Patients and diagnosis

Participants were consecutively recruited from people admitted as in-patients of the Department of Psychiatry and Psychotherapy at the University of Lübeck. Each participant gave written informed consent after having been carefully informed about the study. The study was approved by the local ethics committee. Participants had to meet the following inclusion criteria:

1. (a) stable neuroleptic medication for at least 1 week;

2. (b) no neurological disease;

3. (c) no significant history of head trauma;

4. (d) no other medication that could potentially induce movement disorder.

Operational psychiatric lifetime diagnoses according to DSM–IV (American Psychiatric Association, 1994) were established using the German version of the Mini-International Neuropsychiatric Interview (MINI; Reference Sheehan, Lecrubier and SheehanSheehan et al, 1998) and the Structured Clinical Interview for Personality Disorders (SCID–II; Reference Spitzer and WilliamsSpitzer & Williams, 1987). Diagnoses were divided into four categories: organic psychiatric disorders, including all forms of alcohol and drug dependence; psychotic disorders; affective disorders; and other Axis I or Axis II disorders. Psychopathological symptom severity was rated on the Brief Psychiatric Rating Scale (BPRS; Reference Overall and GorhamOverall & Gorham, 1962). Each participant underwent a complete neurological examination with particular emphasis on specific signs of acute dystonic reaction, parkinsonism, akathisia and tardive dyskinesia. The following rating scales were used to assess the severity of extrapyramidal symptoms (Reference van Harten, Hoek and Matroosvan Harten et al, 1997): the Abnormal Involuntary Movement Scale (AIMS; National Institute of Mental Health, 1975); the Tsui Rating Scale for Cervical Dystonia (Reference Tsui, Eisen and StoesslTsui et al, 1986) and the Burke Rating Scale for Primary Torsion Dystonias (Reference Burke, Fahn and MarsdenBurke et al, 1985); Part III of the Unified Parkinson's Disease Rating Scale (UPDRS; Movement Disorder Society Task Force on Rating Scales for Parkinson's Disease, 2003); the global score on the Hillside Scale (Reference Fleischhacker, Miller and SchettFleischhacker et al, 1991) and the Barnes Akathisia Scale (Reference BarnesBarnes, 1989). History of extrapyramidal symptoms on previous neuroleptic treatment was explored in a structured interview developed by our group, covering the typical symptoms of acute dystonic reaction, parkinsonism, akathisia and tardive dyskinesia. To check the reliability of their answers, participants were asked about the consequences of the development of their movement disorder (for example, change in medication or administration of biperiden). In addition, clinical records covering former treatment phases were reviewed to complete the information. Family history of movement disorders was obtained by means of a structured interview developed in our department, specifically covering symptoms of Parkinson's disease (stiffness of movement, gait problems, tremor, change in facial expression, lateralisation of symptoms), dystonia and psychiatric disorders in first-degree to third-degree relatives. This structured interview had previously been shown to provide reliable data in a large epidemiological study (Reference Klein, Vieregge and HagenahKlein et al, 1999). A diagnosis of a primary movement disorder was only made if the criteria for the disorder were clearly fulfilled by the symptom description given by the patient. Relatives were asked to also undergo the structured interview whenever possible.

