A multi-source approach to determine SMA incidence and research ready population

Spinal muscular atrophy (SMA) is an autosomal recessive neuromuscular disorder. In SMA, a mutation in the survival motor neuron gene (SMN1) at locus 5q13.2 leads to degeneration of alpha motor neurons, resulting in progressive muscular weakness [1]. The majority of patients (92%) have a homozygous SMN1 deletion. In the remaining patients, point mutations are found or SMA is caused by mutations in other genes [2]. A homologous copy of the SMN1 gene, the SMN2 gene, is presented at the same chromosome, which is capable of producing about 10–20% of full-length SMN protein [3, 4]. SMN2 is presented in varying copying copies, which plays a role in the heterozygosity of the phenotype [5‐7].

The clinical classification system is based on the age of symptom onset and the maximum motor function achieved [8, 9]. Type I SMA (Werdnig–Hoffmann disease) has an onset in the first months of life. Patients are never able to sit without support and without ventilatory support most patients will not survive after 2 years [10, 11]. Type II patients, with onset between six and 18 months of age, reach the ability to sit independently. Type III (Kugelberg–Welander disease) is less severe, with onset after 18 months of age. Patients gain the ability to walk independently and usually survive into adulthood [8, 9, 12].

According to the literature, SMA due to SMN1 mutations has an incidence of approximately 1 in 10,000 newborns [13‐18] and a prevalence of approximately 1–2 per 100,000 persons [13, 19]. Most patients suffer from SMA type I [15]. However, no worldwide studies have been performed. Numbers are mainly based on small studies, many of which predate genetic testing and with classification schemes that have changed over the years, highlighting the need for contemporary data.

This study aimed to estimate the worldwide incidence of SMA and the research ready and accessible population, using a by combination of multiple sources, including genetic laboratories and patient and clinical registries.

Spinal muscular atrophy is one of the leading genetic causes of infant mortality and represents a significant healthcare burden. With the development of promising new therapies for this condition [24, 25], comes the need for an improved understanding of its epidemiology and the access to specialized care. TREAT-NMD is a global network that plays a key role in addressing these important issues. To date, no global epidemiological studies of genetically confirmed SMA have been performed. Information is scarce and derives mainly from a limited number of regional studies, often predating genetic testing results, or from estimations based on carrier frequencies obtained from larger population cohorts.

In the absence of large-scale surveillance for SMA, which appears not feasible at present, a novel, international, multi-source approach was used. This approach enabled us to estimate the SMA incidence in multiple countries and to gain insight regarding the portion of the SMA population which is able and willing to participate in SMA research.

Whilst we initially contacted laboratories across the globe, response rates in European countries exceeded other parts of the world. There are several potential reasons for the lack of response or data collection from other countries. Whilst a reliable database (Orphanet) listing the majority of laboratories in European countries exists, for other continents, this is not the case, and our identification of genetic laboratories from non-European countries may, therefore, not have been as robust. Second, publicly owned laboratories, which are common in Europe, more often provided data than those which were privately owned. Furthermore, the response rate was highly improved by contacting laboratories via native speakers and local contacts, which was supported by the infrastructure provided by TREAT-NMD. This observed variability in response rates highlights the importance of multicentre, multinational collaboration in the rare disease field.

Estimated incidence of genetically confirmed SMA patients in 18 European countries ranged from 1 in 3900 to 16,000. There are two sources of data to compare our findings with: studies of genetically confirmed cases observed in the clinic and studies of carrier rates, which in some cases provide projections of potential cases in the population. Incidence estimates based on carrier screening yield higher estimates than population-based studies of observed cases. Wilson and Ogino projected an incidence of 1 per 6000 live births (~16.7 per 100,000 live births) from a carrier frequency estimate and a summary incidence of 9.7–10.1 per 100,000 live births, estimated from 15 studies of clinically diagnosed cases observed between 1960 and 1996 [15, 18]. Similarly, Jedrzejowska et al. observed in Poland a birth incidence of 1 per 9749 births (10.3 per 100,000 live births) but projects an incidence of 1 per 4900 live births (20.4 per 100,000 live births) from carrier frequencies [14]. There are several reasons that could cause differences between population-based incidence and incidence projected from carrier frequencies. The latter could be an underestimation because of de novo mutations (~2% of SMA patients [26]), limitations of diagnostic testing that cannot detect point mutations (~5% of all mutations [2]), and multiple copies of SMN1 on the same chromosome [27], resulting in higher false-negative rates if only SMN1 copy numbers are counted [28]. However, it can also be an overestimation due to greater genetic testing among persons with a higher risk of SMA, a high rate of foetal death due to the disease severity, and lethality of the absence of SMN1 and its homologue SMN2, absent in 10–15% of the general population [16]. Furthermore, there are reports of unaffected individuals with no functional SMN1 copies [29‐31]. High rates of consanguineous marriages in some countries/communities may contribute to the variation in estimations.

