Subtle cognitive deficits in adults with a previous history of Sydenham's chorea during childhood




To evaluate the neuropsychological profile and health-related quality of life (HRQOL) of adults who had rheumatic fever (RF) during childhood with and without Sydenham's chorea (SC).


Three groups of patients were assessed: adults who had RF with SC during childhood (SC group), adults who had RF without SC during childhood (RF group), and controls (CT group). A range of neuropsychological tests looked at several cognitive domains. HRQOL was measured through a Brazilian version of the Short Form 36 (SF-36) health survey.


Twenty patients were included in the SC group, 23 patients in the RF group, and 19 patients in the CT group. The 3 groups were homogeneous regarding sex (P = 0.078), age (P = 0.799), schooling (P = 0.600), socioeconomic status (P = 0.138), intelligence quotient (P = 0.329), and scores for anxiety (P = 0.156) and depression (P = 0.076). The SC group demonstrated inferior performance in tests that assessed attention (Digit Span Forward [P = 0.005], Corsi Block Forward [P = 0.014]), speeded information processing (Trail Making A [P = 0.009], Symbol Search [P = 0.042]), and executive functions and working memory (Corsi Block Backward [P = 0.028]), and higher scores for attention deficit scale (P = 0.030) when compared with the RF and CT groups. They also showed a tendency toward lower scores in the physical aspects, vitality, emotional aspects, and mental health domains of the SF-36. The RF group had a lower score for the general health domain than the CT group (P = 0.030).


Patients who had SC during childhood can exhibit inferior performance in tasks that evaluate attention, speeded information processing, executive functions, and working memory in adult life. Therefore, there is indirect evidence of the persistence of dysfunction in cerebral circuits involved with the basal ganglia. They also presented a worse self-evaluation in HRQOL that was not related to cognitive impairments.


Sydenham's chorea (SC) is the most common form of acquired chorea during childhood (1), and is present in ∼30% of patients with rheumatic fever (RF) (2–4). Clinically, SC is characterized by motor manifestations such as involuntary movements of the extremities, muscular hypotonia, dysarthria, gait disturbance, and tics (5, 6), as well as behavioral manifestations, such as obsessive-compulsive symptoms, attention deficit disorder, and emotional lability (7, 8).

Some studies suggest that its etiopathogenesis is related to immunologic dysfunction located at the basal ganglia, mediated by antibodies that cross-react against antigens in these structures after a group A streptococcus pharyngitis infection (9, 10). Autopsy findings (11) and studies of anatomic (12, 13) and functional (14) neuroimaging show involvement of the basal ganglia and their different frontostriatal circuits.

The basal ganglia are subcortical structures that are related to the brain cortex through the frontostriatal circuits. The motor and oculomotor circuits are responsible for the production of somatic and eye movements, respectively. The orbitofrontal circuit plays an important role in the suppression of inappropriate behavior, the anterior cingulate circuit mediates motivated behavior, and finally, the dorsolateral circuit is involved in the construction of cognitive functions, particularly attention, working memory, and executive functions (15). One can hypothesize that the selective involvement of the different frontostriatal circuits provides evidence of the variability of neuropsychiatric manifestations in SC patients (16). Teixeira et al developed and validated a rating scale for clinical evaluation of SC patients and observed a poor correlation between behavior, activities of daily living, and motor function items, suggesting the selective involvement of the different circuits (16).

The few studies regarding cognitive functions in patients with SC were developed during the acute phase of the disease and showed cognitive deficits in abilities such as attention, executive functions, and phonemic verbal fluency (17–19). Therefore, there is insufficient evidence on the repercussion of cognitive impairment associated with SC in adulthood. The purpose of the current study was to evaluate the neuropsychological profile and health-related quality of life (HRQOL) of adult patients who had SC during childhood or adolescence and now are asymptomatic. A battery of standardized neuropsychological tests and specific questionnaires were applied in order to investigate possible cognitive damage resulting from this involvement.


Patients and inclusion criteria.

