- Email: [email protected]
Neurohormonal responses to d -fenfluramine in healthy elderly subjects. A placebo-controlled study
Psychoneuroendocrinology 25 (2000) 139 – 150 www.elsevier.com/locate/psyneuen
Neurohormonal responses to D-fenfluramine in healthy elderly subjects. A placebo-controlled study Rajamannar Ramasubbu a,*, Alastair Flint b, Gregory Brown c, George Awad c, Sidney Kennedy c a
Department of Psychiatry, Uni6ersity of Ottawa, Royal Ottawa Hospital, 1145 Carling A6enue, Ottawa, Ont. K1Z 7K4, Canada b Department of Psychiatry, Uni6ersity of Toronto, The Queen Elizabeth Hospital and The Toronto Hospital, Toronto, Ont., Canada c The Clarke Institute of Psychiatry, Uni6ersity of Toronto, Toronto, Ont., Canada Received 28 September 1998; accepted 3 June 1999
Abstract Considering age-related changes in serotonin (5HT) function, we examined normative data of prolactin (PRL) and cortisol (CORT) responses to D-fenfluramine (D-FEN) in healthy elderly subjects. Twenty-three healthy male and female volunteers aged 60 – 86 participated in a single-blind, placebo-controlled, fixed-order, crossover-design challenge test. Two baseline PRL and CORT values and the responses of these hormones to 30 mg of oral D-FEN and placebo over a 4 h period were measured on two separate sessions. PRL and CORT responses were significantly greater following D-FEN than after placebo. Peak PRL responses (maximum change from baseline following D-FEN) were relatively robust compared to peak CORT responses. Peak PRL concentration was positively correlated with plasma D-nor-FEN concentration. Gender and aging had no effect on hormonal responses in the elderly. Although the weight adjusted dose used in this study was higher than the therapeutic dose of D-FEN, PRL responses were modest and only two participants experienced side effects. D-FEN is a safe serotonergic probe and PRL responsivity to D-FEN is a reliable index of central 5HT function in the elderly. An age-related decline in serotonergic function
This paper was presented at the 52nd Annual Scientific Convention of the Society of Biological Psychiatry, May 14–18, San Diego, CA. * Corresponding author. Tel.: + 1-613-722-6521, ext. 7076; fax: +1-613-722-5048. E-mail address: [email protected] (R. Ramasubbu)
0306-4530/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 6 - 4 5 3 0 ( 9 9 ) 0 0 0 4 4 - X
140
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
must be considered in determining the dose requirement for maximal hormonal responses to D-FEN challenge tests in the elderly. © 2000 Elsevier Science Ltd. All rights reserved. Keywords:
D-fenfluramine;
Prolactin; Cortisol; Serotonin; Elderly
1. Introduction Abundant evidence indicates that serotonergic neurotransmission may play a stimulatory role in secretion of prolactin (PRL), adrenocorticotropic hormone (ACTH) and cortisol (CORT) (Van de Kar, 1991). On this basis, PRL and CORT responses to fenfluramine (FEN), a serotonergic agonist have been employed as indirect indices of brain 5HT function in a neuroendocrine challenge paradigm. It has been suggested that this approach is superior to cerebrospiral fluid (CSF) and platelet studies as it provides an accurate measure of the overall functional status of the central 5HT system (Yatham and Steiner, 1993). FEN has been widely used in humans and animals to study 5HT function of the brain. It releases 5HT, inhibits 5HT reuptake and stimulates post-synaptic receptors directly and indirectly (Costa et al., 1971; Fuxe et al., 1975). Thus, this probe provides a measure of net presynaptic and post-synaptic serotonergic functioning. A racemic form of D,L-FEN induces a dose dependent increase in PRL that is attentuated by a non-selective 5HT antagonist metergoline (Quattrone et al., 1983), suggesting that this response is mediated by the serotonergic system although it is possible that dopamine antagonist properties of the L-isomer might influence this effect (Invernizzi et al., 1989) and complicate the interpretation. Despite recent studies showing that the effect of 60 mg of oral DL-FEN on PRL release is not mediated by dopaminergic system (Coccaro et al., 1994), the interindividual variations in handling racemic compound into D-isomer and L-isomer might affect its selectivity for 5HT. Further, CORT responsivity to D,L-FEN has been attributed to a nonserotonergic mechanism, direct action by 5HT on adrenal gland (Van de Kar et al., 1985), and stress response to unpleasent side effects associated with D,L-FEN (O’Keane and Dinan, 1991). D-FEN has been reported to be highly specific in its serotonergic properties as it lacks the additional L-isomer. Hence, PRL and CORT responses induced by D-FEN might be due to specific activation of 5HT systems. Further, 30 mg of oral D-FEN is well tolerated without any reported unpleasant side effects justifying its use in an aging population who may be especially sensitive to the effects of drugs. Age is considered to be an important source of variance for 5HT mediated hormonal changes as normal aging seems to be associated with a significant reduction in 5HT receptor function (5HT1A, 5HT2A) (Marcusson et al., 1984a,b). Age-related changes in drug disposition may alter hormonal responses to D-FEN, yet little information is available on neuroendocrine responsivity to FEN challenge in the elderly. Previous reports concerning age-related changes in PRL responsivity to FEN in normal healthy volunteers produced inconsistent results probably due to differences in age distribution of subjects. McBride et al. (1990) reported an
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
141
age-related decline in D,L-FEN induced PRL responses in normal subjects with a wide age distribution (21 – 74). However, Muldoon et al. (1996) found no relationship between PRL responses and age in relatively young subjects with the age range of 25–60. Further, the later study omitted a placebo-control condition and hence failed to match the circadian variation and stress effects related to the procedure on hormonal responses (Thompson et al., 1994). Furthermore, both studies used the non-specific D,L-FEN as a challenge drug. To our knowledge this is the first study to report on PRL, CORT and behavioural responses to an oral D-FEN challenge compared to placebo in healthy elderly subjects. We examined characteristics of PRL and CORT responses to D-FEN in order to determine the effects of gender, age, body weight and drug concentration on these hormonal responses in an elderly population.
2. Materials and methods
2.1. Subject selection Twenty-three healthy volunteers (15 women, 8 men) aged 60–86 participated in the study. The mean age of the group was 70.4 years. The subjects were recruited through local distribution of posters, advertisements in newspapers and by word of mouth. Monetary incentives were provided for their participation in the study. Subjects were excluded if they had the following medical disorders: diabetes mellitus requiring insulin or hypoglycemic agents, moderate and severe hypertension (diastolic blood pressure \105 mmHg) (Williams, 1994), coronary artery disease, hypothyroidism, serious liver disease, cancer, stroke, epilepsy or any other severe medical or neurological disorders. Physical examination supplemented by appropriate laboratory tests were performed to rule out any of the above mentioned conditions in participating subjects. The medical and surgical history of the participants include arthritis, osteoporosis, mild glaucoma, cataract, partial thyroidectomy with normal thyroid function, prostatectomy, ectopic beats, varicose veins, elevated triglycerides, mild hypertension, asymptomatic cardiac murmur, cholycystectomy, chondrocalcinosis, sinusitis, inguinal hernia, appendicectomy, removal of ovarian cystic fibroid tumor. Further, they were not receiving medications that were known to produce changes in the 5HT system or PRL concentrations. These included ketanserin, methyldopa, oral contraceptives, estrogen, glucocorticoids, thyroid supplementation, metaclopropamide, cimetidine, antipsychotic and antidepressant medications. The medications permitted in the study but witheld on the day of the challenge tests were as follows: asprin, ibuprofen, vitamin E, vitamin C, calcium, acetoaminophen, diclofenac sodium, except for levobunolol and pilocarpine eye drops. The alcohol consumption never exceeded 20 units/week (a single unit contains 8 g of ethanol) in men and 14 units/week in women (Ritson and Chick, 1988). At the time of the study only two subjects were smokers (one smoked a pack a day regularly and the other was an ocassional smoker). Subjects
142
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
were free of axis I psychiatric disorders as determined by a clinical evaluation performed by a trained psychiatrist (RR) using a semistructured psychiatric interview. Signed informed consent was obtained from all subjects as approved by a local ethics committee.
2.2.
