Rare diseases are often hard to spot. They can evade detection until irreversible organ damage or disability has already set in. Last month, in the hope of preventing just this type of harm, the UK’s health secretary, Wes Streeting, announced a 10-year plan to make genetic testing for hundreds of rare conditions part of standard newborn screening in England. The world is likely to follow, with numerous feasibility programmes already under way, including in the US and Australia. Streeting’s plan is to “leapfrog” disease before it becomes symptomatic. But how scientifically sound is this, exactly?
The genome is a list of letters that feels as if it could be read like a book, but it is a book in a language so new that only a small number of words have been deciphered. And, like any language, even those deciphered words could have multiple meanings. What is known of the risks associated with some gene variants is drawn from decades of studying families at high risk of certain conditions. But we have little experience in population-based genetic testing in low-risk individuals. There is no doubt the type of screening planned will help some children and families, but it could also lead to unnecessary tests and treatment for a great many others.
There is much more to developing most conditions than a single genetic factor. A variant in the HNF4A gene illustrates the problem. People with a strong family history of a rare form of diabetes who carry this variant have a 75 per cent risk of developing the condition. However, the risk of diabetes in a person with this same variant who doesn’t have a family history of the condition is only 10 per cent. We cannot assume that any gene variant will behave the same way in every population. Maybe those families with the HNF4A variant and high rates of diabetes are missing a protective gene that hasn’t been discovered yet. Perhaps there is something in their shared environment that, when combined with their genetic risk, leads to diabetes.
The planned newborn-screening programme assumes that gene variants linked to disease convey a similar high risk in everybody, but that is unlikely to be correct. The work of looking for disease variants in healthy populations has only just started. Until it is complete, we cannot know how many people carry pathological variants that don’t lead to conditions because they are protected by other factors. Do we really want newborns to be the population on whom we test our genetic hypotheses?
That is to say nothing of the ethical issues that will arise from this programme. How does one obtain informed consent from the parents of newborn children when testing for hundreds of conditions simultaneously? In the not-too-distant future, we could have a genetic database of every living person. How will this be protected and used going forward?
Of course, newborn screening is nothing new. The difference here is the huge range of conditions to be screened, the challenge of interpreting the results and the sensitivity of the information gathered. I am concerned parents will feel obliged to accept this testing, but won’t be adequately apprised of all the unknowns. I am concerned that the vital early stages of life will be disrupted by hospital visits that might prove unnecessary. I am concerned that parents and paediatricians will be weighed down by a decision to subject a currently healthy child to potentially invasive tests and treatments.
The sensible thing is to gather more information on the prevalence and behaviour of disease variants in the general population before genetic testing becomes a speculative screening tool in children. While some will benefit, those success stories may turn out to be tiny compared with all the potential harms.
Suzanne O’Sullivan is a neurologist and author of The Age of Diagnosis: Sickness, health and why medicine has gone too far
Topics:
