Capturing phenotypic heterogeneity in MPS I: results of an international consensus procedure

Abstract Background Mucopolysaccharidosis type I (MPS I) is traditionally divided into three phenotypes: the severe Hurler (MPS I-H) phenotype, the intermediate Hurler-Scheie (MPS I-H/S) phenotype and the attenuated Scheie (MPS I-S) phenotype. However, there are no clear criteria for delineating the different phenotypes. Because decisions about optimal treatment (enzyme replacement therapy or hematopoietic stem cell transplantation) need to be made quickly and depend on the presumed phenotype, an assessment of phenotypic severity should be performed soon after diagnosis. Therefore, a numerical severity scale for classifying different MPS I phenotypes at diagnosis based on clinical signs and symptoms was developed. Methods A consensus procedure based on a combined modified Delphi method and a nominal group technique was undertaken. It consisted of two written rounds and a face-to-face meeting. Sixteen MPS I experts participated in the process. The main goal was to identify the most important indicators of phenotypic severity and include these in a numerical severity scale. The correlation between the median subjective expert MPS I rating and the scores derived from this severity scale was used as an indicator of validity. Results Full consensus was reached on six key clinical items for assessing severity: age of onset of signs and symptoms, developmental delay, joint stiffness/arthropathy/contractures, kyphosis, cardiomyopathy and large head/frontal bossing. Due to the remarkably large variability in the expert MPS I assessments, however, a reliable numerical scale could not be constructed. Because of this variability, such a scale would always result in patients whose calculated severity score differed unacceptably from the median expert severity score, which was considered to be the 'gold standard'. Conclusions Although consensus was reached on the six key items for assessing phenotypic severity in MPS I, expert opinion on phenotypic severity at diagnosis proved to be highly variable. This subjectivity emphasizes the need for validated biomarkers and improved genotype-phenotype correlations that can be incorporated into phenotypic severity assessments at diagnosis.