James Cook and Olivia Seifert
Background
Fabry disease is a rare genetic X-linked disease caused by pathogenic genetic variants in the GLA gene, which can cause fat molecules called sphingolipids to accumulate to dangerous levels within bodily organs and eventually lead to life-threatening complications such as stroke and kidney failure. Fabry disease symptoms can present later in life – in the “late-onset” form of the disease – or from early childhood – in what is known as “classical” Fabry disease. As with other rare genetic conditions, of which there are thought to be over 7000, accurately determining the prevalence of Fabry disease is challenging.
The difficulty in determining how rare is ‘rare’
Estimates for rare genetic disease prevalence are often based on reported patient numbers compiled from literature reviews, data registries and surveillance programs. However, not all patients will be accurately diagnosed, or diagnosed at all, in part because the genetic and phenotypic complexities of rare genetic diseases mean that their symptoms are not always recognised by medical professionals.
As such, relying solely on reported patient numbers impacts the accuracy of prevalence estimates for rare diseases, and likely leads to an underestimation of their true prevalence.
For Fabry disease, epidemiological studies thus far have widely varied in their estimation of the disease’s prevalence: estimates based on reported clinical cases range from 1 in 40,000 to 1 in 170,000 individuals [1,2], whilst recent estimates based on newborn screening are much higher, ranging from 1 in 1,250 to 1 in 21,973 individuals [3-6].
Determining robust prevalence estimates using genetic database mining methodology
By analysing large genetic datasets that contain the genetic profiles of thousands of individuals from diverse ancestries, it is possible to determine how many individuals carry the relevant genetic variants associated with a particular rare disease of interest.
These figures can then be extrapolated to determine how many individuals in other populations of interest would also be expected to carry these genetic variants to produce robust prevalence estimates that can complement and refine existing estimates based off reported patient numbers.
In this study on Fabry disease, HealthLumen analysed the prevalence of 8 causal genetic variants mostly associated with the late-onset form of the disease in the gnomAD v4.1 genetic database, based on the inclusion of these variants in a previous Fabry disease study using data from the UK Biobank [7].
These figures were then mapped to the US population using data from the US Census Bureau and by using the Hardy-Weinberg equation, taking into consideration Fabry’s X-linked mode of inheritance, to produce estimates of the number of people carrying these 8 Fabry genetic variants within the US population – in both men and in women, and across different ethnic groups.
By applying penetrance figures – the likelihood that a carrier will actually present symptoms of the disease at some point within their lifetime – obtained from the literature, the number of symptomatic individuals in the US population was then calculated.
Results: Fabry disease prevalence in the US considerably exceeds previous estimates
The study estimated that the carrier and symptomatic populations of Fabry disease in the US in 2024 are: 12,024 male carriers (or 1 in 14,022 males) who will all develop symptoms, and 24,845 female carriers (or 1 in 6,978 females), of whom 17,392 will develop symptoms.
Individuals within the European (non-Finnish) ancestry group, followed by the South Asian, East Asian and African/African American ancestry groups, were found to have the highest number of carriers of the 8 pathogenic Fabry variants studied.
To our knowledge, this is the first study to estimate the number of Fabry disease carriers and symptomatic individuals in the US using publicly available gnomAD v4.1 data. By analysing just 8 of the potentially hundreds of causal variants within the GLA gene, this study suggests that Fabry disease may be over 3 times as prevalent as is currently believed by the US National Institute of Health, and has a prevalence in line with that suggested by newborn screening studies and the UK Biobank analysis.
This study reinforces the growing consensus that estimates of rare disease patient populations are likely underestimated. It also underscores the immense potential of large genetic databases to generate more accurate prevalence estimates. As these datasets expand, incorporating more data and increasing their power and diversity, their ability to provide robust insights will continue to improve.
Implications for pharma and healthcare stakeholders
This study suggests that there is a potentially large unrecognised and underserved Fabry patient population, uncovering a significant therapeutic market opportunity for those developing treatments for the disease.
Genetic database analysis offers a novel solution to address such data gaps, providing accurate insights to help pharma and biotech companies reach previously undiagnosed and untreated individuals, mitigate risk, and create cost effective drug development strategies – including through the optimisation of clinical trial design and resource allocation.
Moreover, if Fabry disease is more common than currently believed, this would suggest its diagnosis rate is even lower than current estimates suggest – raising urgent questions about how healthcare systems can streamline diagnostic pathways. Additionally, a larger Fabry patient population would indicate a greater economic burden of undiagnosed and untreated disease, which could support the case for increased funding for diagnosis and treatment. To inform these discussions, modelling techniques such as microsimulation modelling can help quantify the health and economic impact of new diagnostic and therapeutic interventions, providing data-driven insights to those lobbying for change.
To discuss the results of this study and how these methodologies could be applied to other rare diseases, book in a meeting with us today.
HealthLumen have recently published this work in medRxiv. Read the full publication here.
References: