In a recent Review article (Dopa-responsive dystonia — clinical and genetic heterogeneity. Nat. Rev. Neurol. 11, 414–424; 2015)1, Wijemanne and Jankovic proposed that the differential diagnosis of dopa-responsive dystonia (DRD) can be narrowed by defining three clinical presentations: classic DRD, DRD with parkinsonism, and early-onset atypical DRD. As a non-medical academic with genetically confirmed autosomal dominant DRD that was misdiagnosed for nearly 10 years, I support the authors' methodical approach to differential diagnosis. However, owing to my own presentation of DRD, I propose a fourth clinical presentation: atypical late-onset DRD with Parkinson disease (PD).

My own atypical phenotype — adult-onset and rapidly generalizing DRD with PD — has not, to my knowledge, been described in the literature. My illness had a similar age of onset (41 years) and similar characteristics to the progressive neurological disease that my paternal grandmother died from at 53 years of age. My father had late-onset parkinsonism, and an aunt has late-onset vocal and unilateral hand tremors. Prior to successful bilateral deep brain stimulation (DBS), I had motor fluctuations and crippling biphasic levodopa-induced dyskinesia.

My phenotype differs from the three proposed in the recent literature2,3,4,5,6, which roughly correspond to those proposed by Wijemanne and Jankovic1: a classic nondegenerative, benign form of dystonia that responds to low, sustained doses of levodopa; a degenerative form (late-onset and without dystonia) with levodopa-induced dyskinesia, parkinsonism and abnormal dopaminergic imaging; and an intermediate form. I believe that the adoption of the term 'atypical', as suggested by Wijemanne and Jankovic, would be valuable in clinical practice.

Several recently published articles discuss the clinical and genetic heterogeneity of DRD phenotypes in relation to the genetics of PD and parkinsonism2,3,4,5,6,7,8,9, with implications for the clinical presentations of DRD proposed by Wijemanne and Jankovic. I have the p.K224R mutation in GCH1, the same genotype as that of some unrelated individuals with PD but not DRD2,4. Segawa et al.10 have proposed that the location of a mutation in GCH1 influences the phenotype of DRD, so the suggestion that some GCH1 mutations cause PD could have considerable implications for people who carry a GCH1 mutation but do not develop DRD in childhood. Mencacci et al.2 suggest that administration of levodopa to such individuals in childhood might protect against later onset of PD, and they point to animal models in which “levodopa has been shown to promote recovery of nigrostriatal denervation.” However, an intermediate phenotype was observed in a patient who had been treated with levodopa since childhood: he had abnormal dopaminergic imaging, but no PD symptoms or dyskinesia5. This phenotype could indicate that dopamine replacement is not neuroprotective, and that the patient had preclinical PD3,5, but could similarly indicate that dopamine replacement suppressed the emergence of PD.

Recognition among clinicians that GCH1 mutations can manifest with a range of phenotypes, including adult-onset generalized dystonia and PD, is vital. Use of the word 'atypical' by Wijemanne and Jankovic1 in their classification recognizes that the classic paediatric DRD phenotype is not the only one, and recognition of the range of phenotypes can transform the lives of patients.