D -fenfluramine
challenge test
The challenge test was performed in the Clinical Investigation Unit at The Toronto Hospital utilizing a placebo-controlled design. Tests were carried out on 2 consecutive days, except for two subjects who had 7 days between tests. For all subjects oral D-FEN, which has longer carry over effects, was administered on the second session. Only the principal investigators were aware of the order of presentation to subjects of placebo and drug conditions. Participating subjects, the research nurse who administered clinical scales, medications, and drew blood samples during the challenge test, and the laboratory technicians who performed hormonal assays were blind to the order of drug and placebo conditions. Subjects reported to the test centre at 0800 h after an overnight fast. Height and weight were measured and expressed in meters and kilograms to facilitate body mass index calculations. On each occasion, subjects were served a low tryptophan breakfast which consisted of apple sauce, jello, and orange juice. The provision of food was aimed at minimizing hypoglycemic effects on PRL and CORT during the 5 h of the study. Subjects were allowed to drink only water and were required to remain awake during the test. Nicotine, and caffeine consumption were prohibited for the duration of the test. An intravenous cannula was inserted in the anterior cubital vein of subjects resting in a reclined position and a physiological saline drip was commenced. The first blood samples were drawn through a three-way stop cock in the intravenous line following 30 min of adaptation and 30 min later (at 0 time) just before the oral administration of 30 mg of D-FEN or identical placebo capsules. These two samples were obtained for baseline PRL and CORT assay. Thereafter, blood samples were drawn and immediately centrifuged every 60 min for 4 h (i.e. four samples post-challenge) and plasma was stored at − 25°C until analysis. A fixed dose of 30 mg of D-FEN which is equivalent to the dose of D-isomer in 60 mg of oral D,L-FEN (Silverstone et al., 1987) was employed, on the assumption that higher dose may produce a higher frequency of side effects in the elderly and too low a dose may be ineffective in inducing hormonal changes due to an age-related decline in 5HT function. The sampling period for PRL and CORT assays was limited to 4 h which is shorter than the standard 5 h test in order to minimise the discomfort of prolonged testing. In addition, previous placebo-controlled studies indicated that the maximal PRL response occurs at 4 h after oral D-FEN administration (Quattrone et al., 1983). Plasma was obtained at 3 h after administration of D-FEN for assays of FEN and nor-FEN levels. This time point was selected because the pharmacokinetics of oral D-FEN show a maximal plasma level of its active metabolites D-FEN and nor-D-FEN 2 – 4 h post D-FEN (Campbell, 1991).
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
143
2.3. Beha6ioural and cardio6ascular response Subjects were asked to rate themselves using a 10 cm Visual Analogue Scale, containing four parameters (sleep, anxiety, sad, high). This self rating scale was completed at time 0 and every hour for 4 h on the two challenge sessions. Subjects were asked to report any unpleasant side effects including light headedness, cold, nausea, diarrhea, dry mouth, visual impairment, anorexia, headache or bodyache over the duration of the test and also after the test. Heart rate and blood pressure were also recorded at each time point.
2.4. Assays Plasma PRL and CORT concentrations were determined by quantitative enzyme immunoassay methods using transferable solid phase technology. The kits were supplied by Sychron Enzyme linked immunosorbent assay (SYNELISA, USA). The intra-assay and inter-assay coefficients of variation for plasma PRL were 5.0 and 7.8%, and for plasma CORT, were 5.2 and 7.1%, respectively. The lower detection limit for plasma PRL was 2 ng/ml and for plasma CORT was 0.5 ng/ml. Levels of FEN and nor-FEN were assayed by gas liquid chromatography using nitrogen sensitive detectors (Kerbs et al., 1984). The intra-assays and inter-assays coefficients of variance were reported to be 5.2 and 7.9%, respectively. The lower detection limit was 2 ng/ml.
2.5. Data analysis Statistical analyses were performed using the computer program SPSS. Lilliefors test (modified Kolmogorov – Nor test) was used to ascertain the normality of distribution. PRL and CORT data were calculated for three outcome measures: (a) Net changes in PRL and CORT were calculated by subtracting the baseline (mean of −30 and 0 time values) from hormonal changes at each time point after administration of placebo and D-FEN; (b) In order to take into account the circadian effects, placebo-controlled net changes were obtained by subtracting net changes during placebo condition from net changes at each time point post D-FEN; (c) To examine the magnitude of hormonal responses, peak hormonal responses (differences between the mean baseline level and the highest hormonal concentration after D-FEN administration) were computed. Negative values (values below baseline) were also taken into account for the analysis. Comparisons of net changes in placebo and drug conditions within subjects were performed using two-way analysis of variance with repeated measures (ANOVA) to evaluate the significance of the main effects of drug (D-FEN versus placebo), time (changes over each time point) and also drug× time interactions. Paired Student’s t-tests were used to analyse simple effects when main effects were significant. Between group comparison for sex (female versus male), and age (65–69 versus \ 70) of net changes in placebo and drug conditions and placebo-controlled net changes were performed using ANOVA and significant interactions revealed by
144
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
ANOVA were analyzed by unpaired Student’s t-tests. Intergroup comparisons of continous data and categorical data were performed with t-tests (two tailed) and chi-squared tests, respectively. Pearson’s product moment correlational analysis was used to evaluate relationships among normally distributed variables (neuroendocrine measures, metabolites D-FEN and nor-D-FEN, age and weight). The area under the response curve (AUC) was calculated using the trapezoid method for both PRL and CORT responses to D-FEN and placebo from time 0 until time 240 min (AUC 60 – AUC 240). AUC values were compared using Student’s t-tests. The statistical significance of all tests was set at PB .05 (two tailed). All values were reported as the mean9SD unless stated otherwise.
3. Results
3.1. Hormonal responses 3.1.1. Prolactin Figure 1 summarises the PRL response to D-FEN and placebo over 240 min. The increase in PRL levels began to appear 2 h post D-FEN and maximal PRL response was noted at the 4 h time point. Mean baseline PRL concentrations did not differ in drug and placebo conditions (t = − 1.20, df= 22, P= .24). The ANOVA on net PRL changes over 240 min revealed main effects of drug (F=13.83, df= 1,22, P= .001), time (F =14.43, df=3,66, P= .00), and also a significant drug by time interaction (F =18.60, df=3,66, P =.000) indicating an increase in PRL concentration in response to D-FEN. Post hoc tests showed that net PRL changes post D-FEN were significantly higher at 180 (t= 3.71, df =22, P= .000) and 240 min
Fig. 1. Mean plasma prolactin concentration after 30 mg of oral D-fenfluramine (solid line) and placebo (dashed line) in 23 elderly healthy volunteers compared to placebo. The prolactin response to Dfenfluramine was significantly elevated (ANOVA).
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
145
Fig. 2. Mean plasma cortisol concentration after 30 mg of oral D-fenfluramine (solid line) and placebo (dashed line) in 23 elderly healthy volunteers. There was a significant difference in cortisol response between placebo and D-fenfluramine (ANOVA).
(t=4.47, df = 22, P=0.00) compared to post placebo net PRL changes at the same time points. AUC 60 – 240 of net PRL responses were significantly higher following D-FEN than placebo (t =3.19, df = 22, P= .004). Among 23 subjects, 16 (69.6%) had positive peak responses and the relative increase of peak PRL response of the entire sample was 26% above the baseline estimate. The peak PRL concentration was positively correlated with plasma nor-D-FEN levels (r= .467, P=.03), mean basal CORT concentration (r= .427, P = .04). However there were no significant correlations of peak PRL response with combined plasma D-FEN and nor-D-FEN concentration (r= .341, P = .12), D-FEN levels (r=.284, P=.20), age (r= − .056 P=.80) or body mass index (r = −.323, P= .13) or mean baseline PRL concentration (r =.202, P =.35) or sex (r = .379, P =.07).
3.1.2. Cortisol CORT responses to D-FEN and placebo are shown in Fig. 2. The maximal CORT response was observed at 3 h post D-FEN and was not present at 4 h. Although baseline CORT concentrations were not statistically different between the two challenge sessions, there was a trend indicating higher baseline CORT concentration on the first day compared to the second day (t= − 1.95, df=22, P=.064) suggesting a mild stress response. When baseline CORT concentration was entered into an analysis of covariance (ANCOVA), it showed the main effect of drug (F = 6.61, df=1,22, P =.017) and drug by time interaction (F= 4.13, df= 3,66, P= .01) on net CORT changes. The effect of D-FEN on CORT changes was significant at 180 min (t =3.46, df=21, t= .002) and 240 min time points (t=2.37, df= 21, P = .027). AUC 60 – 240 of net CORT responses was greater following D-FEN than placebo (t = 2.69, df= 22, P= .013). However, mean CORT concentrations for the whole group did not rise above the baseline. The majority (53%) of the group had negative CORT responses. There was a positive correlation of peak
146
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
CORT response with plasma nor-D-FEN concentration (r=.541, P= .01), but there were no significant correlations with plasma D-FEN concentration (r= .193, P= .39), combined plasma D-FEN and nor-D-FEN concentration (r= .281, P= .21), body mass index (r = −.328, P = .13), age (r= .258, P= .23), sex (r= − .190, P= .38) or mean baseline CORT concentration (r= −.280, P= .19).
3.1.3. Gender effect ANOVA revealed that there was no main effect of gender (F= .67, df= 1,21, P=.42), or gender by time (F =1.41, df= 3,63, P= .248), or gender by drug interaction (F =1.05, df =1,21, P =.378), or gender by time by drug interaction (F = 2.27, df =3,63, P =.09) on baseline corrected net PRL changes. Similarly there was no gender effect (F = .00, df= 1,21, P= .96) or gender by time interaction on placebo controlled net PRL changes (F= .64, df= 3,63, P= .60). With respect to CORT responses, there was no gender effect (F= 1.83, df= 1,21, P=.19) or gender by time interaction (F= 1.35, df = 3,63, P= .27) on placebo controlled CORT responses. Further, there was no significant effect of gender (F = .06, df= 1,21, P =.82) on baseline net CORT changes or significant interactions between gender and time (F =1.80, df= 3,63, P= .156) or gender and drug (F= 1.83, df= 1,21, P =.190) or gender by drug by time (F= 1.35, df= 3,63, P = .27). In sum, our results suggest the absence of gender differences in net serotonergic responsivity in late life. 3.1.4. Age effect In order to determine age-related changes in D-FEN induced hormonal responses, the group was divided into young–old (65–69 years) and old–old (\70 years) on the basis of a median-split of age (69 years). Comparisons of hormonal changes between young – old and old –old were performed using ANOVA. There were no significant effects of age (F =.07, df=1.21, P= .789); age by drug (F = .41, df=1,21, P =.528); age by time (F=2.03, df= 3,63, P= .118) or age by drug by time interaction (F =1.59, df =3,63, P= .201) on net CORT changes. Similarly, young – old subjects (65 – 69 years) did not differ from old–old subjects (\ 70 years) in placebo corrected CORT changes (effect of age F= .41, df = 1,21, P= .53, age by time F =1.59, df =3,63, P =.20). There were trends indicating a weaker PRL response with advancing age (age by time interaction for placebo corrected PRL changes: F =2.50, df=3,63, P= .068; age by time by drug interaction for baseline corrected PRL changes: F=2.50, df = 3,63, P =.068). 3.2. Beha6ioural and cardio6ascular response D-FEN did not increase self-rated anxiety, sad effect, mood elevation or alertness. Two subjects complained of feeling cold following D-FEN administration. In regard to unpleasant side-effects, one male subject reported dizzy spells which lasted for 24 h after the test. One female subject was found to have an irregular heart beat after D-FEN administration. None of the subjects had nausea or vomiting. ANOVA revealed a significant drug by time interaction for heart rate and
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
147
systolic blood pressure in drug and placebo conditions. There were, however no significant main effects of drug or drug by time interaction for diastolic blood pressure. Paired t-tests revealed that heart rate at time points 120 (t= −3.95, df=15, P = .001), 180 (t = − 4.35, df= 15, P= .001), and 240 (t= − 3.3, df= 15, P= .005), and systolic blood pressure at time point 60 (t= − 3.62, df= 18, P=.002), and 120 (t = −3.60, df= 17, t= .002) were significantly lower in post D-FEN than in post placebo conditions. Thus our results suggest that cardiovascular responses were significantly higher during the first challenge test (placebo condition) than the second challenge test (D-FEN condition).
4. Discussion This placebo-controlled investigation indicates that PRL responses were more sensitive than CORT responses to acute administration of 30 mg of oral D-FEN in healthy elderly subjects. Thus, the present data clearly shows that a fixed dose of D-FEN induces a moderate increase in peak PRL concentration but weak CORT responses neutralizing only the circadian fall seen with placebo. This observation is consistent with the findings of Gorard et al. (1993) that 30 mg of oral D-FEN is more potent in stimulating PRL release than CORT in a normal sample with a wider age distribution (range 18 – 54 years). Considering the findings of the present study involving an elderly population and previous observations in non elderly subjects it might be suggested that PRL response to D-FEN is a more sensitive measure of central 5HT function than CORT response across the life span. However, Maes et al. (1991) did not find significant differences in PRL and CORT responses to 45 mg of oral D-FEN in ten healthy young normals. Similarly, Goodwin et al. (1994) failed to demonstrate either PRL or CORT responses to 30 mg of D-FEN in ten healthy young male volunteers. In contrast, some studies using a similar dose of D-FEN demonstrated substantial increase in both PRL and CORT concentrations in healthy normals (O’Keane and Dinan, 1991). The discrepancies in these results might be due to errors of inference resulting from the lack of a placebo control (O’Keane and Dinan, 1991), smaller sample size (Maes et al., 1991; Goodwin et al., 1994) and gender differences in serotonergic responsivity (Goodwin et al., 1994). Further, the interpretation of negative responses to D-FEN reported in previous studies is also made more uncertain by the failure to examine pharmacokinetic aspects in relevance to D-FEN induced hormonal responses (O’Keane and Dinan, 1991). Our findings indicate that there are no gender differences in PRL responses to D-FEN in elderly. The previous documentation of greater PRL responses in younger women have been attributed to a stimulatory role of estrogen on PRL responsiveness to D-FEN during mid phase of the menstrual cycle (O’Keane et al., 1991). However, the present study involved only post menopausal elderly women. Our study design would not permit us to examine the effects of old age versus young age on hormonal responses to D-FEN. However, in the light of our findings that there were no differences in hormonal responses between young–old and
148
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
old–old and there was an absence of correlation between age and PRL or CORT response, it can be suggested that advanced aging may have no major impact on D-FEN induced hormonal response in the elderly. This is partly in agreement with previous observations that the negative correlation between aging and PRL responses to D,L-FEN may not be linear and that the effect of aging on 5HT system would be completed by the third decade of life (McBride et al., 1990). With regard to side effect profile, our results corroborate with previous observations reporting lower incidence of unpleasant side effects with 30 mg of oral D-FEN as a challenge agent in younger subjects (O’Keane and Dinan, 1991). However, our findings contradicts the notion that there might be an age dependant decline in drug disposition. Predominantly, serotonergic mechanisms have been attributed to side effects of fenfluramine. Hence, it is possible that as age associated decline in 5HT receptors may account for lower incidence of side effects and absence of behavioural responses with D-FEN in elderly subjects, despite using a therapeutic dose. In sum, our findings strongly suggest that D-FEN is a safe drug to employ as a serotonergic probe in the elderly population. It is of interest to note that although the weight adjusted dose (0.44 mg/kg) used in this study is higher than the standard dose (0.2–0.3 mg/kg) typically prescribed as a weight reducing agent, (Guy-Grand et al., 1989) the hormonal responses were modest. This may reflect differential dose response relationships for weight reduction and hormonal response or an age-related decline in 5HT1A and 5HT2 post synaptic receptors (Marcusson et al., 1984a,b) as 5HT mediated PRL secretion seems to be dependent on 5HT1A or 5HT2A/C or interaction of both. Given the fact that the D-FEN is better tolerated in the elderly, it might be suggested that higher therapeutic doses in the range of 45–60 mg of D-FEN be may required to maximise the hormonal responses in the elderly. Surprisingly, our results indicate a positive correlation between peak PRL responses and plasma concentration of D-nor-FEN, while there was no correlation between peak PRL responses and D-FEN concentration. Although these findings remain unclear at this juncture, this may reflect the differences in activity between D-FEN and nor-D-FEN. nor-D-FEN has a relatively higher affinity for 5HT1C (Campbell, 1991) and the involvement of 5HT1C in the mediation of PRL release has been demonstrated in preclinical studies (Rittanhouse et al., 1990). Further, it is possible that plasma level of D-FEN may reach to the maximum beyond the time point selected for metabolite assay in this study or there might be a time lag between maximal D-FEN concentration and peak PRL response. The order of drug administration is an important methodological issue to be considered. In a fixed-order design, differences in stress response to the procedure between first and second challenge sessions are not balanced, which may exhibit an order effect, therefore, stress related CORT/PRL responses and cardiovascular response tend to be higher in the first than in the second session. This might have minimised the magnitude of the differences between hormonal responses to D-FEN and placebo. There is some evidence to support this suggestion as our results have shown a significant decrease in heart rate and systolic blood pressure in the second session, although the effect of D-FEN on cardiovascular responses can not be
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
149
excluded. However, since there is no statistically significant differences in baseline CORT, PRL and anxiety responses between the two challenge tests, the possibility of stress effect on hormonal responses is less evident in this study. In conclusion, we found that D-FEN is a safe putative probe to use when assessing the net serotonergic functioning in elderly subjects. Further, this study confirms that PRL responsivity to D-FEN is a reliable measure of central 5HT function in the aged. Given the age-related loss of serotonergic post-synaptic receptors (Marcusson et al., 1984a,b) and low concentration of drug in human brains relative to plasma concentrate for a given dose it might be desirable to employ a dose in the upper therapeutic range in the challenge paradigm, against the conventional wisdom of dose reduction in the elderly, in order to maximise hormonal responses to D-FEN. Failure to do so may increase the risk of masking subsensitivity of a relatively deficient 5HT system in the elderly due to understimulation, which may have an important research implication particularly when D-FEN challenge study is designed to confirm diminished serotonergic activity in psychiatric disorders of old age, especially in late onset depression.
Acknowledgements This research is supported in part by a grant from the Canadian Psychiatric Research Foundation. Authors thank staff of the Clinical Investigation Unit, Toronto Hospital, Toronto for assisting with the challenge tests and Servier Ltd for supplying D-fenfluramine and placebo capsules. We also acknowledge Thomas Cooper, Nathan Kline Institute of Psychiatric Research, New York State for measuring levels of D-fenfluramine and D-nor-fenfluramine. Dr David Streiner provided statistical consultation and Paul Miceli assisted with statistical analyses.
References Campbell, D.B., 1991. Dexfenfluramine: an overview of its mechanisms of action. Rev. Contemp. Pharmacother. 2, 93–113. Coccaro, E.F., Kavoussi, R.J., Hauger, R., 1994. PRL and HVA responses to D- and D,L-fenfluramine in humans. Biol. Psychiatry 35, 615–747. Costa, E.A., Groppetti, A., Refuelta, A., 1971. Action of fenfluramine on monoamine stores of rat tissue. Br. J. Pharmacol. 41, 57–64. Fuxe, K., Farnebo, L.O., Hamberger, B., Ogren, S.O., 1975. On the in vivo and in vitro actions of fenfluramine and its derivatives on central monoamine neurons, especially 5 hydroxytryptamine neurons and their relation to the anorectic activity of fenfluramine. Postgrad. Med. J. 551, 35 – 45. Gorard, D.A., Taylor, T.M., Medbak, S.H., Perry, L.A., Libby, G.W., Farthing, M.J., 1993. Plasma prolactin, adrenocorticotrophic hormone, and cortisol after administration of D-fenfluramine or placebo to healthy subjects. Int. Clin. Psychopharmacol. 8, 123 – 128. Goodwin, G.M., Murray, C.L., Bancroft, J., 1994. Oral D-fenfluramine and neuroendocrine challenge: problems with the 30 mg dose in men. J. Affect. Disord. 30, 117 – 122. Guy-Grand, B.J.P., Crepaldi, G., Lefebvre, P., Apfelbaum, M., Gries, A., Turner, P., 1989. International trial of long-term dexfenfluramine in obesity. Lancet ii, 1142 – 1145.
150
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
Invernizzi, R., Bertorelli, R., Consolo, S., Garattini, S., Samanin, R., 1989. The effects of the L-isomer of fenfluramine and dopamine mechanism in rat brain. Eur. J. Pharmacol. 164, 241 – 248. Kerbs, H.A., Cheng, L.K., Wright, G.S., 1984. Determination of fenfluramine and nor-fenfluramine in plasma using a nitrogen sensitive detector. J. Chromatr. Biomed. Appl. 310, 412 – 417. Maes, M., D’Hondt, P., Suy, E., Minner, B., Vandervorst, C., Raus, J., 1991. HPA-axis hormones and prolactin responses to dextro-fenfluramine in depressed patients and healthy controls. Prog. NeuroPsychopharmacol. Biol. Psychiatry 15, 781– 790. Marcusson, J., Oreland, L., Winblad, B., 1984a. Effect of age on human brain serotonin (S1) binding sites. J. Neurochem. 43, 1699–1705. Marcusson, J., Morgan, D.G., Winbaild, B., Finch, C.E., 1984b. Serotonin-2 binding sites in human frontal cortex and hippocampus. Selective loss of S-2A sites with age. Brain Res. 311, 51 – 56. McBride, P.A., Tierney, H., DeMeo, M., Chen, J.S., Mann, J.J., 1990. Effect of age and gender on CNS serotonergic responsivity in normal adults. Biol. Psychiatry 27, 1143 – 1155. Muldoon, M.F., Manuck, S.B., Jansma, C.L., Moore, A.L., Perel, J., Mann, J.J., 1996. D,L-fenfluramine challenge test: experience in nonpatient sample. Biol. Psychiatry 39, 761 – 768. O’Keane, V., Dinan, T.G., 1991. Prolactin and cortisol responses to D-fenfluramine in major depression: evidence for diminished responsivity of central serotonergic function. Am. J. Psychiatry 148, 1009 – 1015. O’Keane, V., O’Hanlon, M., Webb, M., Dinan, T., 1991. D-fenfluramine/prolactin responses throughout the menstural cycle: evidence for an oestrogen-induced alteration. Clin. Endocrinol. 34, 289 – 292. Quattrone, A., Tedeschri, G., Aguglia, U., Scopacasa, F., Direnzo, G.F., AnnunZiato, L., 1983. Prolactin secretion in man: a useful to evaluate activity of drugs on central 5-hydroxy tryptaminergic neurones. Studies with fenfluramine. Br. J. Clin. Pharmacol. 16, 471 – 475. Rittanhouse, P.T., Bakkum, E.A., Herbert, G., Betha, C.L., Van de Kar, L.D., 1990. Serotonin receptor subtypes mediating neuroendocrine responses to DOI. Pharmacologist 32, 185. Ritson, E.B., Chick, J.D., 1988. In: Kendell, R.E., Zealley, A.K. (Eds.), Dependence on Alcohol and Other Drugs Comparison to Psychiatric Studies. Churchill Livingstone, Edinburgh, pp. 476 – 504. Silverstone, T., Smith, G., Richards, R., 1987. A comparative evaluation of dextrofenfluramine and dl-fenfluramine. In: Bender, A.F., Brooks, L.J. (Eds.), Bodyweight Control: The Physiology, Clinical Treatment and Prevention of Obesity. Churchill Livingstone, Edinburgh, pp. 240 – 246. Thompson, P.A., Maes, M., Meltzer, H.Y., 1994. Effect of the placebo control condition in neuroendocrine challenge studies. Psychiatry Res. 52, 317 – 326. Van de Kar, L.D., Urban, J.H., Richardson, K.D., Bethea, C.L., 1985. Pharmacological studies on the serotonergic and nonserotonin-mediated stimulation of prolactin and corticosterone secretion by fenfluramine. Neuroendocronology 41, 283– 288. Van de Kar, L.D., 1991. Neuroendocrine pharmacology of serotonergic (5-HT) neurons. Annu. Rev. Pharmacol. Toxicol. 31, 289–320. Williams, G.H., 1994. Hypertensive vascular disease. In: Isselbacher, K.J., Braunwald, E., Wilson, J.D., Martin, J.V., Fauci, A.S., Kasper, D.L. (Eds.), Harrison’s Principles of Internal Medicine, 13th edn. McGraw Hill, USA, pp. 1116–1131. Yatham, L.N., Steiner, M., 1993. Neuroendocrine probes of serotonergic function: a critical review. Life Sci. 53, 447–463.
.
Neurohormonal responses to D-fenfluramine in healthy elderly subjects. A placebo-controlled study Rajamannar Ramasubbu a,*, Alastair Flint b, Gregory Brown c, George Awad c, Sidney Kennedy c a
Department of Psychiatry, Uni6ersity of Ottawa, Royal Ottawa Hospital, 1145 Carling A6enue, Ottawa, Ont. K1Z 7K4, Canada b Department of Psychiatry, Uni6ersity of Toronto, The Queen Elizabeth Hospital and The Toronto Hospital, Toronto, Ont., Canada c The Clarke Institute of Psychiatry, Uni6ersity of Toronto, Toronto, Ont., Canada Received 28 September 1998; accepted 3 June 1999
Abstract Considering age-related changes in serotonin (5HT) function, we examined normative data of prolactin (PRL) and cortisol (CORT) responses to D-fenfluramine (D-FEN) in healthy elderly subjects. Twenty-three healthy male and female volunteers aged 60 – 86 participated in a single-blind, placebo-controlled, fixed-order, crossover-design challenge test. Two baseline PRL and CORT values and the responses of these hormones to 30 mg of oral D-FEN and placebo over a 4 h period were measured on two separate sessions. PRL and CORT responses were significantly greater following D-FEN than after placebo. Peak PRL responses (maximum change from baseline following D-FEN) were relatively robust compared to peak CORT responses. Peak PRL concentration was positively correlated with plasma D-nor-FEN concentration. Gender and aging had no effect on hormonal responses in the elderly. Although the weight adjusted dose used in this study was higher than the therapeutic dose of D-FEN, PRL responses were modest and only two participants experienced side effects. D-FEN is a safe serotonergic probe and PRL responsivity to D-FEN is a reliable index of central 5HT function in the elderly. An age-related decline in serotonergic function
This paper was presented at the 52nd Annual Scientific Convention of the Society of Biological Psychiatry, May 14–18, San Diego, CA. * Corresponding author. Tel.: + 1-613-722-6521, ext. 7076; fax: +1-613-722-5048. E-mail address: [email protected] (R. Ramasubbu)
0306-4530/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 6 - 4 5 3 0 ( 9 9 ) 0 0 0 4 4 - X
140
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
must be considered in determining the dose requirement for maximal hormonal responses to D-FEN challenge tests in the elderly. © 2000 Elsevier Science Ltd. All rights reserved. Keywords:
D-fenfluramine;
Prolactin; Cortisol; Serotonin; Elderly
1. Introduction Abundant evidence indicates that serotonergic neurotransmission may play a stimulatory role in secretion of prolactin (PRL), adrenocorticotropic hormone (ACTH) and cortisol (CORT) (Van de Kar, 1991). On this basis, PRL and CORT responses to fenfluramine (FEN), a serotonergic agonist have been employed as indirect indices of brain 5HT function in a neuroendocrine challenge paradigm. It has been suggested that this approach is superior to cerebrospiral fluid (CSF) and platelet studies as it provides an accurate measure of the overall functional status of the central 5HT system (Yatham and Steiner, 1993). FEN has been widely used in humans and animals to study 5HT function of the brain. It releases 5HT, inhibits 5HT reuptake and stimulates post-synaptic receptors directly and indirectly (Costa et al., 1971; Fuxe et al., 1975). Thus, this probe provides a measure of net presynaptic and post-synaptic serotonergic functioning. A racemic form of D,L-FEN induces a dose dependent increase in PRL that is attentuated by a non-selective 5HT antagonist metergoline (Quattrone et al., 1983), suggesting that this response is mediated by the serotonergic system although it is possible that dopamine antagonist properties of the L-isomer might influence this effect (Invernizzi et al., 1989) and complicate the interpretation. Despite recent studies showing that the effect of 60 mg of oral DL-FEN on PRL release is not mediated by dopaminergic system (Coccaro et al., 1994), the interindividual variations in handling racemic compound into D-isomer and L-isomer might affect its selectivity for 5HT. Further, CORT responsivity to D,L-FEN has been attributed to a nonserotonergic mechanism, direct action by 5HT on adrenal gland (Van de Kar et al., 1985), and stress response to unpleasent side effects associated with D,L-FEN (O’Keane and Dinan, 1991). D-FEN has been reported to be highly specific in its serotonergic properties as it lacks the additional L-isomer. Hence, PRL and CORT responses induced by D-FEN might be due to specific activation of 5HT systems. Further, 30 mg of oral D-FEN is well tolerated without any reported unpleasant side effects justifying its use in an aging population who may be especially sensitive to the effects of drugs. Age is considered to be an important source of variance for 5HT mediated hormonal changes as normal aging seems to be associated with a significant reduction in 5HT receptor function (5HT1A, 5HT2A) (Marcusson et al., 1984a,b). Age-related changes in drug disposition may alter hormonal responses to D-FEN, yet little information is available on neuroendocrine responsivity to FEN challenge in the elderly. Previous reports concerning age-related changes in PRL responsivity to FEN in normal healthy volunteers produced inconsistent results probably due to differences in age distribution of subjects. McBride et al. (1990) reported an
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
141
age-related decline in D,L-FEN induced PRL responses in normal subjects with a wide age distribution (21 – 74). However, Muldoon et al. (1996) found no relationship between PRL responses and age in relatively young subjects with the age range of 25–60. Further, the later study omitted a placebo-control condition and hence failed to match the circadian variation and stress effects related to the procedure on hormonal responses (Thompson et al., 1994). Furthermore, both studies used the non-specific D,L-FEN as a challenge drug. To our knowledge this is the first study to report on PRL, CORT and behavioural responses to an oral D-FEN challenge compared to placebo in healthy elderly subjects. We examined characteristics of PRL and CORT responses to D-FEN in order to determine the effects of gender, age, body weight and drug concentration on these hormonal responses in an elderly population.
2. Materials and methods
2.1. Subject selection Twenty-three healthy volunteers (15 women, 8 men) aged 60–86 participated in the study. The mean age of the group was 70.4 years. The subjects were recruited through local distribution of posters, advertisements in newspapers and by word of mouth. Monetary incentives were provided for their participation in the study. Subjects were excluded if they had the following medical disorders: diabetes mellitus requiring insulin or hypoglycemic agents, moderate and severe hypertension (diastolic blood pressure \105 mmHg) (Williams, 1994), coronary artery disease, hypothyroidism, serious liver disease, cancer, stroke, epilepsy or any other severe medical or neurological disorders. Physical examination supplemented by appropriate laboratory tests were performed to rule out any of the above mentioned conditions in participating subjects. The medical and surgical history of the participants include arthritis, osteoporosis, mild glaucoma, cataract, partial thyroidectomy with normal thyroid function, prostatectomy, ectopic beats, varicose veins, elevated triglycerides, mild hypertension, asymptomatic cardiac murmur, cholycystectomy, chondrocalcinosis, sinusitis, inguinal hernia, appendicectomy, removal of ovarian cystic fibroid tumor. Further, they were not receiving medications that were known to produce changes in the 5HT system or PRL concentrations. These included ketanserin, methyldopa, oral contraceptives, estrogen, glucocorticoids, thyroid supplementation, metaclopropamide, cimetidine, antipsychotic and antidepressant medications. The medications permitted in the study but witheld on the day of the challenge tests were as follows: asprin, ibuprofen, vitamin E, vitamin C, calcium, acetoaminophen, diclofenac sodium, except for levobunolol and pilocarpine eye drops. The alcohol consumption never exceeded 20 units/week (a single unit contains 8 g of ethanol) in men and 14 units/week in women (Ritson and Chick, 1988). At the time of the study only two subjects were smokers (one smoked a pack a day regularly and the other was an ocassional smoker). Subjects
142
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
were free of axis I psychiatric disorders as determined by a clinical evaluation performed by a trained psychiatrist (RR) using a semistructured psychiatric interview. Signed informed consent was obtained from all subjects as approved by a local ethics committee.
2.2.
