Clinical categorization is an important aid to the aetiological diagnosis of dystonia, given the plethora of genetic, neurodegenerative and acquired causes; nevertheless, a proportion of cases does not fit into any category and is defined as idiopathic. Several classification schemes have been employed to categorize the various forms of dystonia [7, 8]. The previous classification scheme was based on three axes: age at onset, distribution of dystonia and aetiology. The aetiological axis included the following categories: primary, dystonia plus, secondary and heredodegenerative dystonias. However, this type of classification has some inconsistency, partly due to the increasing knowledge on monogenic forms of dystonia, previously inappropriately defined as ‘primary’, and also to the misuse of the terms ‘primary’ and ‘secondary’ which were erroneously used for any form of isolated dystonia or any cause associated to dystonia, regardless of being acquired, degenerative or genetic. Consequently, a new classification scheme based on two axes has been released in 2013 by a Consensus Committee established under the auspices of the Dystonia Medical Research Foundation, the Dystonia Coalition and the European Dystonia Cooperation in Science and Technology Action [2]. The current classification of dystonia includes two main axes of classification: clinical characteristics and aetiology (Fig. 6.2). This approach to classification assists clinicians in a more precise categorization of patients and ultimately improves guidance towards the aetiology. The various sections constituting the current classification of dystonia should be combined and integrated in order to select the appropriate diagnostic work-up and avoid unnecessary investigations.

In the clinical characteristics axis, the following features are included: age at onset, distribution of symptoms, temporal pattern (disease course and variability) and presence of additional clinical features. According to these phenomenological criteria, dystonia might be defined as isolated (when it is the only movement disorders, except dystonic tremor), combined (when it is associated to other movement disorders such as myoclonus or parkinsonism) and associated with other neurological or systemic manifestations (i.e. spasticity, visual or hearing impairment, cognitive decline, psychiatric or behavioural symptoms). Another source of difficulty is the heterogeneity of dystonia manifestations and particularly the different distribution of symptoms; indeed, dystonia may be focal, segmental (i.e. affecting contiguous body segments), multifocal (i.e. affecting multiple noncontiguous body segments), generalized (trunk and at least two other sites are involved) or restricted to one body side (hemidystonia) (Fig. 6.3). Nevertheless, phenomenological features of dystonia are usually clustered. Whereas childhood-onset dystonia usually starts in the lower limb and often generalizes, onset in late adulthood is associated with focal dystonia, commonly involving the cranial or cervical district, and rarely generalizes, although it might spread to an adjacent body part [9, 10]. Paroxysmal appearance is peculiar of paroxysmal dystonias and dyskinesias (see later), which may depend on specific triggers.

Considering the aetiology axis, the presence of nervous system pathology, either as neurodegeneration or as static lesions (nonprogressive neurodevelopmental anomalies or acquired lesions) should be considered. Among neurodegenerative diseases, Parkinson’s disease (PD) and some atypical parkinsonisms (multiple system atrophy [MSA], progressive supranuclear palsy [PSP], corticobasal syndrome [CBS]) may have dystonia as a clinical manifestation with predilection for the lower limbs in PD, the cranial district in MSA, the neck region (retrocollis) in PSP and the upper limb (fixed dystonia) in CBS. Finally, the approach to the patient with dystonia should also carefully consider the family history, since monogenic defects have been found to underlie many forms of dystonia [11], either isolated or combined (most of them belong to the DYT classification, Table 6.2) or associated with other neurological symptoms (i.e. neurodegeneration with brain iron accumulation [NBIA]; Wilson’s disease; Huntington’s disease). Of note, the screening for a genetic mutation should not just rely on positive familial history, since it might have low sensitivity for the extreme variability of dystonia [14] and low penetrance of few genes such as DYT1 [15] and DYT6 [16]. A valid aid to screen for a specific mutation considers age at onset, distribution of symptoms and presence of signs of neurodegeneration on neuroimaging; indeed, dystonia due to PANK2 gene mutation affects the oromandibular district, is associated to other neurological manifestations and might exhibit specific findings (eye of tiger sign) on brain MRI (Fig. 6.4).

