The apolipoprotein E ϵ4 allele is not associated with psychiatric symptoms or extrapyramidal signs in probable Alzheimer's disease

The apolipoprotein E (apo E) ϵ4 allele has been shown to modify risk and age at onset of Alzheimer's disease (AD).^1-3 Neuropathologic studies have found more severe AD pathology in subjects carrying the apo E ϵ4 allele than in those without the allele. This includes more cortical neurofibrillary tangles and senile(neuritic) plaques (SP),⁴ and a more severe cholinergic deficit in the frontal lobes.⁵ Moreover, neuro-imaging studies have shown greater atrophy in medial temporal structures in AD patients carrying the apo E ϵ4 allele.⁶

The development of psychiatric symptoms appears to be related to more severe neuropathologic changes^7-9 and to specific neurotransmitter imbalances, notably acetylcholine¹⁰ and serotonin.¹¹ Consequently, given that patients carrying the apo E ϵ4 allele have more severe structural and neurochemical abnormalities, psychiatric symptoms should be more prevalent in AD patients with the apo E ϵ4 allele than in those without the apo E ϵ4 allele. However, the results of studies addressing this hypothesis are discrepant. On the one hand, Ramachandran et al.¹² found, in a small group study (N = 46), that patients with the ϵ3/4 genotype had more depression and psychosis than AD patients with the ϵ3/3 genotype. By contrast, in a larger cohort (N = 120) Lyketsos et al.¹³ found no relationship between the presence of the apo E ϵ4 allele and major depression, minor depression, hallucinations, or delusions.

As a further consequence of the more severe pathology associated with apo E genotypes, extrapyramidal signs (EPS) should be more frequent in patients carrying the apo E ϵ4 allele. Postmortem studies have shown a relationship between the pathologic changes associated with Parkinson's disease (PD) (e.g., ubiquitin-positive neuritic changes in CA2/3 of the hippocampus) and the apo E ϵ4 allele in AD patients.¹⁴ Furthermore, patients with Lewy body disease and Lewy body variant of AD carrying the apo E ϵ4 allele have increased SP and cerebral amyloid angiopathy.^15,16 Although Lehtovirta et al.¹⁷ found no relationship between the apo E ϵ4 allele and rigidity, hypokinesia, tremors, and myoclonus, the sample size was small (N = 58) and four of the patients were taking antipsychotic medication.

The purpose of the present study was to examine the relationship between the presence of the apo E ϵ4 allele and the presence of psychiatric symptoms and EPS in AD patients. The sample is large, well characterized, and our ability to examine specific symptoms (as well as syndromes) permit a more powerful examination of these hypotheses.

Methods We reviewed the clinical records of 201 patients with probable AD¹⁸ who underwent apo E genotyping at the Alzheimer's Disease Research Center of the University of Pittsburgh. Each patient was assessed with an extensive medical, neurologic, psychiatric, social work/nursing, and neuropsychologic evaluation.¹⁹ None of the patients had signs of significant cerebrovascular disease, and those with a Hachinski Ischemic Score (HIS) greater than 4²⁰ or with evidence of possible infarction on neuroimaging studies were excluded from the study.

Apolipoprotein E genotyping. Apolipoprotein E genotyping was determined using a polymerase chain reaction amplification of a portion of the gene followed by restriction enzyme analysis with Hha I, as described previously.²¹

The ϵ4/4 genotype was present in 19 of the 201 patients (9.5%),ϵ3/4 in 97 (48%), ϵ3/3 in 70 (39%), ϵ2/4 in three (1.5%),ϵ2/3 in three (1.5%), and ϵ2/2 in one (0.5%). Because the ϵ2 allele appears to have a protective role in AD,²² and the number of ϵ2 allele carriers was quite small, we excluded from this study those patients who were carrying the ϵ2 allele (e.g., ϵ2/4,ϵ2/3, ϵ2/2), resulting in a total of 194 subjects.

Psychiatric assessment. The psychiatric evaluations were conducted by psychiatrists using a semistructured interview with both patients and primary caregivers.²³ This interview encompasses the DSM-III-R criteria for axis I disorders.²⁴ In addition, the psychiatrists completed the Behavior Rating Scale (BRS) for dementia of the Consortium to Establish a Registry for Alzheimer's Disease.²⁵ The BRS is a 51-item psychopathologic scale that reviews relevant psychiatric signs and symptoms of demented patients, and rates the symptomatology from 0 to 8 based on the frequency of occurrence. For the purpose of this study, signs and symptoms were recorded as present or absent. None of the patients in this study had a history of depression, psychotic illness, or substance abuse before the onset of dementia. Details of the clinical criteria for the diagnosis of delusions, hallucinations, and depression have been published previously.²⁶

Extrapyramidal signs. The neurologic examination was performed by both a neurologist and a physician's assistant (PA) trained in the medical evaluation of elderly demented individuals; details of this examination have been published previously.²⁷ The EPS were rated using individual items of the New York University (NYU) scale for parkinsonism.²⁸ The NYU scale was completed after the neurologic examination, and both the PA and the neurologist had to agree on the characteristics and severity of each sign. The EPS included in this study were: resting and postural tremors, rigidity, bradykinesia, postural instability, and abnormal gait. None of these patients exhibited dyskinesia. For the purpose of this study, the extrapyramidal syndrome was considered present when the patient exhibited at least one of the listed signs. Patients taking neuroleptic medication (N = 16) at the time of the examination were not entered in the analysis of EPS, even when it was thought that medication was not the cause of EPS.

Statistical analysis. Demographic and clinical variables were analyzed using ANOVA with post hoc comparisons (Tukey's HSD test), and chi-squared analysis. The association between the apo E ϵ4 allele (e.g.,ϵ3/4, ϵ4/4) and the presence of major depression, psychosis, and EPS were analyzed using a logistic regression analysis adjusted by age at onset, education, duration of the symptoms, and Mini-Mental State Examination(MMSE) score.²⁹ Subjects with the apo E ϵ3/3 genotype were used as the reference.

Results. Demographic characteristics. There were more men than women with the ϵ4/4 genotype than with the ϵ3/3(χ² = 3.7, df = 1, p = 0.005) and ϵ3/4(χ² = 6.4, df = 1, p = 0.01) genotypes (women:ϵ3/3, 73%; ϵ3/4, 77%; ϵ4/4, 42%). There were no differences among groups in the rate of family history of dementia (χ² = 0.44, df = 2, p = 0.80). Patients with the ϵ4/4 genotype were younger than those with the ϵ3/3 (t = 2.50, df = 96, p = 0.01) orϵ3/4 (t = 2.48, df = 112, p = 0.01) genotypes (age at onset[years]: ϵ3/3, 69.4 ± 8.6; ϵ3/4, 68.7 ± 7.8; andϵ4/4, 63.8 ± 8.1). There were no significant differences as a function of genotype in terms of years of education (F = 0.41, df = 2, p = 0.66) or duration of symptoms (F = 0.77, df = 2, p = 0.46).

Neuropsychiatric characteristics. No statistical differences were noted among groups in MMSE²⁹ (F = 0.32, df = 2, p = 0.72), Mattis Dementia Rating Scale³⁰ (F= 0.96, df = 2, p = 0.38), Blessed Dementia Rating Scale³¹ for activities of daily living (F = 1.50, df = 2, p = 0.22), Hamilton Depression Scale³² (F= 1.78, df = 2, p = 0.17), or the HIS¹⁷ (F = 1.69, df = 2, p = 0.18).

