In accordance with previous reports 
, RB were a common symptom of GTS as 64.4% of our patients displayed one or more RB. Compared to patients without RB, those with RB had more severe and more frequent complex tics and SIBs, and were more frequently treated with neuroleptics. This is in line with recent reports showing segregated clusters of patients with GTS with different types of psychiatric manifestations 
according to the characteristics of tics 
. It can be concluded that, independently of their phenomenological expression, the presence of RB is a hallmark of the severity of the syndrome.
Patients with GTS with RB could be categorised into three groups. In the first ‘tic-like’ group (24.1%), RB were performed without distress or anxiety and in response to an ‘urge to do’, that is in accordance with the definition of tics. The characteristic RB in this group were touching, counting, symmetry rituals and ‘just right’ phenomena. In the second, ‘OCD-like’ group of patients (20.4%), RB were performed to reduce anxiety, experienced as unwanted and senseless, and performed with the aim to protect the patients from real or potential negative events. Moderate to severe distress was observed when patients attempted to suppress or delay RB, a situation that corresponds to the criteria of compulsion such as those observed in anxious-type, non-tic related OCD. Characteristic RB in this group were checking and washing behaviours. The third, ‘mixed’ group (13.0%) could be identified in patients with both ‘tic-like’ and ‘OCD-like’ types of RB.
The clinical features of patients with GTS differed according to the type of RB observed. In the ‘OCD-like’ group compared to the ‘tic-like’ group (and to the non-RB patients), a higher score for complex tics and less successful socio-professional adaptation (higher score of the overall impairment item of the YGTSS and lower score of the GAF) was noted. In the mixed RB group compared to the ‘tic-like’ and ‘OCD-like’ groups, all clinical scores were higher with the exception of self-injurious behaviours and echophenomena. Taken together, these data suggest that ‘OCD-like’ RB differs from ‘tic-like’ RB, not only from a phenomenological point of view, but also in terms of treatment response and socio-professional adaptation. This contrasts with patients in the ‘tic-like’ RB group who had similar characteristics as the non-RB patients with rather limited impact on treatment response and socio-professional adaptation.
Several previous studies have examined the phenomenological characteristics of RB in patients with GTS (summarised in Table S2
), many of which agree with our study insofar as symmetry, touching, counting and ‘just right’ phenomena were typical RB in patients with GTS. Also, in patients with GTS classified as having anxious-type OCD, washing and compulsions differed significantly from the ‘tic-like’ group. However, all of these studies have compared OCD and GTS patient groups and have not ascertained the prevalence of RB in GTS with regard to their ‘tic-like’ or ‘OCD-like’ characteristics, with the notable exception of Cath et al. 
, although a different etiology of ‘compulsive’ symptoms was hypothesised by several authors 
. Accordingly, most studies to date persist in classifying RB in patients with GTS as OCD. Current reviews still consider OCD as the major co-morbidity of GTS with rates up to 50% in the adult population 
, although the concept of ‘tic-like OCD’ has recently been put forward 
. In contrast to the studies summarised in Table S3
, only Shapiro and Shapiro 
have clearly expressed the idea that some RB in patients with GTS may simply represent complex motor tics and termed them ‘impulsions’. However, we and other authors (Table S2
) disagree with the low percentage of anxious-type OCD in patients with GTS as observed by Shapiro and Shapiro 
Obviously, the phenomenological distinction of RB in patients with GTS is important to consider, since the treatment of tics and OCD symptoms differs. Previous reports suggested that the presence of tics and GTS in OCD patients reduces the response of RB to symptomatic treatment, whether with SSRIs or cognitive-behavioural therapy 
. The distinction of two types of RB, either of the ‘tic-like’ or of the ‘OCD-like’ type, strongly suggest that the former should respond to neuroleptic treatments and habit reversal training and the latter to SSRIs or exposure-based cognitive behavioural therapy 
. In addition, multiple studies of OCD symptoms suggested existence of four to six principal OCD-dimensions 