Three groups were evaluated: adults who presented with RF with SC during childhood (SC group), adults who presented with RF without SC during childhood (RF group), and a control group (CT group) constituted by apparently healthy individuals. All patients had a minimum followup time of 6 months at the pediatric rheumatology outpatient unit of the Federal University of São Paulo (UNIFESP). The 1992 update of the Jones criteria was used to diagnose RF and SC (20). For the patients with an SC diagnosis, other possible causes of chorea in childhood, such as juvenile Huntington's chorea, benign hereditary chorea, drug-induced chorea, and chorea due to autoimmune disorders were excluded (1). The RF group of patients was included in the study because, although they had not presented with chorea, it has been previously reported that ∼13% of such patients can develop anti–basal ganglia antibodies (9). They were also used as positive controls in order to discard the possible influence of underlying disease in the results.

The participants in the CT group were adults who did not meet diagnosis criteria for neuropsychiatric disorder and had not been using psychoactive drugs in the last 3 months, all matched for age and sex according to the patients in the other groups. This study was approved by the UNIFESP Research Ethics Committee, and written informed consent was obtained from each subject.

Inclusion criteria included 1) age ≥18 years, 2) last episode of SC ≥5 years prior to the study, 3) absence of other neurologic (detected by physical examination) and/or psychiatric disease in childhood besides SC, 4) absence of psychoactive drugs, illicit drugs, or alcoholism (detected by interview), and 5) absence of severe carditis.


The study was performed in 2 phases, the first a retrospective and the second a cross-sectional phase. Medical records were reviewed and an initial contact by letter or telephone was made explaining the protocol and inviting the patient to participate in the study. If the invitation was accepted, a written informed consent was obtained and the patient was evaluated by a pediatric rheumatologist (AC; in order to exclude any neurologic and/or clinical disorder) and by a neuropsychologist (SAPB). Of 90 patients contacted, 43 agreed to participate in the study; the other 47 were not able to participate due to work-related difficulties. The same pediatric rheumatologist recruited the individuals of the CT group from the triage room located at the emergency unit of the UNIFESP hospital. They usually presented with minor symptoms (such as fever, cough, skin lesions) and were assessed in a scheduled appointment afterward to be certain that they had no symptoms that could interfere with their performance.

Clinical, demographic, and socioeconomic assessment.

The rheumatologic protocol included the following clinical and demographic data: sex, current age, age at RF onset, major Jones criteria (20), number of SC episodes, and duration and severity of SC according to clinical assessment classified as mild (minimal movements), moderate (movements that are inconvenient for the patient but that do not interfere with his or her personal care), or severe (incapacitating movements requiring the patient to solicit help in conducting daily activities) (21). Socioeconomic status was evaluated using the Brazilian Association of Marketing Research Companies (ABEP) questionnaire, which classifies Brazilian families in classes A (highest), B, C, D, and E (lowest) according to possession of assets (such as a television, radio, washing machine, car, and so on) and the educational level of the head of the family (22). Patients and controls also provided information on their working situation (employed or unemployed) and schooling (years of education). These data were determined during the interview and/or by chart review.

Neuropsychological assessment.

Subsequently, all subjects were assessed with a battery of standardized neuropsychological tests, applied by experienced neuropsychologists (FHdS and SAPB), without previous knowledge of the participant's diagnosis. The following tests were undertaken: 1) intellectual level (estimated by intelligence coefficient [IQ] using the vocabulary and block design subtests of the Wechsler Adult Intelligence Scale, Third Edition) (23), 2) memory (logical memory subtest of the Wechsler Memory Scale, Third Edition) (24), 3) attention and speeded information processing (Trail Making A, Digit Span Forward, Symbol Search, and Corsi Block Forward), 4) language (Verbal Semantic Fluency and Vocabulary), 5) visuospatial (Block Design), and 6) executive functions and working memory (Trail Making B, Verbal Phonemic Fluency, Clock Drawing Test, Letter-Number Sequencing, Digit Span Backward, and Corsi Block Backward). These tests have been described in detail by Lezak et al (25) and Strauss et al (26).

Emotional, behavioral, and HRQOL assessment.

Emotional state was evaluated using the Brazilian version of the Hospital Anxiety and Depression Scale (HADS), which comprises 2 subscales, one for anxiety and another for depression (27). The score for each subscale ranges from 0 to 21. To evaluate subjective symptoms related to attention deficit hyperactivity disorder (ADHD), the Brazilian version of the Adult Self-Report Scale (ASRS) was applied (28). This scale has 18 items (9 for attention deficit and 9 for hyperactivity) and contemplates the symptoms listed in the A criteria on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) from the American Psychiatric Association for the diagnosis of ADHD (29). The higher frequency answers (often and very often) were considered to be clinically significant symptoms. The HRQOL was measured using the Brazilian version of the generic questionnaire Medical Outcomes Study Short Form 36 (SF-36) (30). It is a self-administered questionnaire with 8 domains: physical functioning, vitality, social functioning, general health, pain, physical role, emotional role, and mental health. In this instrument, the subjects received a score in each domain ranging from 0 to 100, where 0 = the worst score and 100 = the best score. Higher scores on the SF-36 scales correspond to better health status.

Statistical analysis.

The SPSS software for Windows, version 15 (SPSS), was used for statistical analysis (31). Descriptive analysis of the demographic and clinical data was conducted. The chi-square test was used to compare categorical variables between the groups. The Kolmogorov-Smirnov test assessed normal distribution of continuous variables. Analysis of variance was performed to compare continuous variables with normal distribution among the groups, followed by Tukey's post hoc test. Kruskal-Wallis and Mann-Whitney tests were used to compare variables that were not normally distributed. P values less than 0.05 were considered statistically significant. In addition, for the neuropsychological tests the effect size (f2) was used to estimate the relationship power between the variables. By convention, f2 effect sizes of 0.02, 0.15, and 0.35 are considered small, medium, and large, respectively (32).


Participant clinical, demographic, and socioeconomic assessment.

Of 62 participants enrolled in the study, 20 belonged to the SC group, 23 to the RF group, and 19 to the CT group. The 3 groups were homogeneous regarding sex (P = 0.078) and current age (P = 0.799), although there was a predominance of women over men in all groups. The patients in the SC group were more likely to have carditis than arthritis, whereas the patients in the RF group had more arthritis than carditis. No patient presented the association of arthritis and chorea. The mean ± SD duration of SC was 4.3 ± 2.8 months. In these patients, there was a predominance of the moderate form (75%), and most patients (65%) presented only 1 chorea episode. No statistically significant difference between the 3 groups regarding the socioeconomic status by the ABEP questionnaire (P = 0.138) and schooling level (P = 0.600) was observed. Most of the participants (SC, RF, and CT groups) had stable employment at the time of the study. Clinical, demographic, and socioeconomic characteristics of the sample are shown in Table 1.

Table 1. Clinical, demographic, and socioeconomic characteristics of patients with RF with or without SC (the SC group and RF group, respectively) and the control (CT) group*
 SC group (n = 20)RF group (n = 23)CT group (n = 19)P
  • *

    RF = rheumatic fever; SC = Sydenham's chorea; NA = not applied; ABEP = Brazilian Association Marketing Research Companies.

  • Neither patient presented erythema marginatum or Maynet's nodules.

Sex, no. of women/men18/214/915/40.078
Age at disease onset, mean ± SD years10 ± 3.29.6 ± 2.4NA
Current age, mean ± SD years24.8 ± 3.224.2 ± 3.624.1 ± 3.20.799
Major Jones criteria    
 Pure chorea80NA
 Carditis mild/moderate54NA
 Arthritis and carditis710NA
Chorea duration, mean ± SD months4.3 ± 2.8NANA
Chorea severity    
Number of SC attacks    
ABEP classification, no. (%)    
 A + B7 (35)4 (17.4)7 (36.8)0.138
 C9 (45)18 (78.3)11 (57.9)0.138
 D + E4 (20)1 (4.3)1 (5.3)0.138
Education, mean ± SD years12.0 ± 2.311.7 ± 2.111.8 ± 3.30.600
Current employment, no. (%)    
 Employed16 (80)19 (82.6)15 (79)
 Unemployed4 (20)4 (17.4)4 (21)

Participant neuropsychological assessment.

