Rett syndrome: a wide clinical and autonomic picture

Background Rett Syndrome is a neurodevelopmental disorder almost exclusively affecting females, characterized by a broad clinical spectrum of signs and symptoms and a peculiar course. The disease affects different body systems: nervous, muscolo-skeletal, gastro-enteric. Moreover, part of the symptoms are related to the involvement of the autonomic nervous system. In the Tuscany Rett Center at Versilia Hospital, we collected data from 151 subjects with a clinical diagnosis of classical or variant RTT syndrome. For each subject, we assessed the severity of the condition with clinical-rating scales (ISS, PBZ), we quantified the performance of the autonomic nervous system, and we performed genetic analysis. We used multivariate statistical analysis of the data to evaluate the relation between the different clinical RTT forms, the cardiorespiratory phenotype, the different genetic mutations and the severity of the clinical picture. Individuals were classified according to existing forms: Classical RTT and three atypical RTT: Z-RTT, Hanefeld, Congenital. A correlation between C-Terminal deletions and lower severity of the clinical manifestations was evident, in the previous literature, but, considering the analysis of autonomic behaviour, the original classification can be enriched with a more accurate subdivision of Rett subgroups, which may be useful for early diagnosis. Results Present data emphasize some differences, not entirely described in the literature, among RTT variants. In our cohort the Z-RTT variant cases show clinical features (communication, growth, epilepsy and development), well documented by specific ISS items, less severe, if compared to classical RTT and show autonomic disorders, previously not reported in the literature. In this form epilepsy is rarely present. In contrast, Hanefeld variant shows the constant presence of epilepsy which has an earlier onset In Hanefeld variant the frequency of apneas was rare and, among the cardiorespiratory phenotypes, the feeble type is lacking. Conclusion A quantitative analysis of the different autonomic components reveals differences across typical and atypical forms of RTT that leads to a more accurate classification of the groups. In our cohort of RTT individuals, the inclusion of autonomic parameter in the classification leads to an improved diagnosis at earlier stages of development.


Background
Rett Syndrome (RTT) is a neurodevelopmental disorder almost exclusively affecting females, characterized by a broad clinical spectrum of signs and symptoms and a peculiar course during life. It occurs with an average frequency of about 1:10,000 in girls [1].
After the first identification of the gene in 1999 [2], MECP2 mutations have been identified in 90-95 % of the Classical RTT cases, where, after a period of apparently normal neuropsychic development with some alterations in the general movements [3], individuals after 6-18 months present an arrest in their development followed by a progressive deterioration of acquired skills such as purposeful hand function and communication, a deceleration of head growth, with the subsequent appearance of stereotypic hand movements. This syndrome is associated with comorbidities including reduced somatic growth, gastro-intestinal problems, osteopenia, gait apraxia, scoliosis, autonomic dysfunction, breathing disturbances and frequent seizures.
According to the new revised criteria, it has been recognized that some individuals present many of the clinical features of RTT, such as regression, but do not fit the criteria established for the diagnosis of Classical RTT [4]. These have been termed "atypical" RTT [5].
Among RTT atypical forms, some specific variants have been described: Preserved Speech Variant or "Zappella variant" (Z-RTT), characterized by milder clinical abnormalities and by the appearance of some degree of speech [6,7]; the "Hanefeld variant" (CDKL5 gene related) in which there is a pathognomonic early onset of seizures [8,9]; the "Congenital variant" (FOXG1 gene related), with onset of the symptoms since after birth [10]. In these last two variants, the mutation and several clinical features are notably different, hence they are considered as distinct clinical and molecular entities [11].
All these forms have different severity degrees in terms of comorbidities, behaviour, prognosis and involvement of the autonomic nervous system. The present study describes the clinical and autonomic features of the RTT girls who attend the Tuscany Rett Center, in order to obtain more information on the correlation between genotype and phenotype.

