sign in sign up
<<
Home , Fenfluramine, Phentermine

International Journal of (1999) 23, 723±732 ß 1999 Stockton Press All rights reserved 0307±0565/99 $12.00 http://www.stockton-press.co.uk/ijo Synergistic interactions between Fen¯uramine and

PJ Wellman1* and TJ Maher2

1Behavioral Neuroscience Program, Texas A&M University, College Station, Texas 77843-4235, USA and 2Division of Pharmaceutical Sciences, Massachusetts College of Pharmacy and Health Sciences, 179 Longwood Avenue, Boston, Massachusetts 02115, USA

`Fen-phen' refers to the off-label combination of the suppressants fen¯uramine and phentermine. The rationale for the fen-phen combination was that the two drugs exerted independent actions on brain satiety mechanisms so that it was possible to use lower doses of each drug and yet retain a common action on suppressing appetite while minimizing adverse drug effects. The focus of the present review is to consider whether fen¯uramine and phentermine exert actions that are additive in nature or whether these two drugs exhibit drug-drug synergism. The fen-phen combination results in synergism for the suppression of appetite and body weight, the reduction of brain levels, pulmonary vasoconstriction and valve disease. Fen-phen synergism may re¯ect changes in the of drug distribution, common actions on membrane ion currents, or interactions between neuronal release and reuptake mechanisms with MAO-mediated transmitter degradation. The synergism between fen¯uramine and phentermine highlights the need to more completely understand the pharmacology and neurochemistry of appetite suppressants prior to use in combination pharmacotherapy for the treatment of obesity.

Keywords: inhibitors; serotonin; drug-drug interactions; isobologram

Introduction lost during treatment with an appetite suppressant is quickly regained when the drug is terminated. Weintraub et al 6 reported on a novel approach to Obesity remains a chronic disorder of increasing the pharmacological treatment of obesity using com- prevalence1 that contributes to early death and binations of appetite suppressants rather than tradi- health disorders.2 Moreover, obesity is known to tional monotherapy. This employed a have considerable psychological costs in the form of combination of fen¯uramine (30 mg prior to dinner) ostracism and discrimination.3 These physical and with phentermine (15 mg in the morning) and com- psychological costs provide ample motivation for pared the weight reducing effects of this combination obese and near-obese persons to lose body weight. with that induced by fen¯uramine alone (20 mg ± 3 Among the obesity treatments developed and times per day) or phentermine alone (30 mg in the implemented over the last 50 y are behavioral mod- morning). Over a 24 week period, the combination of i®cation, exercise, dietary restriction, surgical low doses of these anorexic medications produced a approaches such as gastric stapling and=or banding, weight loss (8 kg) greater than that of placebo (4 kg), and pharmacotherapy.4 The latter approach involves but that was equivalent to the monotherapies (fen- the use of centrally acting drugs to suppress appetite ¯uramine alone or phentermine alone). Moreover, and reduce body weight5 and, more recently, agents fewer adverse effects were reported for the combina- that act peripherally to diminish absorption of dietary tion of fen-phen than for either drug alone. In 1992, fat from the gastrointestinal tract (for example, Orli- Weintraub et al 7 reported on a 4-y study that exam- stat). Many factors have contributed to resistance to ined the impact on body weight of a combination of pharmacotherapy including the beliefs that: obesity is phentermine with fen¯uramine. This study similarly not a disease but a psychological issue, that the noted that the fen-phen combination produced sub- effects of appetite suppressants are modest at best, stantial weight loss in some subjects and that this and that all appetite suppressants share common combination therapy could be used for as long as actions with the dangerous .5 4y.7 Moreover, a major concern remains that any weight The fen-phen approach to pharmacotherapy was widely adopted following the Weintraub et al 7 pub- lication. Sales of each appetite suppressant soared and it is estimated that some 18 million prescriptions were *Correspondence: Paul J Wellman, Behavioral Neuroscience written