Noradrenergic Activity in Anticipatory Nausea JOHN H. FETTING, MD, MICHAEL E. STEFANEK, PHD, VIVIAN R. SHEIDLER, RN, JOHN D. ELSWORTH, PHD, STEVE PIANTADOSI, MD, PHD, AND ROBERT H. ROTH, PHD Two studies were conducted to examine the hypothesis that noradrenergic activity is a cause of the anticipatory nausea associated with cancer chemotherapy. In the first study concentrations of plasma 3-methoxy-4-hydroxyphenyl-glycol (MHPG) on day 1 of cycle 5 of initial chemotherapy were significantly higher in patients with than without anticipatory nausea. To determine whether elevated MHPG reflected a clinically significant causative role for noradrenergic activity in anticipatory nausea, we conducted a randomized, double-blind, placebocontrolled, crossover trial of clonidine for anticipatory nausea. At a dose of clonidine that produced significant side effects and reductions of plasma MHPG. anticipatory nausea was improved only marginally. These studies do not support a causative role for noradrenergic activity in anticipatory nausea that can be reduced by clonidine with an acceptable therapeutic index. Key words: anticipatory nausea and vomiting; norepinephrine; 3-methoxy-4-hydroxyphenylglycol (MHPG); clonidine.
INTRODUCTION
Anticipatory nausea develops in patients who have had postchemotherapy nausea and vomiting after previous cycles of cancer chemotherapy. Anticipatory nausea has been observed in approximately one-third of patients receiving chemotherapy (1). It may be a cause of noncompliance with potentially curative chemotherapy. Anticipatory nausea is classically conditioned (2). Little is
From the Johns Hopkins Oncology Center, Baltimore, Maryland (J.H.F., M.E.S., V.R.S., S.P.); and The Department of Pharmacology, Yale University School of Medicine, New Haven, Connecticut Q.D.E., R.H.R.). Address reprint requests to: John H. Fetting, M.D., The Johns Hopkins Oncology Center, 600 N. Wolfe Street, Room 135, Baltimore, MD 21287. Received for publication May 6, 1992; revision received August 21, 1992
Psychosomatic Medicine 54:641-647 (1992) 0033-3174/92/5406-0641 $03.00/0 Copyright © 1992 by the American Psychosomatic Society
known, however, about the basic mechanisms that produce it. Several lines of evidence suggest a role for central noradrenergic activity. Noradrenergic neurons in the locus coeruleus may be involved in the central response to stress and have been implicated in certain types of anxiety (3, 4). Patients receiving repeated cycles of emetogenic chemotherapy are stressed, and increasing anxiety has been shown to precede the development of anticipatory nausea (5). Increased noradrenergic activity might contribute to anticipatory nausea in two ways. The first involves direct activation of neurons controlling nausea and vomiting. Noradrenergic neurons synapse in and near the medullary centers for vomiting, providing an anatomic basis for a role in anticipatory symptoms (6). Also, intracerebroventricular norepinephrine predictably produces vomiting, suggesting a role for noradrenergic receptor mediated emesis (7). A second way in 641
J. H. FETTING et al.
which stress-induced noradrenergic activity might contribute to anticipatory nausea is via alerting and memory functions likely to be important in developing and identifying relevant conditioned stimuli (8). We report on two studies conducted to examine the hypothesis that noradrenergic activity is a cause of anticipatory nausea and vomiting. In the first study we compared plasma free 3-methoxy-4-hydroxyphenyl-glycol (MHPG) in patients with and without anticipatory nausea. A major metabolite of both central and peripheral noradrenergic activity, MHPG has been employed as a marker of noradrenergic activity (9, 10). Although plasma MHPG originates from both peripheral and central sources, significant correlations between brain, CSF, and plasma MHPG have been observed in human and nonhuman primates, which probably arise not only from mixing between the central and peripheral pools but also from the interactive linkage between the peripheral and central adrenergic systems (11). Our decision to employ plasma MHPG as a marker of central as well as peripheral noradrenergic activity was based on the highly significant correlation between the two and the ease with which plasma MHPG can be obtained (9-11). It was hypothesized that plasma free MHPG would be higher in patients with anticipatory nausea. This hypothesis was confirmed. The second study was a randomized, double-blind, placebo-controlled, crossover trial of clonidine for anticipatory nausea. If the elevation of plasma MHPG found in patients with anticipatory nausea indicated a causal role for noradrenergic activity in anticipatory nausea, then it was hypothesized that clonidine, an alpha-2 noradrenergic receptor agonist, would reduce 642
anticipatory nausea, perhaps by decreasing norepinephrine release through its effect on presynaptic alpha-2 receptors (12).
METHODS In the first study we compared plasma free MHPG concentrations in patients with and without anticipatory nausea on day 1 cycle 5 of their initial chemotherapy regimen. All patients had histories of postchemotherapy nausea and vomiting after their previous four cycles of chemotherapy. Studying patients on cycle 5 provided an adequate number of cycles for conditioning and a large enough pool of potential subjects (5) We limited the study to patients with postchemotherapy nausea and vomiting because such patients frequently suffer from anticipatory symptoms (13). In this fashion we hoped to maximize the fraction of patients with anticipatory symptoms. Patients were excluded if they would be taking any medication likely to affect noradrenergic function during the week preceding chemotherapy administration on day 1 cycle 5 or if they had a medical problem likely to affect noradrenergic function (e.g., heart failure, asthma, fever, etc.). Patients agreed not to take antiemetics before their intravenous was started for chemotherapy administration. They followed a low vanillymandelic acid (VMA) diet for 3 days prior to day 1 cycle 5 to minimize fluctuations in MHPG due to diet. Written informed consent was obtained. In the clinic on day 1 cycle 5. after the chemotherapy intravenous was started, 10 cc of blood was withdrawn for assay of MHPG according to the method of Elsworth et al. (14). Assays were performed without knowledge of whether the patient had experienced anticipatory nausea. After receiving chemotherapy, patients were administered the Morrow Assessment of Nausea and Emesis (MANE) to assess any anticipatory symptoms associated with day 1 cycle 5 (15). Anticipatory symptoms were defined as nausea or vomiting occurring before or during intravenous chemotherapy administration Symptoms developing during infusion of drugs with relatively short emetic latencies (e.g , dacarbazine, platinum), however, were classified as postchemotherapy rather than anticipatory (16). The second study was a randomized, placebo-
Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA controlled, double-blind, crossover trial of clonidine for anticipatory nausea. Patients with anticipatory nausea or vomiting were eligible if they were to receive two or more cycles of chemotherapy. Patients were informed that behavioral techniques that reliably reduce anticipatory nausea were available as an alternative to study participation. The other eligibility requirements and study procedures were the same as with the first study. Clonidine was administered in a dose of 2 MgAg by mouth bid for 72 hours prior to chemotherapy administration. The dose of clonidine was determined in a preliminary Phase I-II study (17). Clonidine was administered for 72 hours prior to chemotherapy so that plasma concentrations would be at or near steady state at the time of chemotherapy administration when anticipatory symptoms were expected to be most severe (18, 19). Patients took their last dose of clonidine in the morning on the day of chemotherapy administration. Patients were randomized to receive clonidine or an identical-appearing placebo prior to their first on-study cycle of chemotherapy. They received the alternative prior to their next cycle of chemotherapy. In the clinic after the chemotherapy intravenous was started, blood was drawn for MHPG determination. After receiving chemotherapy, patients completed the MANE to evaluate their anticipatory nausea and vomiting. Patients also rated the severity of sedation and dry mouth as absent (0). mild (1), modorate (2), or severe (3). These two toxicities were the most commonly reported side effects of clonidine in the Phase I-II study (17). After the assessments of nausea, vomiting, and clonidine side effects, patients were asked about compliance with the study medication. The number oi missed doses were noted. For the first study a sample of 30 patients was planned. Based on previous work it was estimated that half the patients would have anticipatory nausea or vomiting and half would not (1, 5. 13). Published reports of plasma MHPG concentrations suggested that a comparison of 15 patients with and without anticipatory nausea or vomiting would provide more than adequate power to delect potentially clinically significant differences in MHPG concentrations (10, 11, 14). In the first study the statistical significance of differences in mean concentrations of plasma MHPG between patients with and without anticipatory nausea or vomiting was determined by the I test for independent samples of unequal size. The relationship between anticipatory nausea severity and plasma MHPG was examined by the Pearson correlation coefficient.
Psychosomatic Medicine 54:641-647 (1992)
For the second study a sample of 20 patients was planned to provide 80% power to detect a 0.35 difference in efficacy between the clonidine and placebo conditions (assuming a 0.55 frequency of anticipatory nausea or vomiting on the clonidine arm and a 0.90 frequency on placebo) at an a level of 0.05. This study was discontinued, however, after studying 11 patients when it was recognized that the therapeutic index of clonidine was unfavorable (i.e., significant side effects were encountered). Data analysis in the second study featured examination of mean differences in anticipatory nausea, sedation, dry mouth, and plasma MHPG concentration between clonidine and placebo. These analyses tested the hypothesis that the mean difference between clonidine and placebo was zero. Two-tailed levels of statistical significance are reported for both studies.
RESULTS
Twenty-nine patients were recruited for the first study from August 1985 to April 1987. Five were later excluded for the following reasons: one took diazepam before treatment, in one the chemotherapy intravenous could not be started, and three were studied but later ruled ineligible when evaluation revealed that they had not had vomiting after each of their previous treatments. Excluding these three subjects from the analyses did not significantly affect the findings. The remaining 24 subjects included 18 females (75%); the sample had a mean age of 41.6 ± SD 15.3. Diagnoses included: Stage II-IV breast cancer (10), Hodgkin's and non-Hodgkin's lymphoma (8), gynecologic malignancy (3), and pediatric malignancy (3). Patients were receiving a variety of chemotherapy regimens. Eight of 24 (33%) had anticipatory nausea on day 1 of cycle 5. In the eight patients with anticipatory nausea, nausea severity ranged from very mild (1) to moderate (3) on a 0 to 6 scale. Plasma MHPG concentrations are reported on 22 patients 643
|. H. FETTING et al.
since in two patients there were technical difficulties in sample preparation. The relationship between MHPG and anticipatory nausea was examined in two ways. First, MHPG concentrations in patients with and without anticipatory nausea were compared. Mean plasma free MHPG ± SD in patients with anticipatory nausea (n = 7) was 3.9 ± 0.8 ng/ml vs. 2.9 ± 0.8 ng/ml in patients without anticipatory nausea (n = 15], (t = 7.13, d/20, p < 0.025). Second, the correlation between anticipatory nausea severity on a 0 to 6 scale and MHPG concentration was examined. The correlation was r = 0.41 (p < 0.10, df 20). Sixteen patients were recruited for the second study from August 1985 to February 1990. Three patients were subsequently ruled ineligible. In all three ineligibility was related to concomitant use of drugs that could interfere with noradrenergic activity. One patient was admitted to the study incorrectly despite the fact that she was taking antihypertensive medications. Two patients were started on drugs that might affect noradrenergic activity. The remaining 13 patients included 11 females (85%) and had a mean age of 38 ± SD 16. Diagnoses included: Stage I-IV breast cancer (6), colon cancer (2), Hodgkin's and non-Hodgkin's lymphoma (2), sarcoma (2), and acute lymphoblastic leukemia (1). These 13 patients were receiving 10 different chemotherapy regimens. Of 13 eligible patients, 11 completed the study. Two patients withdrew from the study after the first cycle because of side effects they attributed to the study drug. Both had received clonidine. One complained of severe difficulty sleeping and severe urinary frequency; the other developed a severe migraine headache. Results are presented for the 11 patients who 644
completed the study. Self-reported compliance was good. Eight patients reported taking clonidine and placebo as directed. One patient reported missing two doses of placebo; another missed one dose of placebo. A third patient missed two doses of clonidine. In order to test the hypothesis that noradrenergic activity contributes to anticipatory nausea, it was necessary for clonidine to reduce MHPG. It did. There was a significant 0.52 ± 0.7 ng/ml reduction in plasma MHPG from a mean of 3.32 ± 0.89 ng/ml on placebo to 2.80 ± 0.59 ng/ ml on clonidine (p = 0.03). The next analysis compares the severity of anticipatory nausea on placebo and clonidine. On a 0 to 6 severity scale, mean anticipatory nausea on placebo was 2.00 ± 1.09 vs. 1.27 ± 1.68 on clonidine. The mean reduction in anticipatory nausea from placebo to clonidine was 0.73 ± 1.85 (p = 0.22). Although not depicted, there was also no significant difference in anticipatory nausea duration between placebo and clonidine. Anticipatory vomiting occurred in three patients on clonidine and four patients on placebo; mean vomiting severity scores were identical in the two groups. We further examined the relationship between anticipatory nausea and noradrenergic activity by determining the correlation between the difference in anticipatory nausea severity between placebo and clonidine (i.e., AAN) and the difference in MHPG between placebo and clonidine (i.e., AMHPG). The correlation coefficient was r = 0.36 (p = 0.20) and R2 = 0.13. Side effects on clonidine and placebo were next compared. On the 0 (absent) to 3 (severe) scale, mean dry mouth on clonidine was 1.91 ± 0.94 vs. 0.64 ± 0.81 on placebo. The mean difference was 1.27 ± Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA
0.79 (p < 0.001). Mean sedation on clonidine was 2.54 ± 0.52 vs. 1.27 ± 0.79 on placebo. The mean difference was 1.27 ± 0.90 (p < 0.001). Six of 11 patients had severe (i.e., 3) dry mouth or sedation on clonidine vs. none on placebo. There was, however, no hypotension with clonidine. We also examined the correlations between side effects and control of anticipatory nausea. The correlation between the difference in sedation between clonidine and placebo and the difference in anticipatory nausea between clonidine and placebo was r = -0.11 (p = 0.75). The correlation between the difference in dry mouth between clonidine and placebo and the difference in anticipatory nausea between clonidine and placebo was r = -0.63 (p = 0.04).
DISCUSSION The comparison of plasma MHPG concentrations in patients with and without anticipatory nausea from the first study suggests that anticipatory nausea is associated with an increase in noradrenergic activity. It was hoped that the second study would determine whether this increase in noradrenergic activity signals a causative role for noradrenergic neurons in the development of anticipatory nausea. The clonidine study was closed, however, before this question could be definitively answered once it became clear that the therapeutic index of clonidine was unfavorable. The limited power of the second study does not permit us to exclude an effect of clonidine on anticipatory nausea. As a result, the question of whether noradrenergic activity is a cause of anticipatory nausea is not answered by these studies. Psychosomatic Medicine 54:641-647 (1992)
Do these studies permit the conclusion that if noradrenergic activity plays a role in anticipatory nausea, it is one that it too small to be clinically exploitable? We tentatively suggest that they do. Examination of the R2's for the correlations between anticipatory nausea and MHPG in both studies suggests that even if noradrenergic activity plays a causative role in anticipatory nausea, it is small. In the first study the R2 of the correlation between anticipatory nausea severity and MHPG was 0.17. The R2 of the correlation between AAN and AMHPG in the second study was 0.13. Therefore, variance in MHPG, a marker of noradrenergic activity, accounts for only a small amount of the variance in anticipatory nausea severity. These statistics suggest that even if noradrenergic activity is a cause of anticipatory nausea, its contribution is modest. If noradrenergic activity plays only a modest role in anticipatory nausea, then clonidine, even at toxic doses, should only partially reduce anticipatory nausea. The findings of the second study are consistent with this suggestion. The dose of clonidine employed in this study produced a significant reduction in MHPG and severe dry mouth or sedation in six of 11 patients but only minimal effects on anticipatory nausea. The significant correlation between clonidine-induced dry mouth and improvement in anticipatory nausea is of note. This finding may indicate that these phenomena share underlying neurobiological mechanisms. On the other hand, since we did not hypothesize this relationship a priori and multiple tests of significance were performed, the finding may represent nothing more than chance. Two aspects of these studies limit our ability to exclude more confidently noradrenergic activity as a clinically exploit645
J. H. FETTING et al.
able cause of anticipatory nausea. The clonidine might provide more consistent first, already alluded to above, is the small effects on anticipatory nausea but would samples for both studies. Only seven pa- produce more toxicity. Assuming that clotients in the first study had anticipatory nidine reduced anticipatory nausea symptoms thereby limiting the ability to through its effect on inhibitory presynexamine the correlation between MHPG aptic alpha-2 receptors, a lower dose of and anticipatory nausea. The small num- clonidine would be unlikely to control bers in the second study limit the ability anticipatory nausea better even if it did to exclude a role for noradrenergic activ- reduce toxicity. This is because the lower ity in anticipatory nausea. The second dose would produce even less inhibition limitation is the relatively mild anticipa- than the dose used in this study. Certory symptoms experienced by patients in tainly, the therapeutic index of clonidine both studies. In the first study the seven is distinctly inferior to the behavioral patients who experienced anticipatory techniques that are the treatment of nausea rated it from very mild to only choice for anticipatory nausea and vommoderate. In the second study anticipa- iting (2). tory symptoms in the placebo condition Although clonidine proved disappointaveraged 2.0 (i.e., mild). The mild symp- ing, there are other pharmacologic aptoms in both studies almost certainly re- proaches to anticipatory nausea. Perhaps flect the decreasing severity ot anticipa- the most obvious is to control postchemtory nausea and vomiting during the time otherapy nausea and vomiting. The adthe studies were conducted (1). This re- vent of 5HT3 antagonists has marked a duction in the severity of anticipatory significant advance in efforts to control symptoms has resulted from improved postchemotherapy symptoms (20). There control of postchemotherapy nausea and is also evidence from a small clinical trial vomiting. Regardless of cause, the limited that the benzodiazapine, alprazolam, rerange of anticipatory symptoms made it duces anticipatory nausea (21). In a more more difficult to examine the relationship speculative vein, recent reports demonbetween noradrenergic activity and antic- strate that the acquisition of conditioned ipatory nausea. If we had patients with fear responses can be blocked by drugs more severe symptoms in our sample, a that are receptor antagonists of N-methylmodest but significant association be- D-aspartate (22, 23). Perhaps N-methyl-Dtween noradrenergic activity and antici- aspartate plays a causative role in anticipatory nausea may have been detected. patory nausea and N-methyl-D-aspartate Finally, although the second study antagonists (e.g., dextromethorphan) may lacked the power to exclude an effect of reduce anticipatory nausea. clonidine on anticipatory nausea, it at least demonstrates that the therapeutic Supported by American Cancer Society index of clonidine for anticipatory nausea Grant PBR-19. is narrow. The dose of clonidine emWe thank Stephanie Young /or her outployed in the second study produced sig- standing administrative assistance, Grenificant side effects but minimal effects gory Dole for technical assistance, and on anticipatory nausea. A larger dose of John EnterJine for statistical consultation.
646
Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA REFERENCES 1. Stefanek ME, Sheidler VR, Fetting JH: Anticipatory nausea and vomiting: Does it remain a significanl clinical problem? Cancer 62:2654, 1988 2. Burish TG, Carey MP: Conditioned responses to cancer chemotherapy: Etiology and treatment. In Fox BH, Newberry BH (eds) Impact of Psychoendocrine Systems in Cancer and Immunity. Toronto, CJ Hogrefe, 1983 3. Glavin GB Stress and brain noradrenaline: A review. Neurosci Behav Rev 9:233-245, 1985 4. Charney DS, Redmond DE: Neurobiological mechanisms in human anxiety. Neuropharmacology 22.1531-1536, 1983 5. Andrykowski MA, Redd WH: Longitudinal analysis of the development of anticipatory nausea in cancer chemotherapy. J Consult Clin Psychol 53:447-454, 1985 6. Fuxe K, Owman C: Cellular localization of monoamines in the area postrema of certain mammals. J Comp Neurol 125:337-354, 1965 7 Jenkins LC, Lahay D: Central mechanisms of vomiting related to catecholamine response: Anaesthetic implications. Can Anaesth Soc J 18:434-441, 1971 8. Sara SJ: Noradrenergic modulation of selective attention: Its role in memory retrieval. Ann NY Acad Sci 444:178-193, 1985 9. Maas JW: Relationships between central nervous system noradrenergic function and plasma and urinary concentrations of norepinephrine metabolites. In Usdin E, et al. (eds), Frontiers in Biochemical and Pharmacological Research in Depression. New York, Raven Press, 1984, 45-55 10. Jimerson DC. Balcnger JC, Lake CR, et al: Plasma and CSF MHPG in normals. Psychopharmacol Bull 17:86-87, 1981 11. Kopin IJ, Gordon EK, Jimerson DC, et al: Relation between plasma and cerebrospinal fluid levels of 3methoxy-4-hydroxyphenylglycol. Science 219:73-75, 1983 12. Aghajanian GK: Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature 276:186-188, 1978 13. Nerenz DR. Leventhal H, Easterling DV, et al: Anxiety and drug taste as predictors of anticipatory nausea in cancer chemotherapy. J Clin Oncol 4:224-233, 1986 14. Elsworth JD, Roth RH, Redmond DE Jr: The relative importance of 3-methoxy-4-hydroxyphenylethyleneglycol (MHPG) and 3,4 dihydroxyphenylglycol (DHPG) as norepinephrine metabolites in rat, monkey, and man. J Neurochem 41:786-793, 1983 15. Morrow GR. The assessment of nausea and vomiting. Cancer 53(Suppl):2267-2278.1984 16. Andrykowski MA: Definitional issues in the study of anticipatory nausea in cancer chemotherapy. J Behav Med 9:33-41, 1986 17. Fetting JH, Sheidler VR, Stefanek ME, et al: Clonidine for anticipatory nausea and vomiting: A pilot study examining dose-toxicity relationships and potential for further study. Cancer Treat Rep 71:409410, 1987 18. Pettinger WA: Pharmacology of clonidine. J Cardiovasc Pharmacol 2(Supp 1):S21-S28, 1980 19. Frisk-Holmberg M: Clinical pharmacology of clonidine. Chest 83(Suppl):395-397,1983 20. De Mulder PHM, Seynaeve C, Vermoken JB, et al: Ondansetron compared with high-dose metoclopramide in prophylaxis of acute and delayed cisplatm-mduced nausea and vomiting. Ann Int Med 113:834840, 1990 21. Greenberg DB, Surman OS, Clarke J, et al: Alprazolam for phobic nausea and vomiting related to cancer chemotherapy [letter]. Cancer Treat Rep 71:549-550, 1987 22. Staubli U, Thibault O, DiLorenzo M, et al: Antagonism of NMDA receptors impairs acquisition but not retention of olfactory memory. Behav Neurosci 103:54-60, 1989 23. Miserendino MJD, Sananes CB, Melia KR, et al: Blocking of acquisition but not expression of conditioned fear-potentiated startle by NMDA antagonists in the amygdala. Nature 345:716-718, 1990
Psychosomatic Medicine 54:641-647 (1992)
647
INTRODUCTION
Anticipatory nausea develops in patients who have had postchemotherapy nausea and vomiting after previous cycles of cancer chemotherapy. Anticipatory nausea has been observed in approximately one-third of patients receiving chemotherapy (1). It may be a cause of noncompliance with potentially curative chemotherapy. Anticipatory nausea is classically conditioned (2). Little is
From the Johns Hopkins Oncology Center, Baltimore, Maryland (J.H.F., M.E.S., V.R.S., S.P.); and The Department of Pharmacology, Yale University School of Medicine, New Haven, Connecticut Q.D.E., R.H.R.). Address reprint requests to: John H. Fetting, M.D., The Johns Hopkins Oncology Center, 600 N. Wolfe Street, Room 135, Baltimore, MD 21287. Received for publication May 6, 1992; revision received August 21, 1992
Psychosomatic Medicine 54:641-647 (1992) 0033-3174/92/5406-0641 $03.00/0 Copyright © 1992 by the American Psychosomatic Society
known, however, about the basic mechanisms that produce it. Several lines of evidence suggest a role for central noradrenergic activity. Noradrenergic neurons in the locus coeruleus may be involved in the central response to stress and have been implicated in certain types of anxiety (3, 4). Patients receiving repeated cycles of emetogenic chemotherapy are stressed, and increasing anxiety has been shown to precede the development of anticipatory nausea (5). Increased noradrenergic activity might contribute to anticipatory nausea in two ways. The first involves direct activation of neurons controlling nausea and vomiting. Noradrenergic neurons synapse in and near the medullary centers for vomiting, providing an anatomic basis for a role in anticipatory symptoms (6). Also, intracerebroventricular norepinephrine predictably produces vomiting, suggesting a role for noradrenergic receptor mediated emesis (7). A second way in 641
J. H. FETTING et al.
which stress-induced noradrenergic activity might contribute to anticipatory nausea is via alerting and memory functions likely to be important in developing and identifying relevant conditioned stimuli (8). We report on two studies conducted to examine the hypothesis that noradrenergic activity is a cause of anticipatory nausea and vomiting. In the first study we compared plasma free 3-methoxy-4-hydroxyphenyl-glycol (MHPG) in patients with and without anticipatory nausea. A major metabolite of both central and peripheral noradrenergic activity, MHPG has been employed as a marker of noradrenergic activity (9, 10). Although plasma MHPG originates from both peripheral and central sources, significant correlations between brain, CSF, and plasma MHPG have been observed in human and nonhuman primates, which probably arise not only from mixing between the central and peripheral pools but also from the interactive linkage between the peripheral and central adrenergic systems (11). Our decision to employ plasma MHPG as a marker of central as well as peripheral noradrenergic activity was based on the highly significant correlation between the two and the ease with which plasma MHPG can be obtained (9-11). It was hypothesized that plasma free MHPG would be higher in patients with anticipatory nausea. This hypothesis was confirmed. The second study was a randomized, double-blind, placebo-controlled, crossover trial of clonidine for anticipatory nausea. If the elevation of plasma MHPG found in patients with anticipatory nausea indicated a causal role for noradrenergic activity in anticipatory nausea, then it was hypothesized that clonidine, an alpha-2 noradrenergic receptor agonist, would reduce 642
anticipatory nausea, perhaps by decreasing norepinephrine release through its effect on presynaptic alpha-2 receptors (12).
METHODS In the first study we compared plasma free MHPG concentrations in patients with and without anticipatory nausea on day 1 cycle 5 of their initial chemotherapy regimen. All patients had histories of postchemotherapy nausea and vomiting after their previous four cycles of chemotherapy. Studying patients on cycle 5 provided an adequate number of cycles for conditioning and a large enough pool of potential subjects (5) We limited the study to patients with postchemotherapy nausea and vomiting because such patients frequently suffer from anticipatory symptoms (13). In this fashion we hoped to maximize the fraction of patients with anticipatory symptoms. Patients were excluded if they would be taking any medication likely to affect noradrenergic function during the week preceding chemotherapy administration on day 1 cycle 5 or if they had a medical problem likely to affect noradrenergic function (e.g., heart failure, asthma, fever, etc.). Patients agreed not to take antiemetics before their intravenous was started for chemotherapy administration. They followed a low vanillymandelic acid (VMA) diet for 3 days prior to day 1 cycle 5 to minimize fluctuations in MHPG due to diet. Written informed consent was obtained. In the clinic on day 1 cycle 5. after the chemotherapy intravenous was started, 10 cc of blood was withdrawn for assay of MHPG according to the method of Elsworth et al. (14). Assays were performed without knowledge of whether the patient had experienced anticipatory nausea. After receiving chemotherapy, patients were administered the Morrow Assessment of Nausea and Emesis (MANE) to assess any anticipatory symptoms associated with day 1 cycle 5 (15). Anticipatory symptoms were defined as nausea or vomiting occurring before or during intravenous chemotherapy administration Symptoms developing during infusion of drugs with relatively short emetic latencies (e.g , dacarbazine, platinum), however, were classified as postchemotherapy rather than anticipatory (16). The second study was a randomized, placebo-
Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA controlled, double-blind, crossover trial of clonidine for anticipatory nausea. Patients with anticipatory nausea or vomiting were eligible if they were to receive two or more cycles of chemotherapy. Patients were informed that behavioral techniques that reliably reduce anticipatory nausea were available as an alternative to study participation. The other eligibility requirements and study procedures were the same as with the first study. Clonidine was administered in a dose of 2 MgAg by mouth bid for 72 hours prior to chemotherapy administration. The dose of clonidine was determined in a preliminary Phase I-II study (17). Clonidine was administered for 72 hours prior to chemotherapy so that plasma concentrations would be at or near steady state at the time of chemotherapy administration when anticipatory symptoms were expected to be most severe (18, 19). Patients took their last dose of clonidine in the morning on the day of chemotherapy administration. Patients were randomized to receive clonidine or an identical-appearing placebo prior to their first on-study cycle of chemotherapy. They received the alternative prior to their next cycle of chemotherapy. In the clinic after the chemotherapy intravenous was started, blood was drawn for MHPG determination. After receiving chemotherapy, patients completed the MANE to evaluate their anticipatory nausea and vomiting. Patients also rated the severity of sedation and dry mouth as absent (0). mild (1), modorate (2), or severe (3). These two toxicities were the most commonly reported side effects of clonidine in the Phase I-II study (17). After the assessments of nausea, vomiting, and clonidine side effects, patients were asked about compliance with the study medication. The number oi missed doses were noted. For the first study a sample of 30 patients was planned. Based on previous work it was estimated that half the patients would have anticipatory nausea or vomiting and half would not (1, 5. 13). Published reports of plasma MHPG concentrations suggested that a comparison of 15 patients with and without anticipatory nausea or vomiting would provide more than adequate power to delect potentially clinically significant differences in MHPG concentrations (10, 11, 14). In the first study the statistical significance of differences in mean concentrations of plasma MHPG between patients with and without anticipatory nausea or vomiting was determined by the I test for independent samples of unequal size. The relationship between anticipatory nausea severity and plasma MHPG was examined by the Pearson correlation coefficient.
Psychosomatic Medicine 54:641-647 (1992)
For the second study a sample of 20 patients was planned to provide 80% power to detect a 0.35 difference in efficacy between the clonidine and placebo conditions (assuming a 0.55 frequency of anticipatory nausea or vomiting on the clonidine arm and a 0.90 frequency on placebo) at an a level of 0.05. This study was discontinued, however, after studying 11 patients when it was recognized that the therapeutic index of clonidine was unfavorable (i.e., significant side effects were encountered). Data analysis in the second study featured examination of mean differences in anticipatory nausea, sedation, dry mouth, and plasma MHPG concentration between clonidine and placebo. These analyses tested the hypothesis that the mean difference between clonidine and placebo was zero. Two-tailed levels of statistical significance are reported for both studies.
RESULTS
Twenty-nine patients were recruited for the first study from August 1985 to April 1987. Five were later excluded for the following reasons: one took diazepam before treatment, in one the chemotherapy intravenous could not be started, and three were studied but later ruled ineligible when evaluation revealed that they had not had vomiting after each of their previous treatments. Excluding these three subjects from the analyses did not significantly affect the findings. The remaining 24 subjects included 18 females (75%); the sample had a mean age of 41.6 ± SD 15.3. Diagnoses included: Stage II-IV breast cancer (10), Hodgkin's and non-Hodgkin's lymphoma (8), gynecologic malignancy (3), and pediatric malignancy (3). Patients were receiving a variety of chemotherapy regimens. Eight of 24 (33%) had anticipatory nausea on day 1 of cycle 5. In the eight patients with anticipatory nausea, nausea severity ranged from very mild (1) to moderate (3) on a 0 to 6 scale. Plasma MHPG concentrations are reported on 22 patients 643
|. H. FETTING et al.
since in two patients there were technical difficulties in sample preparation. The relationship between MHPG and anticipatory nausea was examined in two ways. First, MHPG concentrations in patients with and without anticipatory nausea were compared. Mean plasma free MHPG ± SD in patients with anticipatory nausea (n = 7) was 3.9 ± 0.8 ng/ml vs. 2.9 ± 0.8 ng/ml in patients without anticipatory nausea (n = 15], (t = 7.13, d/20, p < 0.025). Second, the correlation between anticipatory nausea severity on a 0 to 6 scale and MHPG concentration was examined. The correlation was r = 0.41 (p < 0.10, df 20). Sixteen patients were recruited for the second study from August 1985 to February 1990. Three patients were subsequently ruled ineligible. In all three ineligibility was related to concomitant use of drugs that could interfere with noradrenergic activity. One patient was admitted to the study incorrectly despite the fact that she was taking antihypertensive medications. Two patients were started on drugs that might affect noradrenergic activity. The remaining 13 patients included 11 females (85%) and had a mean age of 38 ± SD 16. Diagnoses included: Stage I-IV breast cancer (6), colon cancer (2), Hodgkin's and non-Hodgkin's lymphoma (2), sarcoma (2), and acute lymphoblastic leukemia (1). These 13 patients were receiving 10 different chemotherapy regimens. Of 13 eligible patients, 11 completed the study. Two patients withdrew from the study after the first cycle because of side effects they attributed to the study drug. Both had received clonidine. One complained of severe difficulty sleeping and severe urinary frequency; the other developed a severe migraine headache. Results are presented for the 11 patients who 644
completed the study. Self-reported compliance was good. Eight patients reported taking clonidine and placebo as directed. One patient reported missing two doses of placebo; another missed one dose of placebo. A third patient missed two doses of clonidine. In order to test the hypothesis that noradrenergic activity contributes to anticipatory nausea, it was necessary for clonidine to reduce MHPG. It did. There was a significant 0.52 ± 0.7 ng/ml reduction in plasma MHPG from a mean of 3.32 ± 0.89 ng/ml on placebo to 2.80 ± 0.59 ng/ ml on clonidine (p = 0.03). The next analysis compares the severity of anticipatory nausea on placebo and clonidine. On a 0 to 6 severity scale, mean anticipatory nausea on placebo was 2.00 ± 1.09 vs. 1.27 ± 1.68 on clonidine. The mean reduction in anticipatory nausea from placebo to clonidine was 0.73 ± 1.85 (p = 0.22). Although not depicted, there was also no significant difference in anticipatory nausea duration between placebo and clonidine. Anticipatory vomiting occurred in three patients on clonidine and four patients on placebo; mean vomiting severity scores were identical in the two groups. We further examined the relationship between anticipatory nausea and noradrenergic activity by determining the correlation between the difference in anticipatory nausea severity between placebo and clonidine (i.e., AAN) and the difference in MHPG between placebo and clonidine (i.e., AMHPG). The correlation coefficient was r = 0.36 (p = 0.20) and R2 = 0.13. Side effects on clonidine and placebo were next compared. On the 0 (absent) to 3 (severe) scale, mean dry mouth on clonidine was 1.91 ± 0.94 vs. 0.64 ± 0.81 on placebo. The mean difference was 1.27 ± Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA
0.79 (p < 0.001). Mean sedation on clonidine was 2.54 ± 0.52 vs. 1.27 ± 0.79 on placebo. The mean difference was 1.27 ± 0.90 (p < 0.001). Six of 11 patients had severe (i.e., 3) dry mouth or sedation on clonidine vs. none on placebo. There was, however, no hypotension with clonidine. We also examined the correlations between side effects and control of anticipatory nausea. The correlation between the difference in sedation between clonidine and placebo and the difference in anticipatory nausea between clonidine and placebo was r = -0.11 (p = 0.75). The correlation between the difference in dry mouth between clonidine and placebo and the difference in anticipatory nausea between clonidine and placebo was r = -0.63 (p = 0.04).
DISCUSSION The comparison of plasma MHPG concentrations in patients with and without anticipatory nausea from the first study suggests that anticipatory nausea is associated with an increase in noradrenergic activity. It was hoped that the second study would determine whether this increase in noradrenergic activity signals a causative role for noradrenergic neurons in the development of anticipatory nausea. The clonidine study was closed, however, before this question could be definitively answered once it became clear that the therapeutic index of clonidine was unfavorable. The limited power of the second study does not permit us to exclude an effect of clonidine on anticipatory nausea. As a result, the question of whether noradrenergic activity is a cause of anticipatory nausea is not answered by these studies. Psychosomatic Medicine 54:641-647 (1992)
Do these studies permit the conclusion that if noradrenergic activity plays a role in anticipatory nausea, it is one that it too small to be clinically exploitable? We tentatively suggest that they do. Examination of the R2's for the correlations between anticipatory nausea and MHPG in both studies suggests that even if noradrenergic activity plays a causative role in anticipatory nausea, it is small. In the first study the R2 of the correlation between anticipatory nausea severity and MHPG was 0.17. The R2 of the correlation between AAN and AMHPG in the second study was 0.13. Therefore, variance in MHPG, a marker of noradrenergic activity, accounts for only a small amount of the variance in anticipatory nausea severity. These statistics suggest that even if noradrenergic activity is a cause of anticipatory nausea, its contribution is modest. If noradrenergic activity plays only a modest role in anticipatory nausea, then clonidine, even at toxic doses, should only partially reduce anticipatory nausea. The findings of the second study are consistent with this suggestion. The dose of clonidine employed in this study produced a significant reduction in MHPG and severe dry mouth or sedation in six of 11 patients but only minimal effects on anticipatory nausea. The significant correlation between clonidine-induced dry mouth and improvement in anticipatory nausea is of note. This finding may indicate that these phenomena share underlying neurobiological mechanisms. On the other hand, since we did not hypothesize this relationship a priori and multiple tests of significance were performed, the finding may represent nothing more than chance. Two aspects of these studies limit our ability to exclude more confidently noradrenergic activity as a clinically exploit645
J. H. FETTING et al.
able cause of anticipatory nausea. The clonidine might provide more consistent first, already alluded to above, is the small effects on anticipatory nausea but would samples for both studies. Only seven pa- produce more toxicity. Assuming that clotients in the first study had anticipatory nidine reduced anticipatory nausea symptoms thereby limiting the ability to through its effect on inhibitory presynexamine the correlation between MHPG aptic alpha-2 receptors, a lower dose of and anticipatory nausea. The small num- clonidine would be unlikely to control bers in the second study limit the ability anticipatory nausea better even if it did to exclude a role for noradrenergic activ- reduce toxicity. This is because the lower ity in anticipatory nausea. The second dose would produce even less inhibition limitation is the relatively mild anticipa- than the dose used in this study. Certory symptoms experienced by patients in tainly, the therapeutic index of clonidine both studies. In the first study the seven is distinctly inferior to the behavioral patients who experienced anticipatory techniques that are the treatment of nausea rated it from very mild to only choice for anticipatory nausea and vommoderate. In the second study anticipa- iting (2). tory symptoms in the placebo condition Although clonidine proved disappointaveraged 2.0 (i.e., mild). The mild symp- ing, there are other pharmacologic aptoms in both studies almost certainly re- proaches to anticipatory nausea. Perhaps flect the decreasing severity ot anticipa- the most obvious is to control postchemtory nausea and vomiting during the time otherapy nausea and vomiting. The adthe studies were conducted (1). This re- vent of 5HT3 antagonists has marked a duction in the severity of anticipatory significant advance in efforts to control symptoms has resulted from improved postchemotherapy symptoms (20). There control of postchemotherapy nausea and is also evidence from a small clinical trial vomiting. Regardless of cause, the limited that the benzodiazapine, alprazolam, rerange of anticipatory symptoms made it duces anticipatory nausea (21). In a more more difficult to examine the relationship speculative vein, recent reports demonbetween noradrenergic activity and antic- strate that the acquisition of conditioned ipatory nausea. If we had patients with fear responses can be blocked by drugs more severe symptoms in our sample, a that are receptor antagonists of N-methylmodest but significant association be- D-aspartate (22, 23). Perhaps N-methyl-Dtween noradrenergic activity and antici- aspartate plays a causative role in anticipatory nausea may have been detected. patory nausea and N-methyl-D-aspartate Finally, although the second study antagonists (e.g., dextromethorphan) may lacked the power to exclude an effect of reduce anticipatory nausea. clonidine on anticipatory nausea, it at least demonstrates that the therapeutic Supported by American Cancer Society index of clonidine for anticipatory nausea Grant PBR-19. is narrow. The dose of clonidine emWe thank Stephanie Young /or her outployed in the second study produced sig- standing administrative assistance, Grenificant side effects but minimal effects gory Dole for technical assistance, and on anticipatory nausea. A larger dose of John EnterJine for statistical consultation.