Methods
Our cohort consists of 151 individuals evaluated in the Tuscany Rett Center, Versilia Hospital (Lido di Camaiore, Italy) from January 2006 to April 2014. During the study six girls died due to complication of illnesses or sudden death or because of drug resistant infections.
Diagnosis of RTT (Classical, Atypical) was made by two independent child neuropsychiatrists with expertise on RTT, and individuals were categorized on the basis of the RTT consensus revised criteria [4] 64.9 % of the girls presented classic Rett, while 35.1 % presented atypical RTT. 20 % of the individuals presented evident autistic signs.
In our cohort all the forms described in the current classification [4] were present, considering the wide variability of the RTT phenotype, the age of onset, the severity of impairments, and the clinical course. We also observed some different type of "Rettoid" variants that are now described as RTT spectrum, and were previously classified as "forme fruste": late childhood regression variant, Rettoid Male, familial atypical variant [12].
In our cohort there are also individuals that can be classified as atypical RTT, but they do not belong to any of the specific forms. These individuals are two boys and four girls with MECP2 mutations, one female with MEF2C mutation and two girls with no mutation. These individuals have been grouped in variant cases and they will be referred to as Atypical RTT not otherwise specified (ARTT-NOS).
Phenotypic data were collected through clinical and neurological evaluations. In all RTT cases, cardiorespiratory parameters and brain activity were investigated. For each patient we assessed: clinical severity, cardiorespiratory parameters, brain activity, and genetic mutation.

Clinical assessment
For clinical assessment of the severity degree, all individuals were evaluated using both the International Scoring System (ISS) [13], and the Pini Bonuccelli Zappella scale: PBZ [14].
International Scoring System (ISS): The ISS scale consists of 21 items regarding the RTT typical features, divided in five subscales: Growth and Development, Musculoskeletal, Movement, Mental-Cortical and Brainstem-Autonomic. Each item score ranges from 2 to 0 as follows: 2 severe abnormality, 1: mild abnormality, 0: no abnormality. Lower the score, the better the clinical condition. The total score range between 0 and 15 has been considered as mild, between 16 and 24 as moderate and above 24 as severe.
Pini Bonuccelli Zappella (PBZ): The PBZ scale takes into account specific items about daily skills, and consists in 26 items grouped into 3 functional domains: movement (MOV): facial expressions, eye contact, praxis, stereotipies, walking, sitting position, muscular tone, tropism and dystonias and other movement disorders; communication and sensitive (C-S) speech production, comprehension, attention to visual, tactile, and auditory stimuli, breathing, convulsions, appetite, sleep, night-time sleep, day-time sleep, aggressivity; daily living (DL) climbing and, descending stairs, feeding/swallowing, feeding autonomy, falls unrelated to seizures. Each item score ranges from 0 (normal) to 4 (more severe).

Cardio-respiratory assessment
All the individuals have been investigated with the Neuroscope system (Medifit Instruments Ltd, London, UK) [15] to analyse their autonomic functions.
The NeuroScope™ employs VaguSoft software (MediFit Diagnostics Ltd, London, UK), allowing to record simultaneously several autonomic cardiovascular and respiratory parameters non-invasively and synchronously with EEG and video monitoring over a 1-h period. The autonomic parameters included: cardiac vagal tone, heart rate, transcutaneous blood gases (TCM4, Radiometer, Copenhagen, Denmark), and respiratory patterns. Thirteen abnormal, awake breathing rhythms have been identified in the RTT syndrome: Breath hold, Central apnoea, Rapid and shallow breathing, Hyperventilation, Tachypnoea, Deep breathing, Biot's breathing, Cheyne-Stokes breathing, Regular breath holds, Protracted inspiration, Atypical breathing, Shallow breathing and Valsalva's manoeuvre. Following analysis of the recording each patient was assigned to a specific cardiorespiratory phenotype according to the predominant pattern of respiratory dysrhythmia [16].
Three abnormal breathing rhythms in the waking state have been identified in the RTT population and categorised into Feeble, Forceful, and Apneustic type of breathings. They constitute three unique cardiorespiratory phenotypes with different levels of blood gases, autonomic tone, physical features, clinical complications and idiosyncratic responses to drugs [17].
The cardio-respiratory assessment has been performed in the Tuscany Rett Center only on individuals with confirmed clinical and genetic diagnosis.