for the fen-phen combination between 1992 Program, Texas A&M University, College Station, Texas 77843- and 1997.8 However, fen¯uramine and the active 4235, USA. E-mail: [email protected] dexfen¯uramine (dexfen) were withdrawn Received 5 November 1998; revised 16 February 1999; accepted 3 March 1999 from the market in September 1997 following Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 724 mounting concerns that use of the fen¯uramines was will be less effective. Of greater interest is the associated with valvular disease9 and that appetite possibility that the two drugs will mutually amplify suppressant use has been associated with the devel- their respective clinical effects resulting in drug-drug opment of primary (PPH: synergism. A synergistic combination produces a 10 ± 15). greater effect than that associated with the simple The use of fen¯uramine or dexfen¯uramine in sum of effects of both drugs alone. It is interesting to combination with phentermine was `off-label' in the note that synergism between phentermine and fen- sense that although each of these drugs had been ¯uramine was an issue raised by Weintraub et al 6 in approved for use in the treatment of obesity, the their original fen-phen study: FDA had not explicitly approved this combination `Our data also suggest a synergistic effect of the of drugs. The FDA had subjected each of these drugs two medications used together. The combination individually to varying degrees of scrutiny for safety treatment group showed a marked, sustained and ef®cacy. There was a presumption on the part of increase in total fullness whereas mean results some that one could rationally predict the safety and from participants receiving the other active treat- ef®cacy pro®le of the combination therapy from ear- ments remained near baseline and similar to the lier research studies using each monotherapy alone. placebo response. It appears that the combination As a consequence, there appear to have been no of half doses of two treatments that by themselves preclinical studies conducted to determine the safety have no effect, even in full doses, results in of the combination of a fen¯uramine with phenter- increased effects ± in this case total fullness.' mine. Unfortunately at the time of development of the (6: pages 1147 ± 1148.) fen-phen pharmacotherapy, our understanding of the pharmacology of fen¯uramine and of phentermine Synergism may apply not only to the clinical out- was incomplete while our understanding of the fen- comes of the fen-phen combination, but also to the phen combination was essentially nonexistent. adverse reactions associated with the combination. Thus, the primary emphasis of the present review is to consider the known synergism between the fen¯ur- amines and phentermine with regard to eating and Rationale for the fen-phen combination body weight, cardiovascular activity (valve disease There exist two different rationales for the fen-phen and PPH), and central nervous system activity. combination therapy. The ®rst is that a combination of low doses of two drugs that share a common clinical effect (for example, suppression of appetite) will summate to produce a desired clinical effect. Neuropharmacology of Fen¯uramine and Phentermine Because adverse side effects are usually a function of The FDA approved Phentermine in 1959 while fen- dose, combining low doses of different drugs should ¯uramine was approved in 1973. The assumption of ideally maintain the desired clinical effect (that is, most investigators was that fen¯uramine acted to those effects will add together) while reducing the reduce feeding by indirectly enhancing serotonin incidence of unwanted side effects. A secondary activity in brain.17 ± 19 Fen and=or its desethylated notion behind combining these two drugs is that metabolite (norfen) blocks the reuptake of serotonin fen¯uramine acts as a depressant while phentermine into axon terminals, and at higher doses releases 5-HT acts as a stimulant.16 Fen¯uramine is a racemic from presynaptic terminals.20,21 Thus, the net effect of mixture of two , levofen¯uramine (levo- these actions is to increase 5-HT activity within the fen) and dexfen. Whereas, levofen has been shown to synapse, leading to suppression of appetite, and even- exert antidopaminergic (for example, sedative) activ- tually to weight loss. Additionally, Curzon and co- ity, phentermine has properties. workers22 have also suggested that fen¯uramine anor- Mixing phentermine with racemic fen¯uramine exia may partly re¯ect direct actions of fen at brain 5- would theoretically cancel out these dopaminergic HT receptors (for example 5 HT2c receptors). effects, so as to minimize patient complaints and Unlike fen¯uramine, phentermine has been viewed improve compliance. as a dopaminergic similar in action and Fen-phen combination therapy, however, rests on mechanism to amphetamine.23,24 An emerging theme the assumption that the two drugs to be combined of the recent research literature is the realization that produce their clinical effects (appetite suppression) phentermine is not only a dopaminergic agonist, but and adverse effects via independent mechanisms. The that this drug, at doses that reduce appetite, exerts essence of combination therapy is that the two drugs important indirect effects on neurotrans- show at least additivity: that is, their combined mission. A secondary emphasis of the present review effects should be predictable knowing the action of is that the conjoint effects of phentermine and of each drug alone. The notion of additivity is crucial fen¯uramine on serotonin may for both the clinical and adverse effects of the help to explain the clinical ef®cacy of the fen-phen combination drugs. If the two drugs are less than combination, as well as those synergistic actions of additive in their clinical effect, then the combination fen-phen that result in toxicity. Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 725 Synergistic Interactions between in nature, the experimenter examines how closely various combinations of the drugs of interest produce Phentermine and Fen¯uramine the expected reductions in food intake. The results from the Roth and Rowland26 study Food intake and body weight con®rm a synergism between dexfen and phentermine Only recently has the interaction between phenter- on food intake in free-feeding rats (see Figure 2). The mine and a fen¯uramine been formally examined in dose-response curves for dexfen and for phentermine animal studies of eating. Roth and Rowland25 exam- are represented along the x and y axes, respectively. ined the impact of daily IP administration of either The solid line represents the theoretical dose additive dexfen (2 mg=kg), phentermine (5 mg=kg) or a mix- line: that is, combinations of doses tested along this ture of dexfen=phentermine (2 mg=kg ‡ 5mg=kg) on line should yield a 50% suppression of eating in rats. food intake and body weight in rats (see Figure 1). The two open circles in the ®gure represent the ED50 Daily administration of 5 mg=kg=d phentermine alone value of one drug in combination with various doses had minimal effects on food intake in rats, whereas of the other drug. In one instance, the investigators 2mg=kg=d dexfen produced a signi®cant reduction in combined 0.5 mg=kg dexfen with various phentermine food intake. More importantly, when 2 mg=kg dexfen doses. The dose additive curve predicts that about was combined with 5 mg=kg phentermine, the combi- 5mg=kg of phentermine would be required to elicit a nation yielded a larger suppression of food intake than 50% reduction in eating in combination with that produced by 2 mg=kg dexfen alone. A similar 0.5 mg=kg dexfen. The actual dose required was synergy between dexfen and phentermine was evident approximately 2.0 mg=kg phentermine. Similarly, with respect to the reduction of body weight in this when 2.5 mg=kg of phentermine was combined with study. various doses of dexfen, less dexfen was required to The ®ndings of Roth and Rowland25 document a elicit a 50% reduction of eating than that predicted by synergism between dexfen and phentermine on eating the dose-additive line. The line (dotted) that bounds and body weight in the rat. In order to further assess either side of the dose-additive line represents the fen-phen synergism over a wide range of drug doses, 95% con®dence interval. In each instance, the combi- Roth and Rowland conducted a follow-up study using nations of dexfen and phen required smaller doses to the isobologram technique to index drug-drug inter- suppress food intake by 50% than that predicted by actions that occur between phentermine and dexfen- the dose-additive line (note that each point lies outside ¯uramine.26 The isobolographic technique considers the 95% CI). This study con®rms and extends their the potency of each of two drugs on food intake.27 A earlier observations and supports the notion that the dose-response curve for appetite suppression is gen- dexfen-phen combination is synergistic with regard to erated for each drug alone and these curves are in turn the suppression of food intake in rats. It will be used to predict the dose-additive function for the two important to further extend these studies using the drugs (that is those doses that in combination would fen-phen combination, as this was the combination be expected to produce a 50% reduction in food most commonly used in people. intake). To examine whether the drugs are additive

Figure 1 Effects of repeated administration of vehicle (open Figure 2 Isobologram of the effects of either 0.5 mg=kg dexfen circle), DEXFEN (open triangle: 2 mg=kg), PHEN (®lled square: in combination with various doses of phentermine or 2.5 mg=kg 5mg=kg) and the combination (®lled diamond) of DEXFEN phentermine in combination with various doses of dex-fen. In (2 mg=kg) and PHEN (5 mg=kg) on mean intake of sweetened each instance, the drug-drug combinations induced greater milk on 11 consecutive days. (From Roth JD and Rowland NE. suppressions of eating than that predicted on the basis of Ef®cacy of administration of dexfen¯uramine and phentermine, dose-additivity. The dotted lines represent the 95% con®dence alone and in combination, on ingestive behaviour and body intervals. (Figure reprinted with permission from Roth and Row- weight in rats. #1998, Springer-Verlag). land,26 courtesy of Elsevier Science, Amsterdam). Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 726 An advantage presumably conferred on the fen- preexposure to phentermine signi®cantly enhanced phen combination is a reduced pro®le of adverse drug the vasoconstriction in response to dexfen. These effects because the combination allowed for lower studies indicate that phentermine increases the vaso- doses of each drug to be used to reduce appetite and constrictive activity of fen¯uramine, an effect that has body weight. This section considers whether the been argued as contributing to the induction of PPH.33 adverse effects of the fen-phen combination exhibit As described below, the ability of phentermine to synergy as was evident for appetite suppression. attenuate the metabolism of serotonin could be the underlying mechanism to explain this ®nding. An issue to be resolved in future epidemiological Primary pulmonary hypertension studies is whether the combination of fen¯uramine Primary pulmonary hypertension (PPH) is a rare, and phentermine leads to a risk for PPH that re¯ects progressive, and generally incurable disorder invol- the additive PPH risk for each drug or whether the risk ving increased pulmonary artery pressure and for PPH with the combination re¯ects the type of increased vascular resistance eventually leading to synergism evident for appetite suppression. heart failure. The incidence rate of PPH is 1 ± 2 cases per million in the general population. Exposure to speci®c appetite suppressants can increase the incidence of PPH. In the 1960's, PPH was linked to Connolly et al 9 reported varying degress of valvular the use of the appetite suppressant .13,28 regurgitation in a series of 24 fen-phen users. Sub- More recently, PPH has been reported to be linked sequent histopathological analyses revealed a thicken- to the chronic use of dexfen and fen¯ura- ing of the affected valves resulting in reduced valve mine10,11,14,15,29 Moreover, exposure to phentermine mobility. Subsequent studies have sought to establish alone has resulted in reported cases of PPH.30 ± 32 the incidence of valve disease as a function of Serotonin has long been thought to play a role in the anorexigen exposure. Although phentermine alone development of PPH.33,34 In the 1970's, a group of rarely produces valve disease, an evolving ®nding German investigators explored the concept that has been that the combination of phentermine with a enhancement of systemic serotonin might induce fen¯uramine produces a greater incidence of valve vasoconstriction and in turn contribute to the devel- disease than that associated with a fen¯uramine alone. opment of PPH. Subsequently, Herve et al 35 reported Since the original report by Connolly et al 9 there that PPH patients exhibit a signi®cant increase in have been a number of studies published and