Increased Growth Hormone Response to Apomorphine in Parkinson Disease Compared With Multiple System Atrophy FREE

THE CLINICAL differentiation of patients with parkinsonian syndromes (PSs) involves a high risk of false-positive diagnosis of Parkinson disease (PD), particularly in early stages of the disease. Twenty-five percent of patients with clinically diagnosed PD had clear evidence of a multiple system atrophy (MSA) at postmortem examination.¹ With respect to appropriate treatment regimens, the distinction of PD from MSA is of increasing importance. The outcome of treatment studies using possibly "neuroprotective" drugs is weakened by the inclusion of patients with nonidiopathic PS, especially when patients with de novo cases are investigated.

In search of a reliable and practicable method for differentiating PD from PS of other causes, a recent study by Kimber et al² focused on the dysregulation of the central noradrenergic pathways in patients with MSA. The blunted response of growth hormone (GH) to clonidine differentiated patients with MSA from patients with PD and healthy subjects.² However, an attempt to replicate these findings was unsuccessful.³

The GH response to dopaminergic substances is a well-established neuroendocrine test to investigate the sensitivity of the overall central dopaminergic system.⁴ The present study aimed to distinguish PD from nonidiopathic PS, particularly MSA, with an endocrinologic challenge test reflecting the central dopamine receptor function. We therefore investigated the GH response to the strong dopamine 1–dopamine 2 receptor agonist apomorphine hydrochloride in patients with new clinical diagnoses of PD or MSA and compared the results with those for a healthy control group matched for age.

The study design, experimental procedure, and recruitment of patients are diagrammed in Figure 1.

We recruited 38 patients with PS who had never been treated with dopaminergic agents. Before participating in the experiments, all subjects gave informed consent. The clinical classification was made by experienced neurologists who were unaware of the results of the endocrinologic tests. The results of only the 33 patients diagnosed as having PD or MSA were included in the subsequent data analysis and compared with an age-matched healthy control group. The patients with vascular PS or with progressive supranuclear palsy were excluded (see Figure 1).

The tests with somatorelin (GH-releasing factor), a low dose of apomorphine (low-apomorphine test), or a high dose of apomorphine (high-apomorphine test) were performed on 3 consecutive days. The somatorelin test and the low-apomorphine test were performed on the first 2 days (randomized schedule). On the third day, the patients participated only in the high-apomorphine test. Eleven patients with PD and 14 patients with MSA participated in the somatorelin test. According to the protocol described by Corn et al,⁵ the subjects received either apomorphine hydrochloride, 0.005 mg/kg of body weight subcutaneously (Teclapharm GmbH, Lüneburg, Germany), or somatorelin, 1 µg/kg of body weight subcutaneously (Ferring Arzneimittel GmbH, Kiel, Germany). The high-apomorphine test^{6 - 7} (3 mg subcutaneously) was performed with a pretreatment of domperidone, 20 mg 3 times daily orally (Motilium; Byk Gulden Lomberg Chemische Fabrik GmbH, Konstanz, Germany), that started after the completion of the preceding endocrine challenge. The response to the high-apomorphine injection was evaluated by a clinical rating according to the Unified Parkinson's Disease Rating Scale, Part III (motor examination), before and 20 minutes after the injection: a score of 30% or more was rated as positive; 10% to 30%, equivocal; and 10% or less or an increase, negative.⁶ All patients were treated subsequently with levodopa at a dosage of at least 400 mg 3 times daily or an equivalent dose of a dopamine agonist (cabergoline or pergolide mesylate, at least 3 mg 3 times daily). A positive levodopa response included a subjective and objective change in the clinical symptoms within a period of at least 3 months (improvement of the Unified Parkinson's Disease Rating Scale, Part III, score ≥30%).

The blood specimens were analyzed with commercially available radioimmunoassay kits (Figure 1) (GH: Nichols Institute, San Juan Capistrano, Calif [sensitivity, 8.8 pmol/L]; prolactin: ICN Pharmaceuticals, Costa Mesa, Calif [sensitivity, <87 pmol/L]; cortisol: ICN Pharmaceuticals [sensitivity, 4160 pmol/L]; corticotropin: Nichols Institute [sensitivity, 0.22 pmol/L]). The interassay and intra-assay variation coefficients of the radioimmunoassays were all less than 8%. All samples from a given subject were measured in duplicate in 1 assay.