Genetic analysis and mutations classification
Genetic analysis has been performed on all the individuals with clinical RTT phenotype clinical presentation, first for MECP2, (sequencing of the coding region and Multiple Ligand Probe Amplification) and, whenever negative or in case of early infantile onset epilepsy for CDKL5. In two females FOXG1 and MEF2C molecular analysis was performed upon specific clinical suspicion [18,19]. The majority of genetic tests have been performed in different molecular laboratories. All MECP2 and CDKL5 mutations have been classified into 3 principal categories according to the nature of the sequence variation and its subsequent effect on the protein sequence: substitutions (nonsense, missense), deletions and frameshift. Among the frameshift mutations we focused on the mutations located in the C-terminal segment of the gene. The group of the nonsense mutations is furthermore divided into early truncating and late truncating mutations according to the position of the mutation in the gene [20].

Description of the sample and of the different clinical RTT forms
Our sample consists of 151 individuals: 149 females and 2 males, 118 individuals present mutations in MECP2, 12 females are mutated in CDKL5, 1 in FOXG1 and 1 in MEF2C. In 19 individuals all the molecular analysis resulted negative. The mean age in our cohort is 12 years (range 1-49, median 10).
Genetic tests and instrumental evaluations have been performed in all individuals. However, for some few individuals, some data (like auxological parameters at birth or development in the first year of life) were lacking. This is the reason why the tables below have different numbers.

Statistical analysis
Data are described as mean and standard deviation (SD), median and percentiles for continuous variables and as absolute and relative frequencies for categorical variables.
Parametric analysis (ANOVA with Bonferroni) for continuous variables and the χ 2 test or Fisher's exact test for categorical variables, were used to measure differences between groups.
P value < or equal to 0.05 were considered statistically significant and all P values were based on two tailed tests. Statistical analysis was performed using STATA 9.0 for Windows (StataCorp, College Station, TX).

ISS (subscales and items) and clinical forms
In order to evaluate the clinical severity of the common features in the RTT disorder, a modified version of the International Scoring System was used [16].  Cortical investigation is related to three main parameters: intellectual disabilities, language impairment and epilepsy. We analyzed intellectual impairment, language and epilepsy. For epilepsy, we considered its onset and its distribution across clinical forms. For cognitive level and language, we could only use two scores: profound or mild impairment, since there were no girls with a "no compromise" value. However there were some Z-RTT girls able express themselves relatively well, with extensive vocabulary and spontaneous construction of sentences. Speech Since the first description of Andreas Rett one of the key symptoms was the absence or loss of speech. At the onset of the disease some children had already developed a vocabulary of a few words or even sentences with two words. To this stage usually follows a stop phase and then regression, however 12.6 % of females in our sample recovered and amplified the vocabulary, using more than 10 words at age of 5; they constitute the Zappella variant. In this form we recognized the presence of moderate speech consisting of a few words, and eventually the construction of complex sentences, sometimes echolalic, sometimes purposeful, with an extensive vocabulary. In some the language skills extended to the understanding of writing.