pre- plasma serotonin and exhibit a greater release of sented at meetings that report on the incidence of serotonin from aggregated platelets. Moreover, pul- valvulopathies following treatment-with fen, dexfen monary arterial tissue taken from PPH patients exhi- and=or phen. These studies report a wide range of bits greater sensitivity to serotonin relative to tissue valvulopathy incidence for anorexigen exposure ran- from control patients.36 One proposed animal model ging from greater than 35% incidence down to 2 ± 4% of PPH indicates a link between serotonin and pul- (a value equal to control). These studies utilize a wide monary hypertension. For example, rats injected with range of protocols. Not all the studies report on the the toxic alkaloid monocrotaline (MCT) develop ele- incidence in control groups or when utilized, some vated pulmonary pressures37 within 21 days after control groups are not appropriately age=weight=sex MCT treatment. Kanai et al 38 noted that plasma matched, the durations of exposure vary widely, and serotonin levels are increased in rats 12 h after MCT the criteria used to de®ne valvulopathy severity are treatment and that chronic pretreatment of MCT rats often inconsistent from study to study. Only a few of with the serotonin antagonist DV-7028 suppressed these studies have explicitly compared the incidence medial pulmonary artery hypertrophy, right ventricu- of valve disease for a fen¯uramine alone versus a lar hypertrophy, and pulmonary artery pressure. group exposed to a fen¯uramine used in combination As noted above, phentermine use is known to with phentermine. produce PPH in some patients and this drug is In the ®rst study by Khan et al,40 the odds ratio for known to interact with and to modulate serotonin cardiac valve abnormalities that met their case de®ni- activity within the pulmonary system. Seiler et al 34 tion by echocardiography in patients taking dexfen reported that phentermine prolongs the vasoconstric- alone was 12.7, whereas for those patients taking tive action of serotonin in the isolated rat lung. More dexfen plus phentermine, or fen plus phentermine, recently, Reeve et al 39 examined pulmonary artery the odds ratio increased to 24.5 and 26.3, respectively. (PA) pressure in the isolated rat lung by administering While these authors acknowledge that the variation in a series of vasoconstrictive stimuli (angiotensin II or duration of treatment between the groups makes hypoxia), as well as a series of infusions of vehicle interpretation dif®cult, the addition of phentermine followed by dexfen, or phentermine followed by appears to signi®cantly increase the likelihood of dexfen. This report notes that phentermine pretreat- valvulopathy. In the second study by Wee et al 41 ment (10 and 100 uM) enhanced the vasoconstrictive patients who for various reasons had undergone echo- activity of dexfen (10 and 100 uM). Although phen- cardiographic analysis prior to being prescribed fen- termine alone had no impact on PA pressure, phen or dexfen were reevaluated following cessation Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 727 of therapy. Of the 46 patients included in the study, thought to exert direct actions at serotonin receptors. only two had worsening or newly developed valvulo- Whereas fen-phen-associated valve disease appears to pathies that met FDA criteria (mild aortic or moderate involve primarily the left-side of the heart,9 carcinoid mitral regurgitation). Both of these patients had taken disease involves right side heart valves46 and ergota- fen-phen. None of the patients who had taken dexfen mine alters valves on either side of the heart.47 Fish- developed valvulopathies. However, despite the desir- man33 argues that the difference in pattern of valve ability of this study design, so few patients were disease re¯ects the role of the lung in the inactivation included so as to prevent a de®nitive conclusion of serotonin. In carcinoid syndrome, the lung retains regarding the in¯uence of phentermine addition. the ability to remove circulating serotonin, thus the A study by Gross et al 42 reported that aortic valve damage is primarily on the right side whereas insuf®ciency (AI) in fen-phen patients was signi®- during fen-phen use, the lung is impaired in the ability cantly altered by duration of exposure (greater than 9 to clear serotonin, in part because phentermine inhibits months) and by dose of the combination (greater than MAO and this impairs the metabolism of serotonin 60 mg). Given that phentermine alone probably does within the endothelium. not induce valve disease, the ®nding that dose combi- While none of the above studies possesses the ideal nations greater than 60 mg=d increase the likelihood experimental design, or is large enough to yield a of valve disease indicates that phentermine is not inert de®nitive conclusion, each appears to support a role of when given in combination with fen¯uramine. This phentermine in increasing the likelihood of valvulo- latter ®nding supports the notion that mixing phenter- pathy in patients taking one of the fen¯uramines. mine with fen¯uramine increases the impact of fen- Unfortunately, the appropriate study that would con- ¯uramine on valve disease. trol for all the experimental variables and have the A study by Gardin et al 43 similarly noted an appropriate statistical power cannot be undertaken increase in the incidence of mild or greater aortic since the fen¯uramines are no longer available for regurgitation associated with fen-phen. While the use in humans. control group incidence was 4.1%, the dexfen and fen-phen group incidences were 8.9% and 13.7%, respectively. In a study by Shively et al 44 the incidence of Brain and peripheral serotonin valvulopathy meeting FDA criteria was 3% in non- Serotonin (5-HT) is an indolamine transmitter found treated controls and 7% in patients treated with in the brain as well as the periphery. The raphe nuclei dexfen. However, when the potentially confounding are among the primary sources of 5-HT in factor of concomitant use of drugs reported to inhibit brain and these neurons project diffusely to frontal monoamine oxidase (for example, estrogens, thyroid) cortex, striatum, hippocampus, , and was excluded, the incidence of valve disease brain stem. Acute exposure to very large doses of decreased to 4%, a value not signi®cantly different fen¯uramine reduces brain levels of serotonin.48 from control. The signi®cance of this ®nding relates to Although most of bodily 5-HT is found in the periph- the observation that phentermine is an inhibitor of ery within gut and blood platelets, few studies have monoamine oxidase, the enzyme that inactivates sero- examined the impact of fen¯uramine alone or in tonin.45 This ®nding lends further support to the combination with phentermine on peripheral 5-HT notion that valve disease associated with fen-phen levels and 5-HT activity. The importance of this re¯ects enhanced systemic serotonin activity. remarkable gap is evident given recent suggestions Valve disease similar to that noted by Connolly et that fen-phen alterations of systemic 5-HT may repre- al 9 has been noted in carcinoid disease as well as in sent a primary mechanism by which to explain the toxicity.46,47 Carcinoid syndrome involves impact of fen-phen therapy on the incidence of valve hypersecretion of serotonin whereas ergotamine is disease and of PPH.33,34

Table 1

PHEN dose FEN Dose Impact of FEN Evidence of Authors Species mg=kg (CUMUL) mg=kg (CUMUL) Brain regions on regional 5-HT synergism?

Baumann et al 49 Mouse 7 (56) 3 (24), 10 (80), 30 (240) Forebrain ? None Halladay et al50 Rat 12 (96) 16 (128) Striatum 7 68% None Lew et al 51* Rat 5 (20), 20 (80) 3.125 (12.5), 12.5 (50) Nuc accumbens 7 16% PHEN5: YES PHEN 20: NO Striatum 7 27% PHEN5: YES PHEN 20: NO Amygdala 7 38% PHEN5: YES PHEN 20: NO Hypothalamus 7 43% PHEN5: No PHEN 20: NO Fr. Parietal cortex 7 50% FHEN5: YES PHEN20: NO Hippocampus 7 53% PHEN5: NO PHEN 20: NO McCann et al 52 Mouse 20 (160), 40 (320), 10 (80) Striatum 7 15% PHEN 20: NO PHEN 40: YES Hypothalamus 7 17% PHEN20: YES PHEN40: YES Hippocampus 7 27% PHEN 20: NO PHEN 40: NO Cortex 7 37% PHEN 20: NO PHEN 40: NO