Statistical procedure

Pearson's chi-squared tests for independence were performed to investigate the relationship between categorical outcome variables (lifetime occurrence of extrapyramidal symptoms, including reported and currently observed symptoms; reported extrapyramidal symptoms during previous treatment phases; currently observed extrapyramidal symptoms; extrapyramidal symptoms on typical neuroleptic medication; extrapyramidal symptoms on atypical neuroleptic medication; lifetime occurrence and currently observed acute dystonic reaction, parkinsonism, akathisia and tardive dyskinesia) and possible categorical predictor variables (gender; age; psychiatric diagnostic category; dosage range; duration of exposure to any neuroleptic medication; duration of exposure to typical neuroleptic medication; duration of exposure to atypical neuroleptic medication; positive family history of primary movement disorder). To achieve the required expected cell frequency of more than 5 in χ²-tests, we defined three different age groups: 18–40 years, 41–60 years and >60 years. For the same reason, duration of exposure to neuroleptic medication was categorised as <6 months, 6 months to 5 years, and >5 years. For easier comparison, drug dosages were defined as low, medium or high, based on current clinical practice. Chi-squared values are reported with two-tailed probabilities. Relationships between possible predictor and outcome variables with P<0.5 revealed by χ²-tests were entered in a step-wise logistic regression analysis to identify predicted probabilities for extrapyramidal symptoms (probability to enter at 0.05). By use of a logistic regression analysis, all predictor variables are considered within one testing procedure, including intercorrelations, which reduces the probability of type I errors. All statistical procedures were performed using the Statistical Package for the Social Sciences (version 11.0).

Lifetime prevalence of extrapyramidal symptoms

Lifetime prevalence of extrapyramidal symptoms, including both reported and currently observed symptoms, was 65% (Table 2). Acute dystonic reactions occurred most commonly (41%), followed by parkinsonism (37%), akathisia (19%) and tardive dyskinesia (4%). It should be noted that several patients suffered from more than one type of extrapyramidal symptom. Of those who had experienced such symptoms, more patients had been exposed to typical than to atypical medication. Details of mean age and distribution of gender are given in Table 2.

Table 2 Clinical characteristics and lifetime prevalences of extrapyramidal symptoms

	n	Age, years	Gender (male/female)	Duration of exposure
		Mean (s.d.)	n/n	<6 months n	6 months to 5 years n	<5 years n
Lifetime prevalence of
Any EPS	65	37.6 (14.7)	34/31
Acute dystonic reaction	41	35.6 (13.8)	20/21
Parkinsonism	37	38.0 (13.5)	20/17
Akathisia	19	35.8 (13.8)	11/8
Tardive dyskinesia	4	36.8 (11.0)	3/1
Symptoms experienced under exposure to
Any neuroleptic	65			15	30	20
Typical neuroleptic	47	38.7 (15.2)	25/22	15	17	15
Atypical neuroleptic	34	37.3 (15.9)	17/17	13	16	5

Currently observed extrapyramidal symptoms

At the time of examination, most patients were taking an atypical neuroleptic medication (Table 3). Extrapyramidal symptoms were diagnosed in 34% of patients. The most commonly observed symptom was parkinsonism (19%), followed by acute dystonic reaction (15%), akathisia (5%) and tardive dyskinesia (3%). Again, some patients were diagnosed with more than one type of extrapyramidal symptom. Table 3 gives further clinical characteristics and details concerning dosages of neuroleptic medication.

Table 3 Clinical characteristics and prevalences of extrapyramidal symptoms at the time of examination

	n	Age, years Mean (s.d.)	Gender (male/female) n/n	Symptom score Mean (s.d.)	Neuroleptic dosage
					Low n	Medium n	High n
Current exposure to
Typical neuroleptic	14	44.4 (13.9)	6/8	39.8 (10.9)1	4	7	3
Atypical neuroleptic	78	39.2 (14.9)	36/42	38.1 (10.9)1	18	41	19
Both	8	39.2 (14.7)	6/2	41.3 (12.5)1	0	4	4
Currently observed EPS
Any EPS	34	41.1 (17.9)	16/18	40.0 (10.2)1	6	21	7
Current EPS under exposure to
Typical neuroleptic	4	56.8 (19.2)	1/3	42.8 (2.6)1	2	2	0
Atypical neuroleptic	25	39.5 (17.6)	11/14	38.6 (10.7)1	4	17	4
Both	5	36.4 (14.3)	4/1	46.0 (10.6)1	0	2	3
Acute dystonic reaction	15	38.0 (20.0)	7/8		2	10	3
Burke scale
Severity				6.8 (8.3)
Handicap				3.4 (2.7)
Tsui scale				3.8 (4.6)
Parkinsonism	19	43.4 (18.3)	9/10		2	12	5
UPDRS				29.8 (19.1)
Akathisia	5	33.0 (4.7)	3/2		2	1	2
Hillside scale				3.0 (1.6)
Barnes scale				10.0 (7.5)
Tardive dyskinesia	3	31.7 (5.1)	2/1		–	2	1
AIMS				5.0 (2.8)

Reported symptoms and family history of primary movement disorders

Fifty-three patients (29 men and 24 women, mean age 36.3 years, s.d.=13.2) reported that they had experienced extrapyramidal symptoms during previous treatment phases. Information about the family history of primary movement disorders was available for 98 participants and could be assessed for a total of 1316 relatives, 438 of whom were first-degree relatives. Thirty-two of these 98 patients (20 men and 12 women) reported a positive family history, resulting in a total of 47 affected relatives, or a prevalence of a primary movement disorder of 3.5% among all relatives. Specifically, the prevalences were 1.1% (n=14) for Parkinson's disease, 1.6% (n=21) for tremor and 0.9% (n=2) for dystonia. Among relatives of patients with lifetime extrapyramidal symptoms (n=848), the prevalences were 1.4% (n=12) for Parkinson's disease, 2.4% (n=20) for tremor and 1.1% (n=9) for dystonia, whereas among the relatives of patients without lifetime extrapyramidal symptoms (n=468), the prevalences were 0.4% (n=2) for Parkinson's disease, 0.2% (n=1) for tremor and 0.6% (n=3) for dystonia. A subgroup analysis was performed using data from 27 (2.1%) first-degree relatives of 21 patients who were seen personally by one of the investigators. In all of them, the presence or absence of a primary movement disorder was confirmed as established by the family history interview (25 relatives without a primary movement disorder, 1 with dystonia and 1 with tremor). In order to avoid multiple inclusion of patients with more than one affected relative in further statistical analyses, we considered only the closest relative of those patients. This resulted in 12 patients with a relative with Parkinson's disease (5 first-degree relatives and 7 second-degree relatives), 11 patients with a relative with tremor (6 first-degree relatives and 5 second-degree relatives) and 9 patients with a relative with dystonia (7 first-degree relatives and 2 second-degree relatives).

Relationship between possible predictor and outcome variables

Family history of primary movement disorder

The family history of primary movement disorders was related to lifetime prevalence of extrapyramidal symptoms (EPS) (χ²=8.35, d.f.=1, P<0.01), currently observed EPS (χ²=8.05, d.f.=1, P<0.01), prevalence of reported EPS (χ²=6.75, d.f.=1, P<0.01) and lifetime prevalence of acute dystonic reaction (χ²=4.69, d.f.=1, P=0.03). Table 4 shows that the prevalence of these four related outcome variables was higher in participants with a positive family history than in patients with a negative history. Furthermore, lifetime prevalence of acute dystonic reaction was related to the subtype of primary movement disorders occurring in relatives (χ²=8.27, d.f.=3, P=0.04). Lifetime acute dystonic reaction occurred in 7 of 9 (78%) patients with a family history of dystonia, but only in 7 of 12 (58%) patients with a family history of Parkinson's disease, 4 of 11 (36%) patients with a family history of tremor and 22 of 44 (50%) patients with a negative family history.

Table 4 Distributions of outcome and possible predictor variables

Possible predictor	Outcome variable
	Lifetime EPS (yes/no) n/n	Reported EPS (yes/no) n/n	Currently observed EPS (yes/no) n/n	EPS with typical neuroleptic (yes/no) n/n	Lifetime ADR (yes/no)	Lifetime parkinsonism (yes/no)	Lifetime akathisia (yes/no)
Family history1,2				P>0.05		P>0.05	P>0.05
Positive	27/5	23/9	17/15		18/14
Negative	36/30	29/37	16/50		22/44
Age				P>0.05		P>0.05	P>0.05
18–40 years	46/11	40/17			31/26
41–60 years	12/18	9/21	P>0.05		7/23
> 60 years	7/6	4/9			3/10
Lifetime exposure to any neuroleptic2			P>0.05		P>0.05
<6 months	15/23	6/32				10/28	2/36
6 months to 5 years	30/8	27/11		P>0.05		13/25	9/29
> 5 years	20/4	20/4				14/10	8/16
Lifetime exposure to typical neuroleptic2			P>0.05
<6 months	17/10	14/13		15/12	11/16	6/21	5/22
6 months to 5 years	18/1	16/3		17/2	12/7	7/12	6/13
> 5 years	15/3	15/3		15/3	10/8	13/5	6/12
No exposure	15/21	8/28			8/28	11/25	2/34
Lifetime exposure to atypical neuroleptic2	P>0.05		P>0.05	P>0.05	P>0.05	P>0.05	P>0.05
<6 months		14/25
6 months to 5 years		28/11
> 5 years		7/2
No exposure		4/9

Predictors revealed by logistic regression analysis

We found that lifetime occurrence of EPS (yes or no) could be correctly predicted in 74% of all cases by knowing the duration of treatment with typical neuroleptics, and family history of primary movement disorders. Lifetime occurrence of acute dystonic reaction (yes or no) could be predicted in 65% of cases taking into account the exposure to typical neuroleptic medication and age. Both lifetime occurrence of parkinsonism (yes or no) and of akathisia (yes or no) were predicted correctly in 72% and 82%, respectively, by the duration of typical neuroleptic medication. Reported EPS (yes or no), regardless of the subtype, were correctly predicted in 82% of cases by the duration of exposure to typical neuroleptic medication, age and family history of primary movement disorders. Currently observed EPS (yes or no) were predicted in 68% of all cases by the family history of primary movement disorders. Extrapyramidal symptoms on typical neuroleptic medication (yes or no) could be predicted in 73% when considering the duration of exposure. As can be seen from the regression coefficients B in Table 5, the probability for the occurrence of symptoms always increased with the duration of exposure to neuroleptic medication, younger age and positive family history.

Table 5 Step-wise logistic regression analysis revealing significant predictor variables for occurrences of extrapyramidal symptoms (d.f.=1 in all tests)

EPS	Predictor	B	s.e.	Wald	P
Lifetime occurrence of EPS	Duration of typical neuroleptic	0.78	0.25	10.65	<0.01
	Age	– 0.85	0.35	5.95	0.02
	Positive family history	1.30	0.60	4.67	0.03
Lifetime occurrence of acute dystonia	Duration of typical neuroleptic	0.54	0.20	7.16	<0.01
	Age	– 0.88	0.36	6.17	0.01
Lifetime occurrence of parkinsonism	Duration of typical neuroleptic	0.53	0.20	7.16	<0.01
Lifetime occurrence of akathisia	Duration of typical neuroleptic	0.63	0.24	6.81	<0.01
Reported EPS	Duration of any neuroleptic	1.98	0.41	23.28	<0.01
	Age	– 1.30	0.45	8.54	<0.01
	Positive family history	1.45	0.62	5.57	0.02
Currently observed EPS	Positive family history	1.27	0.46	7.69	<0.01
EPS under typical neuroleptics	Duration of typical neuroleptic	0.90	0.40	4.97	0.03

