensitivity (no false negatives) and specificity (no false positives). Many ophthalmologists enjoy the fact that we work within an ethically straightforward specialty, but that is not always true, particularly when inherited disease is concerned.2 So now we can offer a robust test to this young girl. But should we? Let‘s assume that we do and that she tests positive. Her parents hear of the result the next month (there‘s an efficient diagnostic laboratory in town too!). They are sad that she has a genetic, potentially blinding disorder, passed down from her father, for which there is no treatment and no means of prevention. In the meantime, they move house, she has a different bedroom which she likes and she‘s no longer afraid of the dark. She no longer requires nappies at night as she has no further wetting of the bed. She goes through school, gets on well, and prepares for her GCSE examinations. At the age of 16, she becomes more aware of her father‘s deteriorating eye condition. This is important to her at this stage in that she is having to consider various career options and wonders whether she too may have a similar problem to her father. She learns that her great uncle also has the disease, and that is was only noticed on having his cataract surgery at the age of 75 years. The great uncle had a genetic test. It is positive. The doctors say that it is not at all unusual for there to be so much variation in RP even between the same members of a family. In fact, the few other families in the world with the same DNA change in rhodopsin have varied greatly in the severity of their disease (other rhodopsin changes are less variable). She discusses the issue with her parents and mentions that, as there‘s still no treatment, she would rather not know whether she has the gene for the disorder until there is a treatment. “Oh we already know you have the mistake in your genes, my dear,” says mum. “We got you tested when you were four.”

    So the basic point is this. The young girl should be investigated for her symptoms. If there is a real suspicion of visual dysfunction, then her acuity should be assessed carefully, she should undergo a refraction and have her eyes examined. She should be treated like any other child presenting to the clinic. If the night blindness is real and disabling (which I doubt), then she could have an electroretinogram, looking for a reduction in amplitude in the Ganzfeld responses. It may be normal, which would be reassuring, but when performed at the age of 4 years would not rule out her later developing RP. If there is the possibility of a presymptomatic DNA test for an untreatable disorder then the patient herself should enter into the decision making when she is old enough to do so. Whose DNA are we testing here?

    However, if a preventive treatment were known or even suspected, then that‘s a different scenario. We are currently doing predictive DNA testing in retinoblastoma.3 The patient herself would stand to gain from having such treatment before the onset of symptoms. She might even avoid the condition altogether despite her imperfect genome. But we‘re not at that stage yet, are we? Maybe one day we will be. Then this dilemma will not exist. We will encourage the parents to consent to their child having a presymptomatic test. We might even save her sight.


    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.