*Rats treated with fen-phen once per hour for 4 hours (all other treatments are 2=day for 4 days) Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 728 Several recent research efforts have examined the coadministration of phentermine in hypothalamus potential interactions between fen¯uramine and phen- and striatum but not in hippocampus or cortex. termine on brain 5-HT activity. These four extant These studies indicate that phentermine, which by studies are summarized in Table 1. The table presents itself did not alter brain serotonin levels, enhances the authors, species used, dose per injection (typically the impact of fen¯uramine on reducing brain seroto- twice per day for 4 d), the cumulative dose of nin levels. phentermine or fen¯uramine, the brain regions exam- Table 1 also presents the impact of fen¯uramine ined in the study, the impact of fen alone on 5-HT alone on changes in regional serotonin levels. One brain concentration, and whether the study observed factor that may partly explain whether fen-phen syner- an outcome consistent with synergism. gism is observed is the basal level of depletion pro- The ®rst study was that of Baumann et al 49 who duced by a test regimen of fen¯uramine. In the Lew et examined the changes in mouse forebrain 5-HT asso- al 51 study, synergism was most likely observed when ciated with twice a day exposure for 4 days to either 3, fen¯uramine alone produced a small decrease in 10 or 30 mg=kg fen¯uramine alone or these fen¯ur- regional serotonin. A similar pattern was evident in amine doses in combination with a single dose of the McCann et al 52 study in which phentermine was 7mg=kg phentermine. The second study of Table 2 most likely to yield synergism when the impact of was carried out by Halladay et al 50 who examined the fen¯uramine alone on serotonin within a brain region impact of 16 mg=kg fen¯uramine (twice per day for 4 was small. These results suggest that the failure to days) on striatal 5-HT levels as well as the impact of observe synergism in the Halladay et al 50 study may 12 mg=kg phentermine given in combination with the re¯ect a ceiling effect on depletion of striatal seroto- 16 mg=kg fen¯uramine regimen. Neither the Baumann nin, thereby masking any possible synergistic effect of study nor the Halladay study noted evidence of phentermine in combination with fen¯uramine. synergism with regard to reductions of brain serotonin The available studies on the interaction of phenter- content. These two studies use different species and a mine with fen¯uramine on central serotonin levels common exposure regimen. The studies, however are utilize widely varying dose ranges and exposure regi- limited by the use of only a single dose of phenter- mens. A key aspect is that such studies tend to use mine, and only examined changes in serotonin in but a high-dose repeated administration regimens. Further single brain region. Moreover, because these severe studies using a wider range of pharmacologically fen¯uramine exposure regimens produced signi®cant relevant dosages and longer exposure durations will acute reductions in brain indole content, demonstra- further indicate the extent to which phentermine and tion of synergistic effects with addition of phenter- fen¯uramine exhibit synergism with regard to altera- mine is made dif®cult. tion of brain serotonin systems. Similar caveats apply The last two entries of Table 1 provide some to future studies of the impact of fen-phen on systemic evidence of a synergistic interaction between phenter- 5-HT activity. mine and fen¯uramine on brain serotonin levels. In the Lew et al 51 study, rats were injected every hour for 4 hours with various combinations of fen-phen and then sacri®ced either 7 d or 28 d later. The rats were treated Potential mechanisms underlying each time with either 3.125 or 12.5 mg=kg fen¯ura- fen-phen synergism mine, with either 5 or 20 mg=kg phentermine, or a combination of 5 mg=kg phentermine with 3.125 mg=kg fen¯uramine or 20 mg=kg phentermine with Whereas the prior sections present the current state of 3.125 mg=kg fen¯uramine. These authors examined the literature on fen-phen synergism, this section turns changes in 5-HT levels in each of six brain regions. to a discussion of the potential explanations for such Four hourly injections of either 3.125 or 12.5 mg=kg synergism. fen induced dose-dependent reductions in brain serotonin in each of the six brain regions. While the 5.0 mg=kg dose of phentermine alone did not alter Pharmacokinetics serotonin content in any of the six brain regions, this A pharmacokinetic explanation would suggest that phentermine dose (a no effect dose for alteration of phentermine exposure alters the absorption, distribu- brain serotonin) enhanced the impact of 3.125 mg=kg tion, metabolism and=or elimination of fen¯uramine. fen dosing regimen on reducing serotonin in four of the The impact of phentermine on fen¯uramine pharma- six brain regions. cokinetics is a key issue. There are data available as to In the last entry of