The experimental protocol was approved by the Ethics Committee for Human Experiments of the Bayerische Landesärztekammer (Munich, Germany).

Parkinson disease was diagnosed according to the United Kingdom Parkinson's Disease Society Brain Bank criteria.⁸ We diagnosed PD if the patient showed a positive or equivocal response in the high-apomorphine test and a clear motor improvement after 3 months of the dopaminergic treatment. We diagnosed MSA according to the consensus statement of the American Autonomic Society and American Academy of Neurology.⁹ The clinical features included (1) a PS with autonomic failure proved by either tilt-table test¹⁰ or clinically relevant urinary dysfunction; (2) negative dopaminergic treatment response in the high-apomorphine test and a negative or equivocal response to a 3-month dopaminergic treatment period; (3) presence of additional pyramidal tract signs, cerebellar signs, predominant gait disorder, falling, and dizziness (Table 1). All patients underwent magnetic resonance imaging to detect vascular lesions and atrophy.

The statistical differences in hormone responses were tested for significance by a 2-factorial multivariate analysis of covariance (MANCOVA, Wilks multivariate test of significance) for repeated measures with time as a within-subject factor and group (PD, MSA, and control group) as a between-subject factor. Age, sex, and hormone concentrations at time 0 (baseline) were entered as covariates in all subjects, and duration (in years) and the severity of the disease (in Hoehn and Yahr stages¹¹ ) only when the patient groups were compared. The MANCOVAs were applied for the hormone levels at single time points and additionally with the use of 2 robust profile characteristics: the area under the time course curve of the hormone concentration (trapezoidal integration) and the Δ values (differences between baseline and maximum values). The latter MANCOVA was done with a 1-factorial MANCOVA for each experimental condition, with group as the only between-subject factor. In case of a significant main or interaction effect of group and/or time, post hoc tests (univariate F tests and tests with contrasts) were performed to identify the pairs of groups or time points with significant differences. To approach normality and homogeneity, all variables entered in the MANCOVAs were log n–transformed. As a nominal level of significance, α = .05 was accepted and corrected (post hoc tests) according to the Bonferroni procedure to keep the type I error at .05 or less.

The analysis of covariance showed both a significant time main effect and group × time interaction effect (Wilks multivariate tests of significance; time: F_20,442 = 8.28, P<.001; group × time: F_20,483 = 2.96, P<.001), which were mainly attributed to the time and group differences of GH and prolactin responses. The GH response differed significantly between the patients with PD and MSA and between the patients with PD and the healthy control subjects at 30, 45, and 60 minutes after the subthreshold dose of apomorphine (Figure 2).

The prolactin secretion was lower in the patients with PD than in those with MSA or the control subjects, although the group differences were not statistically significant (Figure 3).

The covariance analysis showed a significant time effect (Wilks multivariate test of significance; effect of time: F_20,256 = 7.08, P<.001). The high dose of apomorphine resulted in a marked increase in cortisol, corticotropin, and GH concentrations (univariate F test; P<.05). However, there were no significant differences between the groups (Table 2).

After administration of somatorelin (Figure 4), the concentration of almost all hormones showed a time-dependent course (Wilks multivariate test of significance; effect of time: F_20,372 = 2.43, P<.001), which was similar for all groups. The only non–time-dependent hormone was prolactin.

The analysis of the concentration-curve characteristics between the group and the experimental conditions showed a significant group effect in the low-apomorphine condition (Wilks multivariate test of significance; effect of group for low apomorphine: F_16,34 = 3.05, P = .003), where only the area under the curve and Δ values of the GH responses contributed significantly to this group effect (univariate F test; P<.05).

There was no significant influence of the covariates age and sex on the hormone responses in any of the experimental conditions. The covariates duration and severity of disease had no significant effect on the hormone secretion courses. However, the patients with MSA exhibited a higher mean (±SD) Hoehn and Yahr score than those with PD (MSA, 2.7 [1.0]; PD, 2.0 [0.8]), while the duration of the disease was comparable in the 2 groups (MSA, 2.7 [2.0]; PD, 3.3 [2.1]). A significant influence of the baseline values was found only for prolactin in both the low- and high-apomorphine tests.