The neuropsychological tests were applied a mean ± SD of 14.8 ± 3.3 years after the first episode of chorea and 14.6 ± 2.4 years after the first episode of RF. No statistically significant difference regarding the estimated IQ results between the 3 groups was detected (P = 0.329). We observed that patients from the SC group needed more time to execute the Trail Making A when compared with the RF group (P = 0.017) and the CT group (P = 0.022). Patients from the SC group also presented inferior scores in the Digit Span Forward, Symbol Search, and Corsi Block Forward when compared with the CT group (P = 0.037, P = 0.018, and P = 0.014, respectively). In addition, in the Corsi Block Backward, the SC group presented an inferior score in relation to the RF group (P = 0.009). With regard to the effect size, we observed that in these tests, the association magnitude was considered large (f2 = 0.57–0.85). No difference between the groups was observed in the other neuropsychological tests applied. All of these results are shown in Table 2.

Table 2. Neuropsychological tests applied to patients with RF with or without SC (the SC group and RF group, respectively) and the control (CT) group*
Neuropsychological testsSC group (n = 20)RF group (n = 23)CT group (n = 19)PPf2
  • *

    Values are the mean ± SD unless otherwise indicated. RF = rheumatic fever; SC = Sydenham's chorea.

  • Post hoc.

  • Cohen.

  • §

    P < 0.05.

Intelligence quotient85.7 ± 9.185.0 ± 9.181.8 ± 4.90.329  0.693
 Immediate logical memory20.1 ± 6.720.6 ± 4.624.7 ± 6.70.054  0.588
 Delayed logical memory17.9 ± 6.918.3 ± 4.821.1 ± 6.20.191  0.574
Attention and speeded information processing       
 Trail Making A52.0 ± 16.739.9 ± 13.739.7 ± 10.50.009§0.0220.0170.721
 Digit Span Forward5.2 ± 1.34.8 ± 1.56.5 ± 2.00.005§0.037 0.852
 Symbol Search21.1 ± 9.123.4 ± 9.828.5 ± 9.90.042§0.018 0.579
 Corsi Block Forward4.0 ± 1.95.3 ± 1.85.8 ± 2.10.014§0.014 0.752
 Verbal Semantic Fluency27.0 ± 8.428.3 ± 7.831.1 ± 7.90.278  0.642
 Vocabulary33.7 ± 11.330.4 ± 11.632.3 ± 11.30.651  0.750
 Block design23.3 ± 7.323.3 ± 9.122.0 ± 8.50.856  0.872
Executive functions and working memory       
 Clock Drawing Test7.1 ± 2.67.1 ± 2.46.6 ± 2.20.593  0.648
 Letter-Number Sequencing7.7 ± 2.47.3 ± 2.58.5 ± 2.80.375  0.665
 Trail Making B96.8 ± 29.885.1 ± 30.3102.1 ± 46.90.331  0.398
 Verbal Phonemic Fluency15.3 ± 3.416.3 ± 1.215.8 ± 4.70.662  0.789
 Digit Span Backward4.7 ± 1.44.8 ± 0.75.4 ± 2.60.591  0.785
 Corsi Block Backward5.4 ± 1.97.0 ± 1.76.4 ± 1.90.028§ 0.0090.645

Participant emotional, behavioral, and HRQOL assessment.

The 3 groups were homogeneous in the HADS for anxiety (P = 0.156) and for depression (P = 0.076). The SC group had higher scores on the depression subscale than the other groups, but their scores are far from the cutoff of the test, which is not suggestive of depression in the sample. With regard to the ASRS questionnaire, we observed that the SC group presented a statistically significant higher score in the attention deficit domain when compared with the CT group (P = 0.023). In the hyperactivity domain, although no difference had been observed between the groups, both the SC and RF groups had a tendency toward higher scores when compared with the CT group. In the HRQOL assessment, the only domain that showed a statistically significant difference between the groups was that of general health, in which the RF group had a lower score than the CT group (P = 0.030). In the other SF-36 questionnaire domains, there was no statistically significant difference between the 3 groups. However, the SC group had a tendency toward lower scores, especially in the physical role, vitality, emotional role, and mental health domains, followed by the RF group and then the CT group. Emotional, behavioral, and HRQOL data are shown in Table 3.