Psychiatric symptoms and EPS. The frequency of individual psychiatric symptoms, as assessed by the semistructured psychiatric interview and the BRS, did not differ among apo E genotypes in the univariate analyses. The proportion of patients with major depression (ϵ3/3, 15%;ϵ3/4, 18%; ϵ4/4, 10.5%) (χ² = 0.08, df = 2, p = 0.76), delusions (ϵ3/3, 49%; ϵ3/4, 43%; ϵ4/4, 26%) (χ² = 1.1, df = 2, p = 0.28), and hallucinations(ϵ3/3, 24%; ϵ3/4, 13.5%; ϵ4/4, 16%) (χ² = 2.6, df = 2, p = 0.10), or both (either delusions or hallucinations)(ϵ3/3, 54%; ϵ3/4, 44%; ϵ4/4, 26%) (χ² = 2.2, df = 2, p = 0.10) was not different among genotypes.

Extrapyramidal signs were examined in 178 patients free of neuroleptic medication (number of patients by genotype: ϵ3/4, 75; ϵ3/4, 87; ϵ4/4, 16). The frequency of overall EPS (i.e., present/absent) showed a borderline difference across the three subject groups (χ²= 5.4, df = 2, p = 0.06). This appears to be related to the decreased rate of occurrence among patients with the ϵ3/4 genotype (EPS:ϵ3/3, 44%; ϵ3/4, 26%; ϵ4/4, 50%). However, the frequency of individual EPS (e.g., bradykinesia, rigidity, tremors, abnormal gait) was not statistically different among groups. Furthermore, the severity of the EPS, as measured by the NYU scale,²⁷ was not statistically significant among groups (F = 0.87, df = 2, p = 0.68).

Logistic regression analysis. Because of the intercorrelation between the factors that might predict the presence of psychiatric symptoms and EPS (e.g., age, duration of the symptoms, education, severity of the dementia), a logistic regression analysis was completed to examine the effect of the presence of the apo E ϵ4 allele controlling for these factors. Major depression (yes/no), psychotic symptoms (i.e., delusions or hallucinations) (yes/no), and overall EPS (yes/no) were regressed on age at onset, duration of the symptoms, education, severity of dementia (e.g., MMSE score), and the presence of the apo E ϵ4 allele (i.e., ϵ3/4,ϵ4/4). In a multivariate model, the presence of the apo E ϵ4 allele was not associated with any of the three outcome variables(table).

Discussion. These findings indicate that neither psychiatric symptoms nor the presence of EPS were associated with any specific apo E genotype in probable AD patients. As previously reported, patients homozygous for the apo E ϵ4 allele had an earlier age of onset than those without the apo E ϵ4 allele.^2,3,17

Ramachandram et al.¹² found an association between depression, psychosis, and the ϵ3/4 genotype in a small group study that did not include subjects with the ϵ4/4 genotype. By contrast, the present study and one by Lyketsos et al.¹³ were conducted with larger samples and included patients homozygous for the apo Eϵ4 allele. Furthermore, the present analysis controlled for age at onset, education, duration of the symptoms, and severity of the dementia before examining the effect of the apo E ϵ4 allele on syndrome prevalence. Thus we feel we can confidently conclude that the apo E ϵ4 allele is not associated with psychiatric syndromes in probable AD.

We did not find an association between EPS and the apo E ϵ4 allele, which is consistent with the study conducted by Lehtovirta et al.¹⁷ In their small sample, the apo E genotypes did not predict individual EPS (e.g., tremors, bradykinesia, rigidity, dyskinesia), even though some of the patients were taking neuroleptic medication. Thus we feel that the data are also consistent on this point. The apo E ϵ4 allele is not associated with the development of EPS in probable AD.

The presence of the apo E ϵ4 allele influences the preclinical period of AD, accelerating the onset of the disease.^1-4 However, the mechanism that causes onset of dementia does not appear to affect the natural history of the disease. Furthermore, it is difficult to relate to this single biological marker the etiology of all clinical subgroups in AD. It is possible that a combination of the apo Eϵ4 allele with other still undetermined environmental, pathologic, or chromosomal abnormalities will be necessary to define clinical subgroups with worse outcomes or specific symptomatology in the natural history of AD.

Acknowledgment

JTB is recipient of the Research Scientist Development Award (level II)(KO2-MH 01077).