Prevalence rates for extrapyramidal symptoms and primary movement disorders

Based on a sample that represents a typical cohort of in-patients with acute psychotic symptoms, we observed a lifetime prevalence of extrapyramidal symptoms of 65% and a point prevalence of 34%, comparable to rates reported in large epidemiological studies (Reference AydAyd, 1961; Reference SwettSwett, 1975; Reference Owens and JohnstoneOwens & Johnstone, 1982). Compared with these studies, the mean age of our patients was low, and the rate of patients ever having been treated with atypical neuroleptics was high (87%). The percentage of patients who had been exposed to neuroleptics for less than 6 months was high (38%). The overall prevalence of about 3% for primary movement disorders in all relatives seems plausible. Higher prevalence rates of Parkinson's disease, tremor and dystonia among relatives of patients with lifetime extrapyramidal symptoms than among relatives of patients without such symptoms are comparable with results reported by others, although our study was not designed as a population-based case–control study (Reference Marder, Tang and MejiaMarder et al, 1996; Reference Louis, Levy and Mejia-SantanaLouis et al, 2003). The relatives' diagnoses could be confirmed in a small subgroup.

Predictors of extrapyramidal symptoms

In our sample, logistic regression analysis revealed that a positive family history had significant predictive value for the occurrence of extrapyramidal symptoms. The probability of observed symptoms at the time of examination, for which it was the only predictor, as well as of lifetime prevalence of symptoms and reported symptoms, was increased in patients who had a relative with a primary movement disorder. The strongest predictive value was found for the duration of treatment with typical neuroleptics (lifetime occurrence of EPS, acute dystonic reaction, parkinsonism, akathisia, and EPS on typical neuroleptics) as well as with any neuroleptic medication (reported EPS). The probability of extrapyramidal symptoms increased with longer duration of exposure. Furthermore, younger age was also a significant predictor for the occurrence of symptoms, especially for acute dystonic reaction.

Other possible predictors: age, gender and psychiatric diagnosis

We were able to validate young age as a predictor for the occurrence of extrapyramidal symptoms, especially of acute dystonia (Reference AydAyd, 1961; Reference SwettSwett, 1975). This finding gives further support to the genetic influence on occurrence of these symptoms. In contrast, there is no evidence from our data for either diagnostic specificity or gender as a predictor for the occurrence of symptoms.

Methodological limitations

Our study design must be regarded as somewhat explorative, since for most of the relatives the diagnosis of a primary movement disorder could only be established through the family history interview. This procedure might have reduced the validity and reliability of the data, although special efforts were made to rate a family history as positive only if the criteria were unambiguously fulfilled. In a clinical setting, this is also the most practical way to gather information and decide whether a patient is at risk of developing extrapyramidal symptoms because of a positive family history.

We are aware that the way in which we categorised neuroleptic dosages may appear arbitrary. By definition, atypical neuroleptic dosages cannot be converted to chlorpromazine equivalents. We therefore decided to use a categorisation based on clinical experience. The distribution of dosage ranges in our sample appears reasonable (22% on low, 52% on medium and 26% on high medication dosages). Exact dosages could only be assessed for the time of examination.

Genetic considerations

Our data imply two hypotheses. First, our findings underline the notion of genetic susceptibility for secondary extrapyramidal symptoms, and second, they suggest possible shared genetic factors in primary and secondary movement as well as in psychotic disorders.

CLINICAL IMPLICATIONS

1. ▪ Neuroleptic medication with a known low risk of extrapyramidal symptoms (i.e. an atypical neuroleptic) is preferable for patients with a family history of primary movement disorders, especially dystonia.

2. ▪ Atypical neuroleptics are also preferable in patients under 40 years old, particularly to reduce the occurrence of painful acute dystonic reaction.

3. ▪ Patients who have been exposed to long-term typical neuroleptic medication should be switched to an atypical agent even though extrapyramidal symptoms might not have been observed as yet.

LIMITATIONS

1. ▪ Our study design must be regarded as somewhat explorative owing to the family history approach.

2. ▪ Our data-set might have been too small to confirm any dosage effect.

3. ▪ The small sample size might also have been the reason why we did not observe the gender effect described elsewhere of a higher prevalence of acute dystonia in men and of parkinsonism, akathisia and tardive dyskinesia in women.