D -fenfluramine
challenge test
The challenge test was performed in the Clinical Investigation Unit at The Toronto Hospital utilizing a placebo-controlled design. Tests were carried out on 2 consecutive days, except for two subjects who had 7 days between tests. For all subjects oral D-FEN, which has longer carry over effects, was administered on the second session. Only the principal investigators were aware of the order of presentation to subjects of placebo and drug conditions. Participating subjects, the research nurse who administered clinical scales, medications, and drew blood samples during the challenge test, and the laboratory technicians who performed hormonal assays were blind to the order of drug and placebo conditions. Subjects reported to the test centre at 0800 h after an overnight fast. Height and weight were measured and expressed in meters and kilograms to facilitate body mass index calculations. On each occasion, subjects were served a low tryptophan breakfast which consisted of apple sauce, jello, and orange juice. The provision of food was aimed at minimizing hypoglycemic effects on PRL and CORT during the 5 h of the study. Subjects were allowed to drink only water and were required to remain awake during the test. Nicotine, and caffeine consumption were prohibited for the duration of the test. An intravenous cannula was inserted in the anterior cubital vein of subjects resting in a reclined position and a physiological saline drip was commenced. The first blood samples were drawn through a three-way stop cock in the intravenous line following 30 min of adaptation and 30 min later (at 0 time) just before the oral administration of 30 mg of D-FEN or identical placebo capsules. These two samples were obtained for baseline PRL and CORT assay. Thereafter, blood samples were drawn and immediately centrifuged every 60 min for 4 h (i.e. four samples post-challenge) and plasma was stored at − 25°C until analysis. A fixed dose of 30 mg of D-FEN which is equivalent to the dose of D-isomer in 60 mg of oral D,L-FEN (Silverstone et al., 1987) was employed, on the assumption that higher dose may produce a higher frequency of side effects in the elderly and too low a dose may be ineffective in inducing hormonal changes due to an age-related decline in 5HT function. The sampling period for PRL and CORT assays was limited to 4 h which is shorter than the standard 5 h test in order to minimise the discomfort of prolonged testing. In addition, previous placebo-controlled studies indicated that the maximal PRL response occurs at 4 h after oral D-FEN administration (Quattrone et al., 1983). Plasma was obtained at 3 h after administration of D-FEN for assays of FEN and nor-FEN levels. This time point was selected because the pharmacokinetics of oral D-FEN show a maximal plasma level of its active metabolites D-FEN and nor-D-FEN 2 – 4 h post D-FEN (Campbell, 1991).
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
143
2.3. Beha6ioural and cardio6ascular response Subjects were asked to rate themselves using a 10 cm Visual Analogue Scale, containing four parameters (sleep, anxiety, sad, high). This self rating scale was completed at time 0 and every hour for 4 h on the two challenge sessions. Subjects were asked to report any unpleasant side effects including light headedness, cold, nausea, diarrhea, dry mouth, visual impairment, anorexia, headache or bodyache over the duration of the test and also after the test. Heart rate and blood pressure were also recorded at each time point.
2.4. Assays Plasma PRL and CORT concentrations were determined by quantitative enzyme immunoassay methods using transferable solid phase technology. The kits were supplied by Sychron Enzyme linked immunosorbent assay (SYNELISA, USA). The intra-assay and inter-assay coefficients of variation for plasma PRL were 5.0 and 7.8%, and for plasma CORT, were 5.2 and 7.1%, respectively. The lower detection limit for plasma PRL was 2 ng/ml and for plasma CORT was 0.5 ng/ml. Levels of FEN and nor-FEN were assayed by gas liquid chromatography using nitrogen sensitive detectors (Kerbs et al., 1984). The intra-assays and inter-assays coefficients of variance were reported to be 5.2 and 7.9%, respectively. The lower detection limit was 2 ng/ml.
2.5. Data analysis Statistical analyses were performed using the computer program SPSS. Lilliefors test (modified Kolmogorov – Nor test) was used to ascertain the normality of distribution. PRL and CORT data were calculated for three outcome measures: (a) Net changes in PRL and CORT were calculated by subtracting the baseline (mean of −30 and 0 time values) from hormonal changes at each time point after administration of placebo and D-FEN; (b) In order to take into account the circadian effects, placebo-controlled net changes were obtained by subtracting net changes during placebo condition from net changes at each time point post D-FEN; (c) To examine the magnitude of hormonal responses, peak hormonal responses (differences between the mean baseline level and the highest hormonal concentration after D-FEN administration) were computed. Negative values (values below baseline) were also taken into account for the analysis. Comparisons of net changes in placebo and drug conditions within subjects were performed using two-way analysis of variance with repeated measures (ANOVA) to evaluate the significance of the main effects of drug (D-FEN versus placebo), time (changes over each time point) and also drug× time interactions. Paired Student’s t-tests were used to analyse simple effects when main effects were significant. Between group comparison for sex (female versus male), and age (65–69 versus \ 70) of net changes in placebo and drug conditions and placebo-controlled net changes were performed using ANOVA and significant interactions revealed by
144
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
ANOVA were analyzed by unpaired Student’s t-tests. Intergroup comparisons of continous data and categorical data were performed with t-tests (two tailed) and chi-squared tests, respectively. Pearson’s product moment correlational analysis was used to evaluate relationships among normally distributed variables (neuroendocrine measures, metabolites D-FEN and nor-D-FEN, age and weight). The area under the response curve (AUC) was calculated using the trapezoid method for both PRL and CORT responses to D-FEN and placebo from time 0 until time 240 min (AUC 60 – AUC 240). AUC values were compared using Student’s t-tests. The statistical significance of all tests was set at PB .05 (two tailed). All values were reported as the mean9SD unless stated otherwise.
3. Results
3.1. Hormonal responses 3.1.1. Prolactin Figure 1 summarises the PRL response to D-FEN and placebo over 240 min. The increase in PRL levels began to appear 2 h post D-FEN and maximal PRL response was noted at the 4 h time point. Mean baseline PRL concentrations did not differ in drug and placebo conditions (t = − 1.20, df= 22, P= .24). The ANOVA on net PRL changes over 240 min revealed main effects of drug (F=13.83, df= 1,22, P= .001), time (F =14.43, df=3,66, P= .00), and also a significant drug by time interaction (F =18.60, df=3,66, P =.000) indicating an increase in PRL concentration in response to D-FEN. Post hoc tests showed that net PRL changes post D-FEN were significantly higher at 180 (t= 3.71, df =22, P= .000) and 240 min
Fig. 1. Mean plasma prolactin concentration after 30 mg of oral D-fenfluramine (solid line) and placebo (dashed line) in 23 elderly healthy volunteers compared to placebo. The prolactin response to Dfenfluramine was significantly elevated (ANOVA).
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
145
Fig. 2. Mean plasma cortisol concentration after 30 mg of oral D-fenfluramine (solid line) and placebo (dashed line) in 23 elderly healthy volunteers. There was a significant difference in cortisol response between placebo and D-fenfluramine (ANOVA).
(t=4.47, df = 22, P=0.00) compared to post placebo net PRL changes at the same time points. AUC 60 – 240 of net PRL responses were significantly higher following D-FEN than placebo (t =3.19, df = 22, P= .004). Among 23 subjects, 16 (69.6%) had positive peak responses and the relative increase of peak PRL response of the entire sample was 26% above the baseline estimate. The peak PRL concentration was positively correlated with plasma nor-D-FEN levels (r= .467, P=.03), mean basal CORT concentration (r= .427, P = .04). However there were no significant correlations of peak PRL response with combined plasma D-FEN and nor-D-FEN concentration (r= .341, P = .12), D-FEN levels (r=.284, P=.20), age (r= − .056 P=.80) or body mass index (r = −.323, P= .13) or mean baseline PRL concentration (r =.202, P =.35) or sex (r = .379, P =.07).
3.1.2. Cortisol CORT responses to D-FEN and placebo are shown in Fig. 2. The maximal CORT response was observed at 3 h post D-FEN and was not present at 4 h. Although baseline CORT concentrations were not statistically different between the two challenge sessions, there was a trend indicating higher baseline CORT concentration on the first day compared to the second day (t= − 1.95, df=22, P=.064) suggesting a mild stress response. When baseline CORT concentration was entered into an analysis of covariance (ANCOVA), it showed the main effect of drug (F = 6.61, df=1,22, P =.017) and drug by time interaction (F= 4.13, df= 3,66, P= .01) on net CORT changes. The effect of D-FEN on CORT changes was significant at 180 min (t =3.46, df=21, t= .002) and 240 min time points (t=2.37, df= 21, P = .027). AUC 60 – 240 of net CORT responses was greater following D-FEN than placebo (t = 2.69, df= 22, P= .013). However, mean CORT concentrations for the whole group did not rise above the baseline. The majority (53%) of the group had negative CORT responses. There was a positive correlation of peak
146
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
CORT response with plasma nor-D-FEN concentration (r=.541, P= .01), but there were no significant correlations with plasma D-FEN concentration (r= .193, P= .39), combined plasma D-FEN and nor-D-FEN concentration (r= .281, P= .21), body mass index (r = −.328, P = .13), age (r= .258, P= .23), sex (r= − .190, P= .38) or mean baseline CORT concentration (r= −.280, P= .19).