646
Psychosomatic Medicine 54:641-647 (1992)
NORADRENERGIC ACTIVITY IN ANTICIPATORY NAUSEA REFERENCES 1. Stefanek ME, Sheidler VR, Fetting JH: Anticipatory nausea and vomiting: Does it remain a significanl clinical problem? Cancer 62:2654, 1988 2. Burish TG, Carey MP: Conditioned responses to cancer chemotherapy: Etiology and treatment. In Fox BH, Newberry BH (eds) Impact of Psychoendocrine Systems in Cancer and Immunity. Toronto, CJ Hogrefe, 1983 3. Glavin GB Stress and brain noradrenaline: A review. Neurosci Behav Rev 9:233-245, 1985 4. Charney DS, Redmond DE: Neurobiological mechanisms in human anxiety. Neuropharmacology 22.1531-1536, 1983 5. Andrykowski MA, Redd WH: Longitudinal analysis of the development of anticipatory nausea in cancer chemotherapy. J Consult Clin Psychol 53:447-454, 1985 6. Fuxe K, Owman C: Cellular localization of monoamines in the area postrema of certain mammals. J Comp Neurol 125:337-354, 1965 7 Jenkins LC, Lahay D: Central mechanisms of vomiting related to catecholamine response: Anaesthetic implications. Can Anaesth Soc J 18:434-441, 1971 8. Sara SJ: Noradrenergic modulation of selective attention: Its role in memory retrieval. Ann NY Acad Sci 444:178-193, 1985 9. Maas JW: Relationships between central nervous system noradrenergic function and plasma and urinary concentrations of norepinephrine metabolites. In Usdin E, et al. (eds), Frontiers in Biochemical and Pharmacological Research in Depression. New York, Raven Press, 1984, 45-55 10. Jimerson DC. Balcnger JC, Lake CR, et al: Plasma and CSF MHPG in normals. Psychopharmacol Bull 17:86-87, 1981 11. Kopin IJ, Gordon EK, Jimerson DC, et al: Relation between plasma and cerebrospinal fluid levels of 3methoxy-4-hydroxyphenylglycol. Science 219:73-75, 1983 12. Aghajanian GK: Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature 276:186-188, 1978 13. Nerenz DR. Leventhal H, Easterling DV, et al: Anxiety and drug taste as predictors of anticipatory nausea in cancer chemotherapy. J Clin Oncol 4:224-233, 1986 14. Elsworth JD, Roth RH, Redmond DE Jr: The relative importance of 3-methoxy-4-hydroxyphenylethyleneglycol (MHPG) and 3,4 dihydroxyphenylglycol (DHPG) as norepinephrine metabolites in rat, monkey, and man. J Neurochem 41:786-793, 1983 15. Morrow GR. The assessment of nausea and vomiting. Cancer 53(Suppl):2267-2278.1984 16. Andrykowski MA: Definitional issues in the study of anticipatory nausea in cancer chemotherapy. J Behav Med 9:33-41, 1986 17. Fetting JH, Sheidler VR, Stefanek ME, et al: Clonidine for anticipatory nausea and vomiting: A pilot study examining dose-toxicity relationships and potential for further study. Cancer Treat Rep 71:409410, 1987 18. Pettinger WA: Pharmacology of clonidine. J Cardiovasc Pharmacol 2(Supp 1):S21-S28, 1980 19. Frisk-Holmberg M: Clinical pharmacology of clonidine. Chest 83(Suppl):395-397,1983 20. De Mulder PHM, Seynaeve C, Vermoken JB, et al: Ondansetron compared with high-dose metoclopramide in prophylaxis of acute and delayed cisplatm-mduced nausea and vomiting. Ann Int Med 113:834840, 1990 21. Greenberg DB, Surman OS, Clarke J, et al: Alprazolam for phobic nausea and vomiting related to cancer chemotherapy [letter]. Cancer Treat Rep 71:549-550, 1987 22. Staubli U, Thibault O, DiLorenzo M, et al: Antagonism of NMDA receptors impairs acquisition but not retention of olfactory memory. Behav Neurosci 103:54-60, 1989 23. Miserendino MJD, Sananes CB, Melia KR, et al: Blocking of acquisition but not expression of conditioned fear-potentiated startle by NMDA antagonists in the amygdala. Nature 345:716-718, 1990
Psychosomatic Medicine 54:641-647 (1992)
647