Variability between countries included lower incidences in some Northern and Western European countries and higher incidences in other European countries. It is important to note that the responses to the questionnaire indicate the laboratory location and not necessarily the residency of the patient. Some of the variability in incidence rates may, therefore, be accounted for by cross-border testing by the laboratories. In Germany, several laboratories have indicated that they perform cross-border testing, which could account for the higher reported incidence there (26.7 per 100,000). This could also be the case for the relatively higher rate observed in Croatia, where some neighbouring countries do not provide laboratory testing for SMA. Conversely, some laboratories in the countries which show relatively lower incidence rates, e.g., the UK and The Netherlands, also test samples from abroad. In addition, despite only testing nationally, we found a relatively higher incidence in France, Italy, Bulgaria, and Hungary. Therefore, our incidence variability cannot be explained by cross-border testing alone. There are a few other limitations to our method of incidence estimation. Only countries with a high response rate were used when calculating the incidence. Furthermore, not all laboratories in those countries responded to the survey; however, for the countries that we took into account, local experts’ advice was utilised to ensure that the study included the main laboratories, e.g., in Italy, it is not certain that all of the remaining laboratories do test for SMA, and if they do, it will only concern a small number of tests. Second, some calculations are based on a low number of patients and live births, such as in countries with relatively small populations like Bulgaria and Hungary. In those cases, small changes in population and the number of diagnosed patients per year will have a relatively large effect on the incidence rate. Third, many laboratories cannot test for point mutations, which means that those patients might not be included in the calculations. However, some laboratories indicated that these samples were sent elsewhere for further testing and point mutations have a very low occurrence [2]. Other contributing factors for regional variation may include differences in genetic testing availability and screening practices, genetic confirmation of prevalent cases that previously only had clinical diagnosis, gene pools or rates of consanguinity, changes in the population composition, or clinical trial screening. Incidence was especially low in Greek-Cyprus and Ireland. Both countries are very small. In Cyprus, the level of genetic testing is relatively low and there is a high level of misdiagnosis. In case of the Irish SMA patients, it is possible that some of them are diagnosed in the United Kingdom. However, as the Irish population is relatively small, the number of the additional patients that might have been diagnoses in the United Kingdom would not have great impact on the results from the United Kingdom.

To estimate the size of the readily approachable and research ready SMA population, we conducted enquiries into the TREAT-NMD Global Patient Registry and the CSTR. Not unexpectedly, our findings show a subset of the SMA population prevalence found in the literature, approximately 2–5 times less [13, 19]. First, available literature mostly predates genetic testing, whereas the patients we included from registries were only those with genetic confirmation. In addition, some registries have only recently been established, and are expected to contain more patients in the future. The majority of participating registries have been set up with clinical trial recruitment in mind; therefore, patients not interested in trial participation may decide not to sign up. All registries provide data to TREAT-NMD voluntarily and lack of response of some registries may be due to a lack of resources. Furthermore, in the CTSR only specialist medical centres voluntarily enter data and not all SMA patients attend those centres.

Type I SMA, the most severe clinical presentation, is the most common subtype [15]. However, in both the patient registry and the CTSR, less than 20% of cases were classified as type I. Type I patients have a short life expectancy (<2 years of age), which will not only decrease their chance of being alive on the prevalence day, but may also reduce the likelihood of being registered by their parents in the patient registry.

We observed significant variability in the numbers of registered patients. Possible reasons for this include the healthcare infrastructure in the country affecting access to genetic testing and care, the year of setup, budget, number of staff and purpose of the registry (e.g., regulatory requirements or research/autonomous initiatives), and who is responsible for data entry (patients/guardians or professionals). No clear relationship was observed between our findings and any one of these variables (data not shown); it is likely that the variability is due to a combination of factors.

Data derived from the genetic laboratories and the registries represent unique data sets and cannot easily be compared. It is difficult to convert incidence into prevalence unless life expectancy is known, and given that SMA is a heterogeneous disease [8] with differences in standards of care between countries, it is difficult to calculate a clinically meaningful average life expectancy. The calculations are also based on small patient populations. Nonetheless, the number of patients diagnosed genetically is generally comparable or even higher than reported in previous epidemiological studies [13‐17], at least in countries, where genetic testing is readily available.

SMA is a complex neurodegenerative disease which requires a comprehensive, multidisciplinary approach to ensure the best medical care and clinical outcomes [32]. Our findings from the Global Patient Registry and the CTSR indicate that many patients are not registered at specialized neuromuscular centres, and do not self-identify via patient registries, and thus may not have access to research study opportunities, the best standards of care and advanced treatment.

With a growing number of therapies being developed, there is an increasing need for reliable and larger scale SMA incidence estimations. We provide a novel method of estimating SMA incidence utilizing multiple sources. As this is the first time, a higher incidence has been studied and reported in these countries, these findings require replication with a population-based study. This study is a step forward in understanding the epidemiology of SMA and number of patients that are ready to participate in trials for new, innovative therapies or observational research and presents potentially new hypotheses to test with regard to the countries where we identified a higher than anticipated incidence.