Intellectual impairment
The disorder involves a significant cognitive impairment, very difficult to quantify in standard tests: the lack of speech and the eminent dyspraxia implies that many assessments should be done with eye tracking technology or other techniques of augmentative and alternative communication (AAC). These methods are incompatible with short assessments, carried out in our sample.   Table 6). The median age of onset of epilepsy in the entire sample was 50 months with a range between 1 and 216 months. An earlier onset epilepsy was present peculiarly in the Hanefeld variant compared to the other forms (F test = 9.38, p < 0.001) ( Table 7). e) Autonomic domain The autonomic Nervous system is related to several abnormalities: impaired swallowing, peripheral circulation in limbs, mood disorders, sleep disorders, abnormalities in the gastro-esophageal tract, urinary bladder and cardio-respiratory abnormalities (see below: Respiratory abnormalities and cardiorespiratory phenotypes). Here we report gastro-enteric dysfunction and sleep (Table 8).
Gastro-enteric disorders RTT syndrome is often associated with nutritional intake and growth problems. In general, feeding difficulties are complex and involve oromotor, behavioral, nutritional and medical components. Oromotor problems may include: oropharyngeal dysfunction, sensory defects, reduced mobility of the tongue. One or all phases of the deglutition function can be compromised. Furthermore, breath disorders could disturb the swallowing ability; drug used to control epilepsy can induce lack of appetite, sedation and increase  Evaluation with Pbz scoring system PBZ and clinical forms PBZ The PBZ is a scale of evaluation designed to test severity in Classical Rett and its variants. PBZ scoring scale, as ISS, considers motor performances and sensitive and comunicative abilities (S-C), and in addition, takes into account daily living. This scoring system has a  similar trend of ISS, but measures also daily living and it is more sensible for the association analysis with genetic mutations. The scale is susceptible to recording even minimal changes in clinical symptoms over the course of therapy, rehabilitation programmes and environmental enrichment, but also to indicating the severity of different clinical and genetic groups (Table 11). Females with Hanefeld variant presented the worst data for ISS and daily living. Table 12 shows the various classes of PBZ for Classic Rett and variants: the differences were all significant in each area: Motor area: all forms were significantly more severe than the Z-Rett (χ 2 test = 40.53; p < 0.001). Communicative and sensory skill area (SC): in a total of 6 females with Hanefeld presentation, 3 variant were severe and 3 very severe, and showed a significant difference versus all other (χ 2 test = 25.93; p < 0.01), while the Z-Rett rank mainly between mild forms (50 %) and medium (36 %).
Daily living: the Hanefeld variants were placed in 33 % of the serious and in 50 % among the worst forms; the Z-Rett presented better than the other with significant statistical difference (χ 2 test = 25.04; p < 0.01).
Total: there were significant differences in the PBZ scores between different forms: 71.4 % of Z-Rett girls presented a mild score while the majority of the girls presented a medium score (χ 2 test = 39.35; p < 0.001).

Cardiorespiratory evaluation Respiratory abnormalities and cardiorespiratory phenotypes
The originality of our study consists in the evaluation of the autonomic system through the use of Neuroscope which is non-invasive procedure, and provides important information on respiratory and heart parameters, and simultaneous video EEG registration electroencephalographic [15] RTT individuals who carried out the assessment with the neuroscope were 135 /151.
Thirteen abnormal, awake breathing rhythms have been identified in the RTT syndrome-see Methods section.
On the basis of the neurovegetative different features (vagal tone, heart rate, etc..), of the data of the hemogasanalysis and of the rate of the breath abnormalities, we define three different cardiorespiratory phenotypes: Apneustic, Feeble, Forceful [16]. Apneustic breathing is characterized by the prevalence of fragmented breath, in which isolated breath hold or regular and prolonged respiration dominate. Feeble breathing is characterized by the prevalence of a superficial and fast respiration during most of the time. Peculiarity of the Forceful breathing is the prevalence of hyperpnea, tachypnea and deep respiration. The recognition of the different cardiorespiratory phenotypes is important since each of them determines different metabolic consequences and determine different treatments [17]. We performed additional statistical analysis taking into account the age of the subjects in the sample. We considered three age groups: 0-5 years, 6-14 years old, and over 14. We found that the cardiorespiratory phenotypes apneustico and forced were stable over time, while the phenotype feeble increased significantly (chi2 = 23:32 p = 0.001). This trend is in agreement with data in literature [15] which showed that individual respiratory abnormalities (hyperpnea, apneas and breath hold) are more important in younger girls and decline over time, while shallow more represented in older females (third group).
Below we report the percentage of abnormal breathing for each clinical form (Table 13).
There is no significant correlation between percentages of abnormal breathing and clinical forms. Considered the importance of the breathing phenotype in RTT, we investigated how different parameters correlate to   the breathing phenotype starting with BMI. For each patient, we correlated the breathing phenotype determined with clinical assessment with the BMI (Table 14). The comparison among cardiorespiratory phenotypes and BMI was not statistically significant (χ 2 test = 2.13; p = 0.71), however in the majority of cases (42 %) the forced phenotype presented a pathological BMI (<3 rd centile).