3.1.3. Gender effect ANOVA revealed that there was no main effect of gender (F= .67, df= 1,21, P=.42), or gender by time (F =1.41, df= 3,63, P= .248), or gender by drug interaction (F =1.05, df =1,21, P =.378), or gender by time by drug interaction (F = 2.27, df =3,63, P =.09) on baseline corrected net PRL changes. Similarly there was no gender effect (F = .00, df= 1,21, P= .96) or gender by time interaction on placebo controlled net PRL changes (F= .64, df= 3,63, P= .60). With respect to CORT responses, there was no gender effect (F= 1.83, df= 1,21, P=.19) or gender by time interaction (F= 1.35, df = 3,63, P= .27) on placebo controlled CORT responses. Further, there was no significant effect of gender (F = .06, df= 1,21, P =.82) on baseline net CORT changes or significant interactions between gender and time (F =1.80, df= 3,63, P= .156) or gender and drug (F= 1.83, df= 1,21, P =.190) or gender by drug by time (F= 1.35, df= 3,63, P = .27). In sum, our results suggest the absence of gender differences in net serotonergic responsivity in late life. 3.1.4. Age effect In order to determine age-related changes in D-FEN induced hormonal responses, the group was divided into young–old (65–69 years) and old–old (\70 years) on the basis of a median-split of age (69 years). Comparisons of hormonal changes between young – old and old –old were performed using ANOVA. There were no significant effects of age (F =.07, df=1.21, P= .789); age by drug (F = .41, df=1,21, P =.528); age by time (F=2.03, df= 3,63, P= .118) or age by drug by time interaction (F =1.59, df =3,63, P= .201) on net CORT changes. Similarly, young – old subjects (65 – 69 years) did not differ from old–old subjects (\ 70 years) in placebo corrected CORT changes (effect of age F= .41, df = 1,21, P= .53, age by time F =1.59, df =3,63, P =.20). There were trends indicating a weaker PRL response with advancing age (age by time interaction for placebo corrected PRL changes: F =2.50, df=3,63, P= .068; age by time by drug interaction for baseline corrected PRL changes: F=2.50, df = 3,63, P =.068). 3.2. Beha6ioural and cardio6ascular response D-FEN did not increase self-rated anxiety, sad effect, mood elevation or alertness. Two subjects complained of feeling cold following D-FEN administration. In regard to unpleasant side-effects, one male subject reported dizzy spells which lasted for 24 h after the test. One female subject was found to have an irregular heart beat after D-FEN administration. None of the subjects had nausea or vomiting. ANOVA revealed a significant drug by time interaction for heart rate and
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
147
systolic blood pressure in drug and placebo conditions. There were, however no significant main effects of drug or drug by time interaction for diastolic blood pressure. Paired t-tests revealed that heart rate at time points 120 (t= −3.95, df=15, P = .001), 180 (t = − 4.35, df= 15, P= .001), and 240 (t= − 3.3, df= 15, P= .005), and systolic blood pressure at time point 60 (t= − 3.62, df= 18, P=.002), and 120 (t = −3.60, df= 17, t= .002) were significantly lower in post D-FEN than in post placebo conditions. Thus our results suggest that cardiovascular responses were significantly higher during the first challenge test (placebo condition) than the second challenge test (D-FEN condition).
4. Discussion This placebo-controlled investigation indicates that PRL responses were more sensitive than CORT responses to acute administration of 30 mg of oral D-FEN in healthy elderly subjects. Thus, the present data clearly shows that a fixed dose of D-FEN induces a moderate increase in peak PRL concentration but weak CORT responses neutralizing only the circadian fall seen with placebo. This observation is consistent with the findings of Gorard et al. (1993) that 30 mg of oral D-FEN is more potent in stimulating PRL release than CORT in a normal sample with a wider age distribution (range 18 – 54 years). Considering the findings of the present study involving an elderly population and previous observations in non elderly subjects it might be suggested that PRL response to D-FEN is a more sensitive measure of central 5HT function than CORT response across the life span. However, Maes et al. (1991) did not find significant differences in PRL and CORT responses to 45 mg of oral D-FEN in ten healthy young normals. Similarly, Goodwin et al. (1994) failed to demonstrate either PRL or CORT responses to 30 mg of D-FEN in ten healthy young male volunteers. In contrast, some studies using a similar dose of D-FEN demonstrated substantial increase in both PRL and CORT concentrations in healthy normals (O’Keane and Dinan, 1991). The discrepancies in these results might be due to errors of inference resulting from the lack of a placebo control (O’Keane and Dinan, 1991), smaller sample size (Maes et al., 1991; Goodwin et al., 1994) and gender differences in serotonergic responsivity (Goodwin et al., 1994). Further, the interpretation of negative responses to D-FEN reported in previous studies is also made more uncertain by the failure to examine pharmacokinetic aspects in relevance to D-FEN induced hormonal responses (O’Keane and Dinan, 1991). Our findings indicate that there are no gender differences in PRL responses to D-FEN in elderly. The previous documentation of greater PRL responses in younger women have been attributed to a stimulatory role of estrogen on PRL responsiveness to D-FEN during mid phase of the menstrual cycle (O’Keane et al., 1991). However, the present study involved only post menopausal elderly women. Our study design would not permit us to examine the effects of old age versus young age on hormonal responses to D-FEN. However, in the light of our findings that there were no differences in hormonal responses between young–old and
148
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
old–old and there was an absence of correlation between age and PRL or CORT response, it can be suggested that advanced aging may have no major impact on D-FEN induced hormonal response in the elderly. This is partly in agreement with previous observations that the negative correlation between aging and PRL responses to D,L-FEN may not be linear and that the effect of aging on 5HT system would be completed by the third decade of life (McBride et al., 1990). With regard to side effect profile, our results corroborate with previous observations reporting lower incidence of unpleasant side effects with 30 mg of oral D-FEN as a challenge agent in younger subjects (O’Keane and Dinan, 1991). However, our findings contradicts the notion that there might be an age dependant decline in drug disposition. Predominantly, serotonergic mechanisms have been attributed to side effects of fenfluramine. Hence, it is possible that as age associated decline in 5HT receptors may account for lower incidence of side effects and absence of behavioural responses with D-FEN in elderly subjects, despite using a therapeutic dose. In sum, our findings strongly suggest that D-FEN is a safe drug to employ as a serotonergic probe in the elderly population. It is of interest to note that although the weight adjusted dose (0.44 mg/kg) used in this study is higher than the standard dose (0.2–0.3 mg/kg) typically prescribed as a weight reducing agent, (Guy-Grand et al., 1989) the hormonal responses were modest. This may reflect differential dose response relationships for weight reduction and hormonal response or an age-related decline in 5HT1A and 5HT2 post synaptic receptors (Marcusson et al., 1984a,b) as 5HT mediated PRL secretion seems to be dependent on 5HT1A or 5HT2A/C or interaction of both. Given the fact that the D-FEN is better tolerated in the elderly, it might be suggested that higher therapeutic doses in the range of 45–60 mg of D-FEN be may required to maximise the hormonal responses in the elderly. Surprisingly, our results indicate a positive correlation between peak PRL responses and plasma concentration of D-nor-FEN, while there was no correlation between peak PRL responses and D-FEN concentration. Although these findings remain unclear at this juncture, this may reflect the differences in activity between D-FEN and nor-D-FEN. nor-D-FEN has a relatively higher affinity for 5HT1C (Campbell, 1991) and the involvement of 5HT1C in the mediation of PRL release has been demonstrated in preclinical studies (Rittanhouse et al., 1990). Further, it is possible that plasma level of D-FEN may reach to the maximum beyond the time point selected for metabolite assay in this study or there might be a time lag between maximal D-FEN concentration and peak PRL response. The order of drug administration is an important methodological issue to be considered. In a fixed-order design, differences in stress response to the procedure between first and second challenge sessions are not balanced, which may exhibit an order effect, therefore, stress related CORT/PRL responses and cardiovascular response tend to be higher in the first than in the second session. This might have minimised the magnitude of the differences between hormonal responses to D-FEN and placebo. There is some evidence to support this suggestion as our results have shown a significant decrease in heart rate and systolic blood pressure in the second session, although the effect of D-FEN on cardiovascular responses can not be
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
149
excluded. However, since there is no statistically significant differences in baseline CORT, PRL and anxiety responses between the two challenge tests, the possibility of stress effect on hormonal responses is less evident in this study. In conclusion, we found that D-FEN is a safe putative probe to use when assessing the net serotonergic functioning in elderly subjects. Further, this study confirms that PRL responsivity to D-FEN is a reliable measure of central 5HT function in the aged. Given the age-related loss of serotonergic post-synaptic receptors (Marcusson et al., 1984a,b) and low concentration of drug in human brains relative to plasma concentrate for a given dose it might be desirable to employ a dose in the upper therapeutic range in the challenge paradigm, against the conventional wisdom of dose reduction in the elderly, in order to maximise hormonal responses to D-FEN. Failure to do so may increase the risk of masking subsensitivity of a relatively deficient 5HT system in the elderly due to understimulation, which may have an important research implication particularly when D-FEN challenge study is designed to confirm diminished serotonergic activity in psychiatric disorders of old age, especially in late onset depression.