that have been associated with different patterns of heritability and specific genetic polymorphisms 
, as well as with differential response to pharmacological and non-pharmacological treatments 
. Thus, distinction of RB into two different groups also suggests that the clinical expression of RB (‘tic-like’ or ‘OCD-like’) could result from the dysfunction of partially overlapping but distinct neuronal circuits. Both tics and OCD compulsions are considered to result from the dysfunction of cortico-striato-thalamo-cortical circuits 
. In patients suffering from GTS without psychiatric co-morbidities, structural and functional neuroimaging studies as well as transcranial magnetic stimulation studies have shown dysfunction of premotor, sensorimotor, dorsal parietal, dorsolateral prefrontal and cingulated and insular cortical areas 
. Similarly, in OCD patients with predominant ‘symmetry/ordering/counting’ behaviours but without tics, dysfunctions in motor, parietal and insular cortices have recently been described 
. Consequently, the ‘symmetry/ordering/counting’ dimension in OCD has not only close phenomenological characteristics to tics but may also share a similar neuronal basis. If so, this specific OCD dimension probably rather represents an integral part of GTS than being a distinct co-morbid condition.
In patients with OCD displaying washing and checking compulsions, neuroimaging studies have shown dysfunction of orbitofrontal, cingular and temporal cortices as well as of the caudate nucleus 
. Thus, ‘contamination/washing’ and ‘harm/checking’ OCD dimensions differ from tics not only by their phenomenological characteristics, but also in their underlying neuronal circuits and would therefore represent a true co-morbid condition. However, the caudate nucleus could play a prominent role in both disorders, since in two large MRI volumetric studies a diminished volume of the caudate nucleus was correlated with the presence of RB and persistence of tic and OCD symptoms into adulthood 
. In light of our results, we therefore stress the importance of a clear phenotypic characterisation of patients with GTS with regard to the nature of their RB (‘tic-like’, ‘OCD-like’ or both) in future neuroimaging studies.
First, if our sample is large, it is also heterogenous regarding age groups. Of note, 45/166 patients (27%) in our sample were 20 years or younger and may thus present a different phenotypic profile upon further aging. Second, no inferences on pathophysiology can be made based on descriptive cross-sectional research. Also, co-morbidity issues should be solved with the aid of family or twin studies, in which occurrence of tic-like or OCD-like behaviours is investigated in family members of probands with either GTS without OCD, GTS+OCD and OCD alone. These family studies 
have indicated that OCB/OCD and related repetitive behaviours in GTS patients are intrinsic to the pathology, although the debate is ongoing. Third, regarding therapeutic recommendations, it must be noted that OCD symptoms may also respond to neuroleptic treatment; however, pharmacological first line treatement of OCD remain SSRIs 
and neuroleptics are currently rather used as augmentation treatment for treatment-resistant OCD 
. Fourth, it can be argued that our results are induced by the method used since we have predefined classifications of the patients according to the symptoms observed (‘tic-like’ vs ‘OCD-like’) and then show that the same symptoms are distributed unevenly into these two categories. However, 6.3% of RB could neither be clearly attributed to the ‘tic-like’ nor the ‘OCD-like’ group and thus fall into a different category, albeit one that remains to be defined. The most important differential diagnosis in the domain of involuntary movements should be stereotypies which often co-exist with tics.
We suggest the importance of a precise semiological analysis of RB in patients with GTS, which may be particularly important for neurologists unfamiliar with the spectrum of OCD symptoms. We suggest that a substantial part of RB in patients with GTS are complex tics, as initially suggested by Shapiro and Shapiro 
and warrant to be treated as such, either pharmacologically and/or by behavioural therapy. Conversely, neurologists facing OCD-like symptoms in GTS patients should seek treatment advice from their psychiatric colleagues, thus advocating a multidisciplinary approach in diagnosing and treating patients with GTS.