Cardiorespiratory phenotype and clinical forms
We considered the different cardiorespiratory phenotype in the different RTT clinical forms, and we noticed that all the respiratory abnormalities are represented in all the RTT groups but in Hanefeld variant, where we found mainly hyperpnea, tachypnea and breath-hold, while the Feeble Phenotype was lacking.
The distribution of the different cardiorespiratory phenotypes in the sample (n = 134) was the following: apneustic n = 14 (10.4 %), feeble n = 52 (39.5 %) and forceful n = 66, (48.5 %). In three cases no pathological abnormalities were present and in one case the results of the analysis were not evaluable.
Seventeen individuals could not be assessed for technical issues (Table 15).
In only one classical RTT and in 2 ARTT-NOS cases no respiratory abnormalities were recorded. The following tables refer only to cases with a complete autonomic evaluation. There were no cases of Feeble phenotype in females with Hanefeld presentation.

Single parameters in cardio-respiratory phenotype
We looked at specific abnormalities in all the groups considered: Tachipnea, Iperventilation, Breath-hold, abnormal breathing, Valsalva Maneuver, cardiac parameters (HR and CVT) and transcutaneous parameters (PO2 e PCO2) (Table 16).
Valsalva maneuver was independent from the cardiorespiratory phenotype. In the unique case of congenital RTT, no Valsalva's was recorded. Valsalva's was more common in the classical form and even so predominant to hide the cardiorespiratory phenotype [16].
Cardiac activity We also looked at the cardiac activity, and the vagal tone (Tables 17, 18, 19 and 20).
There was a significant difference between forceful and feeble fenotype (F test = 3.08; P < 0.005).
The average heart rate frequency recorded in our protocol was 96.7 bpm. We did not find significant differences.
The average value at rest for vagal tone (CTV) was 5.88.
We found differences in the CVT among different groups, with a significant abnormality in feeble and apneustic versus forced (F = 5.80; P <0.005 Anova with Bonferroni). Individuals with forced phenotype showed values closer to standard [15]. Individuals with forced phenotype had the lowest values of both PO2 (means 90.2; SD 13.3), and PCO2 (means 33.5; SD 12.8); while the weak phenotype and  In particular the mutations identified in the two males were one missense mutation (p. Arg306His, in a mosaic state) and one frameshift (p. G269fs). Furthermore, there were 12 cases with a mutation in the CDKL5 gene (9.0 %), 1 case with a mutation in FOXG1 (0.75 %) and 1 case with a mutation in MEF2C (0.75 %). In this study we found a significant correlation between C-terminal and the Z-RETT variant [21]. 2-Different clinical RTT forms and genes involved (     [6,22,23]. The age-range is from 2 to 20 years. Growth parameters were in the normal range in almost all the girls (except one case for which data are not available). OFC was in the normal range in about 10/19 cases, in 6/19 between the 3rd and 40 % centile and only in 2 cases females showed clear microencephaly. Total ISS score evaluation showed for C terminal deletion cases the majority (13/19) had a mild score, 4/19 case had a moderate score and in 1/19 case a severe score. Epilepsy was absent in 8/19 cases, present but drug controlled in 10/19 cases and only in 1/19 case the epilepsy was drug resistant. Scoliosis was absent in the 8/19 cases and present but of a mild degree in 7 cases. No one RTT girls needed a surgical intervention for scoliosis. Only 4/19 were not able to walk, while 9/19 walked with a support and 6/19 were able to walk alone. Tremors or tremors like were present in half of this sample. In almost all cases, hands stereotypies were present but non dominant. In 6 of them (range 7-20 years of age), all with Z-RTT, the language was composed by >20 words. The cardiorespiratory phenotype was forced in 9 cases, feeble in 8 cases, and apneustic in 1 case; 1 case was not analysed.

4-CDKL5 positive cases
Overall we had 13/151 cases with Hanefeld variant who all but one presented CDKL5 gene mutations. All the cases belonging to this group can be defined as Hanefeld variant RTT. They had a age-range of 1-12 years with a mean age of about 6 years. Total ISS total score evaluation showed that Hanefeld variants had higher severity than all the other RTT forms. According to the PBZ evaluation scales, the scores for the CDKL5 cases were higher (severe) than the MECP2 one (F test = 3.18; p < 0.005) and this is mainly regarding the sensorial   Table 19 Vagal tone (CVT) and clinical form Epilepsy was present in all the cases (100 %) and there was a higher percentage of drug resistant cases compared to all the other RTT subgroups. The median age of onset was two, 2 months with a range from 1 to 7. Among gastrointestinal problems, dysphagia was present in 61 % of cases (with mild degree in the half of cases) and constipation in about the 76 %. Only a small subgroup presented a severe sleep disorder while the majority had only a mild impairment.
The cardiorespiratory phenotype was forced in the majority of cases. No Feeble phenotype was found [9].