Genetic features used for classification include mode of inheritance and molecular genetic data, such as linkage to a known gene locus or identification of a specific genetic defect (Table 6.2). Monogenic defects have been found to underlie many forms of dystonia [17], and recent findings on proteins involved are shedding light on possible disease mechanism. In the past 20 years, molecular studies have improved our knowledge on the pathogenesis of dystonia and suggested numerous potential disease pathways, including dopamine signalling, intracellular transport, cytoskeletal dynamics, transcriptional regulation, cell cycle control, ion channel function, energy metabolism, signal transduction and detoxification mechanisms [18]. For example, the biosynthesis of dopamine is affected at different levels by mutations of gene encoding for enzymes involved in this metabolic pathway (GTP cyclohydrolase 1, tyrosine hydroxylase, sepiapterin reductase). Dopamine signalling is also involved in dystonia caused by GNAL mutations (DYT25), since the mutated gene encodes the stimulatory alpha subunit Gα_olf of a G protein coupled with dopamine type 1 receptors [19]. Monogenic forms of dystonias have been traditionally labelled as ‘DYT’ genes, according to the gene or locus involved (Table 6.2); this molecular classification encompasses clinically and genetically heterogeneous disorders, including erroneously assigned loci, duplicated loci and missing and unconfirmed loci [20]. DYT9 (paroxysmal choreoathetosis with spasticity) and DYT14 have been removed from this list, as they turned out to be the same genes involved, respectively, in DYT18 (paroxysmal exertion-induced dyskinesias) and DYT5 (dopa-responsive dystonia, DRD) [21, 22]. Currently, 23 different types of dystonia are considered ‘DYTs’ and designated DYT1-26. Among them, 16 genes have been identified as causing isolated, combined or paroxysmal dystonia; some of these genes may also cause other diseases, including severe encephalopathies such as hypomyelination with atrophy of the basal ganglia and cerebellum (H-ABC) syndrome (DYT4, TUBB4A mutation) [23], progressive infantile encephalopathy with ptosis and mental retardation (DYT5b, tyrosine hydroxylase mutation). Interestingly, SLC2A1 mutations (DYT18) might be associated with a wide spectrum of clinical manifestations [24, 25] including Glut1 deficiency syndrome (early-onset epileptic encephalopathy: epilepsy, psychomotor delay, ataxia and microcephaly) or generalized dystonia, chorea and ataxia. SLC2A1 encodes the major glucose transporter in the brain, placenta and erythrocytes, and its mutations often cause paroxysmal neurological manifestations: various forms of idiopathic/genetic generalized and focal epilepsies, paroxysmal exercise-induced dyskinesia (PED), hemiplegic migraine and more rarely paroxysmal non-kinesigenic dyskinesia (PNKD). Mutations in the proline-rich transmembrane protein 2 (PRRT2) gene (DYT10) may cause paroxysmal kinesigenic dystonia (PKD), paroxysmal torticollis, benign familial infantile convulsions with paroxysmal choreoathetosis (ICCA) and hemiplegic migraine [26, 27]; mutations in ATP1A3 (DYT12) are the underlying cause of rapid-onset dystonia–parkinsonism and alternating hemiplegia of childhood [28, 29]. Diagnosis of these groups of dystonias is further complicated by phenotypic heterogeneity. For instance, carriers of the same GAG deletion in the TOR1A (DYT1) gene may be unaffected because of reduced penetrance or may present with different degrees of disability, even within the same family, ranging from writer’s cramp to severe generalized dystonia [15, 30]. On the other hand, different gene mutations might produce the similar phenotypes of isolated dystonia: tremulous cervical dystonia (ANO3 in DYT24, GNAL in DYT25) [19, 31] and cranio-cervical dystonia with prominent laryngeal involvement (TUBB4 in DYT4, THAP1 in DYT6) [32, 33]. Nevertheless, the recently discovered ANO3, GNAL and CIZ1 genes appear to be rare causes of adult-onset cervical dystonia.

Despite phenotypic overlap, some clinical clues may serve to support the assessment for specific monogenic dystonias. Early-onset generalized dystonia starting in a lower limb is most commonly associated with mutations in the TOR1A gene (DYT1) or in the GCH1 gene (DYT5). Although mutations in the latter gene cause DRD, with its characteristic response to levodopa, these two conditions can be otherwise similar in a young patient. Early-onset dystonia in an upper limb also suggests TOR1A mutations (especially when the cranial musculature is spared) but may also be a sign of THAP1 mutations (DYT6), the second known form of early-onset dystonia with a tendency to generalize. Indeed, spasmodic dysphonia and prominent cranio-cervical involvement are the clinical hallmarks of DYT6 dystonia.

DRD describes a group of dystonias characterized by their response to small doses of levodopa, due to mutations of genes involved in dopamine synthesis [34]. Heterozygous mutations in the GTP cyclohydrolase 1 (GCH1–DYT5a) gene are the most common cause of DRD, whereas mutations in the tyrosine hydroxylase and sepiapterin reductase genes are less frequent causes but produce a more severe clinical picture than GCH1 mutations. Parkinsonian signs might appear over the course of DRD due to GCH1 mutations; recently, 11 different heterozygous GCH1 variants, some of them already described for DRD, have been associated to increased risk of PD, expanding the phenotypes of this mutation [35] (see also Sect. 6.1).

Paroxysmal dyskinesias (PxDs) are also included among the DYT loci. PxDs are a heterogeneous group of rare conditions characterized by recurrent episodes of involuntary movement disorders lasting only a brief but variable duration [12]. PxDs are classified based on the specific trigger as PKD, PKND and PED (Table 6.3). Some but not all patients affected by PxDs might carry a mutation in one of these genes: PRRT2 (PKD) [26], myofibrillogenesis regulator 1 (MR-1) (PNKD) and SLC2A1 (PED) [24]. Recently, a mutation in calcium-activated potassium channel, subfamily M and alpha member 1 (KCNMA1) associated with a PNKD phenotype and/or epilepsy has been reported in a family. One must be aware that the same gene may produce different PxD phenotypes [37] (Table 6.3).

A number of inherited conditions are associated with dystonia, commonly combined with parkinsonism. Indeed, dystonia is part of the clinical picture of Parkin, PINK1 and DJ-1 mutations (respectively, PARK2, PARK6 and PARK7), which cause early-onset parkinsonism; specifically, lower limb dystonia might be the presenting feature of parkin mutation [38]. Among dystonias associated with other neurological or systemic manifestation, the following conditions manifesting with dystonia should be considered: Wilson’s disease, neurodegeneration with brain iron accumulation (NBIA) (Table 6.4), Huntington’s disease, spinocerebellar ataxias (mainly SCA2, SCA3 and dentatorubro-pallidoluysian atrophy), Lesch–Nyhan syndrome, Mohr–Tranebjaerg syndrome (dystonia–deafness) and mitochondrial diseases (POLG mutation, Leigh syndrome) (Table 6.5). Dystonia can be also caused by acquired causes, such as neuroleptic drugs, vascular pathology, manganese toxicity (i.e. ephedronic encephalopathy) and infections (Table 6.6).