Table 3. Emotional and behavioral aspects and HRQOL of patients with RF with or without SC (the SC group and RF group, respectively) and the control (CT) group*
 SC group (n = 20)RF group (n = 23)CT group (n = 19)P
  • *

    Values are the mean ± SD score unless otherwise indicated. HRQOL = health-related quality of life; RF = rheumatic fever; SC = Sydenham's chorea; HADS = Hospital Anxiety and Depression Scale; ASRS = Adult Self-Report Scale; SF-36 = Short Form 36.

  • P < 0.05.

 Anxiety7.0 ± 4.26.6 ± 3.34.9 ± 2.70.156
 Depression5.4 ± 3.44.2 ± 3.33.0 ± 2.00.076
 Attention deficit1.8 ± 2.31.0 ± 1.50.3 ± 0.60.030
 Hyperactivity2.0 ± 2.02.0 ± 2.01.3 ± 1.40.353
 Functional capacity85.7 ± 14.489.3 ± 14.792.8 ± 12.40.182
 Physical aspects85.9 ± 19.191.3 ± 14.694.0 ± 14.90.168
 Pain70.2 ± 22.370.3 ± 21.976.3 ± 22.30.591
 General health state70.8 ± 21.565.1 ± 21.181.2 ± 15.90.037
 Vitality61.5 ± 23.565.4 ± 25.277.6 ± 16.10.070
 Social aspects85.2 ± 21.286.0 ± 15.582.3 ± 22.80.889
 Emotional aspects81.7 ± 21.985.9 ± 19.389.1 ± 18.40.511
 Mental health63.2 ± 25.271.5 ± 17.075.7 ± 12.60.599


In this study, we evaluated the neuropsychological profile and HRQOL of adults who had RF, with and without SC, during childhood. The 2 groups of patients (SC and RF) and the CT group were homogeneous with regard to sex, age, schooling, socioeconomic status, IQ, and a specific questionnaire for anxiety and depression, which allowed us to conclude that possible cognitive deficits cannot be explained by these factors. Our results showed that patients from the SC group presented, after an average time period of 14.8 years since the first episode of chorea, weaker performance in tests that evaluate attention (Digit Span Forward and Corsi Block Forward), speeded information processing (Trail Making A and Symbol Search), and executive functions and working memory (Corsi Block Backward) when compared with RF and CT groups, suggesting a possible dysfunction at the basal ganglia circuits, specifically the dorsolateral circuit.

There are insufficient published studies about cognitive functions in patients with SC both in the acute phase as well as in the remission phase of chorea. However, in other disorders associated with the basal ganglia, cognitive deficits resulting from dysfunction of the frontostriatal circuits have been demonstrated. Neuropsychological studies in individuals with Parkinson's disease (33), Huntington's chorea (34, 35), and obsessive-compulsive disorder (36) reported similar deficits in cognitive functions such as attention, working memory, and executive functions. Therefore, independently of the etiopathogenesis of basal ganglia dysfunction, these patients presented with similar neuropsychiatric manifestations and cognitive impairment.

With regard to the studies conducted with RF and SC, all of them have been concerned with the acute phase. Sacks and colleagues evaluated, through neuropsychological tests, patients with RF and SC with and without disease activity and patients with RF without SC and without disease activity, and they did not observe significant differences with regard to IQ (37). In the same way, we also have not found this involvement, corroborating that SC seems to not compromise patients' IQ.

We observed that patients from the SC group presented lower scores when compared with the patients from the RF and CT groups in the Symbol Search and Corsi Block Forward tests, which require visuospatial tracking of information. Bird et al, although having used the Bender gestalt test (which predominantly evaluates organic involvement), also observed alteration in the visuomotor function in patients with SC after an average followup time of 8 years, in comparison with controls (38).

It is interesting to observe that studies conducted during the acute phase of SC revealed attention (response inhibition and divided attention) (17) and executive function impairment, evaluated through Tower of Hanoi problem solving (18). Although these studies used different tests from the ones used in the present study, their results suggest that patients with SC can present impairment in executive functions and in attention during the acute phase of the disease, which, according to our results, can persist into adult life in some individuals.