Discussion
The present study was conducted on a population of individuals with Rett syndrome who regularly attended the Tuscany Rett Center. The majority of the cohort is represented by girls living in the same region, but we account a considerable number of girls, and even males with Rett syndrome, coming from various parts of Italy. Hence this is the first study conducted in Italy on a large, unbiased, group of individuals affected by this condition.
The cohort has classic cases as well as variants [4]. We account for a considerable percentage of atypical forms (35 %), greater than data reported in the literature, included between 15 and 30 % [24][25][26]. MECP2 mutations were present in almost all classical and Z-RTT cases, a notable link between these cases where clinical features slowly change from one end to another. The ARTT-NOS show different kind of genetic abnormalities in only 30 % of the cases.
All individuals fulfilled the diagnostic criteria for Rett Syndrome in the classical form or variants: all individuals had a delay in development, different degree of regression and an autistic phase. In 20 % of cases autistic traits were marked even years after diagnosis. The sample consisted of 151 individuals (all females except 2 cases), of which 132 subjects with genetic mutations: 118 with MECP2 mutations; 12 with CDKL5 mutations 2 with FOXG1 and MEF2C mutations. For 19 individuals with clinical diagnosis (12.5 %) no genetic mutation was identified.
Considering the clinical symptoms described in classic cases and in those with the preserved speech variant (Z-RTT) and the corresponding MECP2 mutations, we observed a parallel, progressive sequence in the severity of symptoms, concerning the terminal C deletion in several cases of Z-RTT variant and in a more limited number of classic cases only. Somatic features varied from a minor signs in classic cases to normal physical appearance in Z-RTT; scoliosis followed the same pattern and the same applies to the ability to walk and the use of hands.  Hand stereotypic activities ('hand washing') were intense and present all day in the waking state in classic individuals and became occasional in Z-RTT; the same applies to epilepsy as well as to cardiac and breathing abnormalities, more frequent and intense in severe cases and less pronounced in those with some speaking ability. In contrast, we observed that intensity of general anxiety is inversely correlated with clinical severity [27]. In all these cases a constant regression followed an initial period (several months) when the toddler did not show clear signs of abnormality and was usually considered to be normal by parents although abnormalities of General Movements could be detected by professionals [28,29]. In the present study Z-RTT was more frequently observed (12.6 %) than in previous studies as, for example, only 3 out of 168 individuals in Spain [30]. It has recently reported [31] that most cases of FOXG1 mutation have structural brain abnormalities including corpus callosum malformation, present also in the only congenital case of the present study. The study suggests that this is a distinct clinical entity with different features and the hypothesis is further supported by lack of scoliosis and an entirely different clinical course. In these cases symptoms appear soon after birth and, in contrast with classical and Z-RTT, don't have a regression. Some authors associate these cases to the dyskinetic encephalopathy of infancy [32]. A similar suggestion has been made for the early seizure variant (Hanelfeld) with a different clinical appearance where females present minore frequenza of scoliosis and a persistent gaze avoidance; in addition, in most girls regression occurs in the context of severe seizures. This is different from typical Rett, rare MECP2 mutations and in most case a different mutation CDKL5, data which correspond to a large extent to what has been observed in the present study [33,34].
In our sample we considered both the age of the evaluation and the age of diagnosis: this data are a measure of the time required to make the diagnosis of Rett syndrome, which is different for each clinical form [33], and not necessarily directly correlated to the age of onset of symptoms, but is linked to the complexity of the diagnosis. Important symptoms such as developmental delay, absence and / or loss of language, presence of stereotypies and eating disorders favour the early diagnosis; absence of stereotypies, presence of language, autistic traits and mild motor disorders, lead specialists to incorporate these cases into broader diagnostic categories such as mental retardation or autism and contribute to delay the correct diagnosis [26], similarly to what observed in atypical presentations [35]. In our study, the frequency of labeled autistic traits was present in the 20 % of individuals.