Acknowledgements This research is supported in part by a grant from the Canadian Psychiatric Research Foundation. Authors thank staff of the Clinical Investigation Unit, Toronto Hospital, Toronto for assisting with the challenge tests and Servier Ltd for supplying D-fenfluramine and placebo capsules. We also acknowledge Thomas Cooper, Nathan Kline Institute of Psychiatric Research, New York State for measuring levels of D-fenfluramine and D-nor-fenfluramine. Dr David Streiner provided statistical consultation and Paul Miceli assisted with statistical analyses.
References Campbell, D.B., 1991. Dexfenfluramine: an overview of its mechanisms of action. Rev. Contemp. Pharmacother. 2, 93–113. Coccaro, E.F., Kavoussi, R.J., Hauger, R., 1994. PRL and HVA responses to D- and D,L-fenfluramine in humans. Biol. Psychiatry 35, 615–747. Costa, E.A., Groppetti, A., Refuelta, A., 1971. Action of fenfluramine on monoamine stores of rat tissue. Br. J. Pharmacol. 41, 57–64. Fuxe, K., Farnebo, L.O., Hamberger, B., Ogren, S.O., 1975. On the in vivo and in vitro actions of fenfluramine and its derivatives on central monoamine neurons, especially 5 hydroxytryptamine neurons and their relation to the anorectic activity of fenfluramine. Postgrad. Med. J. 551, 35 – 45. Gorard, D.A., Taylor, T.M., Medbak, S.H., Perry, L.A., Libby, G.W., Farthing, M.J., 1993. Plasma prolactin, adrenocorticotrophic hormone, and cortisol after administration of D-fenfluramine or placebo to healthy subjects. Int. Clin. Psychopharmacol. 8, 123 – 128. Goodwin, G.M., Murray, C.L., Bancroft, J., 1994. Oral D-fenfluramine and neuroendocrine challenge: problems with the 30 mg dose in men. J. Affect. Disord. 30, 117 – 122. Guy-Grand, B.J.P., Crepaldi, G., Lefebvre, P., Apfelbaum, M., Gries, A., Turner, P., 1989. International trial of long-term dexfenfluramine in obesity. Lancet ii, 1142 – 1145.
150
R. Ramasubbu et al. / Psychoneuroendocrinology 25 (2000) 139–150
Invernizzi, R., Bertorelli, R., Consolo, S., Garattini, S., Samanin, R., 1989. The effects of the L-isomer of fenfluramine and dopamine mechanism in rat brain. Eur. J. Pharmacol. 164, 241 – 248. Kerbs, H.A., Cheng, L.K., Wright, G.S., 1984. Determination of fenfluramine and nor-fenfluramine in plasma using a nitrogen sensitive detector. J. Chromatr. Biomed. Appl. 310, 412 – 417. Maes, M., D’Hondt, P., Suy, E., Minner, B., Vandervorst, C., Raus, J., 1991. HPA-axis hormones and prolactin responses to dextro-fenfluramine in depressed patients and healthy controls. Prog. NeuroPsychopharmacol. Biol. Psychiatry 15, 781– 790. Marcusson, J., Oreland, L., Winblad, B., 1984a. Effect of age on human brain serotonin (S1) binding sites. J. Neurochem. 43, 1699–1705. Marcusson, J., Morgan, D.G., Winbaild, B., Finch, C.E., 1984b. Serotonin-2 binding sites in human frontal cortex and hippocampus. Selective loss of S-2A sites with age. Brain Res. 311, 51 – 56. McBride, P.A., Tierney, H., DeMeo, M., Chen, J.S., Mann, J.J., 1990. Effect of age and gender on CNS serotonergic responsivity in normal adults. Biol. Psychiatry 27, 1143 – 1155. Muldoon, M.F., Manuck, S.B., Jansma, C.L., Moore, A.L., Perel, J., Mann, J.J., 1996. D,L-fenfluramine challenge test: experience in nonpatient sample. Biol. Psychiatry 39, 761 – 768. O’Keane, V., Dinan, T.G., 1991. Prolactin and cortisol responses to D-fenfluramine in major depression: evidence for diminished responsivity of central serotonergic function. Am. J. Psychiatry 148, 1009 – 1015. O’Keane, V., O’Hanlon, M., Webb, M., Dinan, T., 1991. D-fenfluramine/prolactin responses throughout the menstural cycle: evidence for an oestrogen-induced alteration. Clin. Endocrinol. 34, 289 – 292. Quattrone, A., Tedeschri, G., Aguglia, U., Scopacasa, F., Direnzo, G.F., AnnunZiato, L., 1983. Prolactin secretion in man: a useful to evaluate activity of drugs on central 5-hydroxy tryptaminergic neurones. Studies with fenfluramine. Br. J. Clin. Pharmacol. 16, 471 – 475. Rittanhouse, P.T., Bakkum, E.A., Herbert, G., Betha, C.L., Van de Kar, L.D., 1990. Serotonin receptor subtypes mediating neuroendocrine responses to DOI. Pharmacologist 32, 185. Ritson, E.B., Chick, J.D., 1988. In: Kendell, R.E., Zealley, A.K. (Eds.), Dependence on Alcohol and Other Drugs Comparison to Psychiatric Studies. Churchill Livingstone, Edinburgh, pp. 476 – 504. Silverstone, T., Smith, G., Richards, R., 1987. A comparative evaluation of dextrofenfluramine and dl-fenfluramine. In: Bender, A.F., Brooks, L.J. (Eds.), Bodyweight Control: The Physiology, Clinical Treatment and Prevention of Obesity. Churchill Livingstone, Edinburgh, pp. 240 – 246. Thompson, P.A., Maes, M., Meltzer, H.Y., 1994. Effect of the placebo control condition in neuroendocrine challenge studies. Psychiatry Res. 52, 317 – 326. Van de Kar, L.D., Urban, J.H., Richardson, K.D., Bethea, C.L., 1985. Pharmacological studies on the serotonergic and nonserotonin-mediated stimulation of prolactin and corticosterone secretion by fenfluramine. Neuroendocronology 41, 283– 288. Van de Kar, L.D., 1991. Neuroendocrine pharmacology of serotonergic (5-HT) neurons. Annu. Rev. Pharmacol. Toxicol. 31, 289–320. Williams, G.H., 1994. Hypertensive vascular disease. In: Isselbacher, K.J., Braunwald, E., Wilson, J.D., Martin, J.V., Fauci, A.S., Kasper, D.L. (Eds.), Harrison’s Principles of Internal Medicine, 13th edn. McGraw Hill, USA, pp. 1116–1131. Yatham, L.N., Steiner, M., 1993. Neuroendocrine probes of serotonergic function: a critical review. Life Sci. 53, 447–463.
.