On the other hand, in another study conducted during the acute phase of SC that evaluated semantic and phonemic verbal fluency in 20 children, the authors observed a significant reduction in phonemic fluency in these patients when compared with the control group, without correlation with dysarthria and chorea severity. However, no difference in semantic fluency was observed between the groups (19). In the current study, we did not observe a difference between patients and controls in either semantic or phonemic fluency, which might suggest that some of the alterations found in the acute phase may not persist during the disease course.

Besides the cognitive alterations, the patients from the SC group also presented a higher frequency of symptoms that suggest ADHD through the ASRS questionnaire in comparison with the CT group. As for hyperactivity, although we have not found a statistically significant difference between the 3 groups, there was a tendency in the SC and RF groups to present higher scores when compared with the CT group. Maia et al demonstrated that ADHD is more frequent in patients with SC when compared with controls and patients with RF without SC (8). Although the objective of our study was only a screening of ADHD symptoms by means of the ASRS (28), which takes as reference the A criteria of DSM-IV for ADHD (29) and not the clinical diagnosis of ADHD, we found results similar to the ones found by Maia et al (8). Moreover, such subjective indicators corroborated the performance in objective measures in the attention tests of our protocol.

With regard to the occupational aspects, our findings did not show a higher rate of unemployment in RF patients, with or without SC, when compared with controls. Schooling and socioeconomic status were also similar for all of the groups. This suggests that cognitive changes and attention deficit symptoms found in our patients who had SC during childhood have not impacted, up to the present time, their professional performance during adult life in a significant way.

Concerning the HRQOL, although we found a statistically significant difference only in the general health domain of the SF-36 between the RF group and the CT group (P = 0.030), a detailed analysis of our results (Table 3) shows that patients with SC presented a tendency to have worse scores in other SF-36 domains, such as physical aspects, vitality, emotional aspects, and mental health. We suppose that patients with RF presented lower scores in general health (SF-36) probably because they grew up with frequent medical visits and continuous medication due to their condition. In view of our small sample size, it is possible that we did not find statistical significance in the other SF-36 domains. In summary, adult patients who had SC during childhood presented a worse self-evaluation of quality of life than the patients with RF without SC and the controls in the CT group.

The current study has some limitations, such as the difficulty in recruiting patients because they were no longer being treated in our outpatient clinic, with many of them no longer living in the same city and others having moved to another address. Typically, the majority of patients with RF are discharged or discontinue treatment, especially in the cases with no cardiac sequel. Therefore, the sample size was reduced. Another limitation of our study is that it was not a prospective investigation; that is, we built on the observations of other studies that showed cognitive changes during the acute phase of SC and verified with our results that these changes persist throughout adult life. However, the present result is important because it shows the need for a longitudinal followup of children who have SC, taking into consideration the risk for cognitive and behavioral deficits that can persist into adult life.

In summary, adult patients who had SC during childhood or adolescence can present lower results in tests that evaluate working memory, attention, and executive functions. These findings are probably a result of a dorsolateral circuit dysfunction and are not related to sex, age, education, socioeconomic status, mood disorder, and IQ.

Although prospective studies with a larger number of patients tracked from the acute phase of SC until their adult life are necessary, our results permit us to conclude that patients with SC should be assessed by a neuropsychologist in the acute phase of the disease and in the followup in order to determine the existence of cognitive and behavioral changes and to offer appropriate treatment, if necessary. We must consider that the neuropsychological rehabilitation that is conducted early on can minimize this impairment and reduce the difficulties that some patients might possibly have during adolescence or adulthood.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Hilário had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Cavalcanti, Hilário, dos Santos, Bolognani, Len.

Acquisition of data. Cavalcanti, Bolognani.

Analysis and interpretation of data. Cavalcanti, Hilário, dos Santos, Bolognani, Bueno, Len.


The authors wish to thank the following neuropsychologists who participated in the study: Rene Viana, Daniela Landucci, and Debora Drumond.