Autonomic nervous system (ANS)
A major contribution of this study is connected to the attention paid to disorders of the autonomic nervous system, which are present in all clinical forms, although with some variability.
Imbalances of ANS, essentially characterized by a vagal tone comparable to that of a newborn and therefore particularly immature [17] implies a constant imbalance of sympathetic parasympathetic activity in favour of the latter and determines the following symptoms: impaired swallowing, abnormalities in the gastro-esophageal tract and urinary bladder, mood disorders, sleep disorders, peripheral circulation in the limbs and abnormalities of breathing frequency.
All clinical forms presented massive respiratory dysfunctions which, although with great variability across single cases, affected 50 % of the whole registration time for the classic forms and 40 % of the Z-Rett and RTT-NOS individuals. The other two variants had higher values: 57.1 % in Hanefeld' s and 37.4 % in the congenital forms; these differences were not statistically significant.
We found cardiorespiratory abnormalities in all our subjects with classic RTT, with the exception of one single case, and two females in the ART-NOS group. The Valsalva maneuver is present in the classical forms and represents 5.5 % of all respiratory abnormalities; in the  Z-Rett the maneuver is almost always absent (0.5 %). There were no differences for HR, CVT, and blood gases that depend more on the cardiorespiratory phenotype than on the clinical form. For the CVT analysis, the study showed high values in individuals with forceful breathing (never described in literature observation). In particular, the values of vagal tone (CVT) in the individuals with forceful were closer to the control values than the CVT of the other individuals. Heart rate, on the contrary, did not show significant differences across different phenotypes [17]. The comparison of cardiorespiratory phenotypes showed higher levels of PCO2 in feeble and in apneustic phenotype than in forced phenotype. This observation has particularly important neurophysiological implications, both in the event of general anesthesia and in cases where, in the worsening of the respiratory disorder, the organism generates a pathological change of the CO2 (either an increase, or a reduction: hypo-and hypercapnia) [18].
Gastrointestinal phenotype was compromised, and severe feeding disorders required in some cases the use of some devices, like percutaneous endoscopic gastrostomy (PEG), useful for long period of time, or the nasogastric tube, for short period of time [36]. In our study, the use of PEG was limited to 3 cases.
To further define the differences in the symptoms related to ANS between the clinical forms, we considered sleep. The prevalence of sleep disorders is particularly high in Rett population, in order of 80 % [37]; daytime napping, night-time laughter, teeth grinding, screaming and night awakenings due to the appearance of seizures were the most frequently reported problems. The major problems were related to difficulty falling asleep or night-time/early morning awakenings. In our cases, the measurement time, representing 38 % was more frequently associated with Hanefeld through against the classic forms and Z-Rett.
Comparison between scoring systems ISS and PBZ: the ISS highlights the reduced impairment of Z-Rett compared to all other forms with statistically significant differences in the overall assessment, as in growth and development area, in the muscle skeletal, in the area of movement and in the cortex. However, it does not show significant differences across all other dependent forms of the Autonomic Nervous System, in contrast to what reported in literature (see Table 12).
The results obtained with the two scoring systems ISS and PBZ are comparable, and PBZ evaluation allowed an evaluation of the daily living. The PBZ is a scale of evaluation designed to test severity in Classical Rett Syndrome and its variants. The scale is susceptible to recording even minimal changes in clinical symptoms over the course of therapy, rehabilitation programmes and environmental enrichment, but also to indicating the severity of different clinical and genetic groups [14]. The evaluation with PBZ scoring system showed a worse clinical presentation for Hanefeld variant; daily living evaluation greatly contributed to the evaluation of the phenotype. Even with the PBZ system the Z-Rett was the mildest form in our cohort for all the areas considered. All the reported differences were significant in each area.
In general, the PBZ scale allows a more sophisticated classification. In fact ISS system allows only two values for defining the cognitive level and language: severe and extreme, but it does not accept that there is no sign of Rett, even though there are some females with Z-Rett, with a good speech and motor performance.
Both PBZ and ISS scales showed a better score for the auxological parameters in the Z-Rett and Hanefeld girls, for all the considered parameters (height, weight, and OFC). The BMI values were significantly different in favour of any other forms, Hanefeld included. The BMI did not correlate to the clinical form, but was associated with cardio-respiratory phenotype: this is consistent with an association between low BMI and forced phenotype because of the energy consumption in this phenotype. The data are consistent with those reported in the literature: mean weight, height, and BMI z scores in subjects with Rett syndrome were below that of their age group and decreased at steadily with age [38].
Cortical activity is clearly related to the onset and severity of epilepsy. Epilepsy was present in 64.2 % of the cases, a value close to the 68.1 % reported from the Rett network database. Drug resistant cases were less frequent: 17.2 % compared to 32.2 % of the above quoted study [39] possibly related to a comparatively greater number of Z-RTT girls where drug resistant epilepsy in unusual. Also, the onset of epilepsy occurs earlier in Hanefeld. In all the females with CDKL5 mutation presented major sleep disruptions compared to other forms: RTT versus other forms: χ 2 test = 83.399; p = 0.0001; ARTT-NOS vs Hanefeld vs classical vs congenita: χ 2 test = 4.82; p = 0.18 not significant. It is possible that in our cohort, the higher number of Z-Rett girls justify some of the differences observed between our and other classifications-for example, the resistant epilepsy described by Krajnc [40]. Finally, all clinical forms considered were indistinguishable from each other as regards the ANS involvement, but overall the scales used confirmed that the Z-RTT had milder phenotype if compared to all the other forms, and have the best quality of life as assessed with the PBZ (χ 2 test = 25.04; p < 0.01).
Genetic analysis showed that there is a correlation between gene involved and different clinical forms but there is a weak correlation between the severity of the phenotype and type of mutation, except for C-Terminal deletion that were statistically associated with clinical forms with the lowest scores of severity scales (ISS and PBZ) in accordance with what is reported in the literature [41,42].