We investigated the extrastriatal dopaminergic sensitivity of patients with PSs by measuring the GH response to a subthreshold dose of the dopamine 1–dopamine 2 receptor agonist apomorphine. The enhanced GH release found in the low-apomorphine test significantly differentiated the patients with clinically diagnosed PD from those with clinically diagnosed MSA and from an age-matched healthy control group. The prolactin response tended to be lower in the patients with PD than in the other 2 groups. The cortisol and corticotropin concentration did not distinguish among the 3 groups.

We hypothesized that previously untreated patients with PD would show an increased sensitivity of the hypothalamic receptors because of a nigrostriatal dopaminergic deficit. To reveal the putative hypersensitivity of the GH-controlling hypothalamic dopaminergic pathways, we used a subthreshold dose of apomorphine that did not induce adverse effects. The difference in the apomorphine-induced GH release between the PD and MSA groups disappeared when a high dose of apomorphine was given to test the dopaminergic treatment response.

We speculate that both the dopamine-induced increase in GH secretion and decrease in prolactin secretion result from a hypothalamic dopaminergic hypersensitivity in PD. This was investigated in a recent positron emission tomographic study that used fluorodopa F 18 and demonstrated significant changes in the dopaminergic metabolism in patients with PD within extrastriatal structures.¹²

The neurophysiologic mechanisms mediating the dopaminergic control of the GH release are still unclear. There is some evidence that dopamine 2 receptors influence the M1 cholinergic control of somatostatin, and thus GH release, since the dopamine-induced GH release is antagonized by pirenzepine hydrochloride.¹³ In addition, dopamine is known to stimulate somatorelin secretion,^{2 ,14} although dopaminergic receptors have not yet been demonstrated on the somatorelin-secreting neurons in humans. Dopamine 1 and 2 receptors have been characterized on pituitary somatotrophs in animal experiments¹⁵ and in neuroimaging studies in patients with pituitary adenomas,¹⁶ which also suggests a direct mechanism at the pituitary cell level.

A recent study using the α₂-adrenoceptor agonist clonidine as a stimulus for GH release focused on the central autonomic deficit as a condition pathognomonic for MSA. The clonidine-induced GH response separated these patients with MSA from those with PD.² However, these results have not yet been replicated by other authors.³ Llau et al¹⁷ compared the response of GH and prolactin with apomorphine in patients with PD before and after a dopaminergic treatment. They failed to detect any significant differences between the groups and concluded that hypothalamic dopaminergic sensitivity is normal in PD. We found comparable absolute values for the GH response in the low-apomorphine test in our patients with PD. Therefore, other factors, such as the patients' clinical classification, may explain the observed differences between Llau and coworkers' study and our results.

To exclude major abnormalities in the regulation of the somatotropic system, we investigated the somatorelin-induced GH response, which was comparable in all 3 groups. The subemetic dose of apomorphine did not affect the hormones of the hypothalamic-pituitary-adrenal system. In contrast, the high dose of apomorphine resulted in an immediate and marked rise in the circulating corticotropin and cortisol levels in all patients, which is probably related to the physical stress induced by the side effects of apomorphine. The pretreatment with domperidone increased the baseline prolactin levels in both patient groups, most likely because of the blockade of pituitary dopamine 2 receptors.

We could not find a significant influence of sex; however, this may have been masked by the unequal sex distribution in our study sample. Other influences such as age and duration and severity of the disease did not account for the observed differences in the GH response to the subthreshold dose of apomorphine.

In summary, the GH response to a subthreshold dose of apomorphine could be a simple and clinically useful additional tool for the differential diagnosis of patients with PS. This challenge test is easy to standardize, inexpensive, and without any adverse effects for the participants. Future studies using larger samples may evaluate the influence of age, sex, and pretreatment with dopaminergic agents.

Accepted for publication November 2, 2000.

Presented in part at the meeting of the European Neurological Society, Milan, Italy, June 9, 1999.

We thank Elisabeth Kappelmann for her excellent technical assistance, Alexander Yassouridis for his helpful comments on the statistical analysis, and Susanne Heim for secretarial services.

Corresponding author and reprints: Elisabeth Friess, MD, Max Planck Institute of Psychiatry, Kraepelinstr 10, D-80804 Munich, Germany (e-mail: friess@mpipsykl.mpg.de).