Conclusions
The present study confirms the variability of Rett clinical phenotype and adds new autonomic and anthropometric variables to clinical data with better results in the Z-RTT girls. Autonomic abnormalities were common in the classical form and present also in a number of cases of the preserved speech variant. In this respect it must be noticed that abnormalities of this kind are common also in other neurodevelopmental disorders and have been described in Autistic Spectrum Disorders [11].
In our cohort we found a broad variety of RTT clinical forms. In the great majority of individuals with classical form and preserved speech variant there was a mutation in the MECP2 gene with a correlation between clinical evaluation and genetic analysis. Almost all cases of Hanefeld variant had a CDKL5 mutation. There was only one case of the congenital form.
An analysis of the MECP2 mutations, showed that Cterminal mutations were associated to milder clinical forms and lower scores in severity scales (ISS and PBZ), confirming what has been observed in other studies [41,42].
However, considered the spectrum of all MECP2 mutations, females with the Z-Rett variant present a milder phenotype and severity seems to be related to the early development of the disorder, suggesting that the natural history of the disease is dependent to the stage of regression: absent or reduced.
The frequency of epilepsy was not different across different clinical forms, although we did not find any patient with drug-resistant epilepsy in the Z-Rett females. Other authors report a decreased frequency of epilepsy in Z-Rett, but in our cohort the number of females with Z-Rett was higher than in other studies and this can influence the final outcome [39,40].
We did not find significant correlation between clinical forms and cardio-respiratory phenotype: however, the forced breathing phenotype was associated to a reduced BMI, suggesting that the high energetic cost of this abnormal breathing leads to a chronic hypoxia.
The breathing abnormalities are not typical of Rett syndrome, however the breathing patterns are an essential element of the diagnosis, for example, the evaluation of the ANS allows the distinction between non-epileptic paroxysm and epilepsy, contributes to the diagnosis in MECP2 negative forms, allows to design a personalized therapy [17], and eventually, taking into account all the limitations, a better prognosis.
Similarly, we did not find a significant correlation between clinical form and cardio-respiratory phenotype, however the forceful breathing phenotype was associated to high CVT values and to a reduced BMI suggesting that the high energetic cost of this abnormal breathing leads to a chronic hypoxia. The non-invasive assessment of brainstem function in Rett may contribute to life-long clinical management and might be as important as the genotype in understanding the pathology and in prognostication.