Table 1, McCann et al 52 treated whether exposure to anorexic agents similar in struc- mice twice per day for 4 days with either 10 mg=kg ture to phentermine alters fen¯uramine bioactivity. fen, 20 or 40 mg=kg phentermine or a combination Hunsinger et al 53 reported that rats given a series of of 10 mg=kg fen with either 20 or 40 mg=kg phenter- daily injections of 4.0 mg=kg=day amphetamine mine. Levels of brain serotonin were not altered exhibited a greater sensitivity to the toxic actions of by phentermine alone. Fen¯uramine reduced brain fen relative to rats given daily injections of vehicle. serotonin levels and this effect was enhanced by Speci®cally, amphetamine preexposure lowered the Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 729 LD50 value for fen from 97 mg=kg to 68 mg=kg (53). of 5-HT in rabbit lung. The observation that phenter- Moreover, amphetamine preexposure increased the mine inhibits MAO was recently con®rmed and brain levels of fen relative to that evident in rats extended by Maher et al 45 who noted that phenter- given vehicle pretreatment. The Hunsinger et al 53 mine is as potent in its inhibition of MAO as is study indicates that exposure to amphetamine alters iproniazid, an MAOI drug used to treat the pharmacokinetics of fen so as to prolong the in some countries. The importance of these ®ndings is duration of action and impact of fen. Given the that MAO represents one of two major mechanisms similarity between amphetamine and phentermine in by which serotonin is inactivated. The other mechan- structure, activity and , it is ism, of course, is its reuptake into the nerve terminals reasonable to expect that phentermine exposure or, in the periphery, into platelets. If phentermine, an would also alter fen levels in brain. As described MAO inhibitor, were to be combined with fen¯ura- below, phentermine, like d-amphetamine, is known mine, a drug that releases serotonin and blocks its to competitively inhibit the enzyme that metabolizes reuptake, the expectation would be a further rise in fen and 5-HT. Additionally, fen, at these high doses, serotoninergic activity. Thus, the combination of may also release and , thus phentermine and a fen¯uramine would produce enhancing `aggregation toxicity' of amphetamine. increases in serotonin activity that would appear as Thus, the increased toxicity and CNS accumulation a synergism between the two drugs. of fen observed above might result, in part, from an As noted above, the incidence of valve disease to interference of metabolism of fen and=or 5-HT pro- the fen¯uramines is reduced to control values when duced by phentermine. subjects are excluded who were exposed to the co- administration of MAO inhibitors with a fen¯ura- mine.44 Further support for the notion that fen-phen Ion channels synergism re¯ects MAO inhibition by phentermine is Changes in ion channel activity and the resulting provided by a study in which pretreatment of rats with changes in membrane potential are key elements in the MAOI clorgyline potentiated the appetite suppres- cell function. Weir and colleagues54 examined the sant effects of fen¯uramine.60 Thus, inhibition of impact of anorexic agents on K‡ channel currents in MAO is suf®cient to interact with and to potentiate smooth muscle cells taken from rat pulmonary artery. the appetite suppressant actions of fen¯uramine. Pre- Both fen and dexfen inhibited K‡ channel current, sumably, such effects would also occur when phen- induced membrane depolarization and produced vaso- termine (an MAOI drug) were coadministered with a constriction in smooth muscle cells. More recently, fen¯uramine (cf.25,26 ). dexfen has been shown to increase Ca‡‡ levels within smooth muscle cells.55 These studies of peripheral cardiac muscle cells have not assessed whether K‡ Release of serotonin by phentermine channel inhibition or elevation of intracellular Ca‡‡ is The conventional wisdom has been that phentermine enhanced by co-administration of fen or dexfen with does not interact with central serotonin neurons in any phentermine and whether these effects occur at ther- signi®cant manner. Three recent studies using the apeutically relevant concentrations. Weir and collea- microdialysis technique suggest that while phenter- gues56 have argued that abnormal ion channel activity mine alone has minimal impact on release of 5-HT, appears to increase the susceptibility to the develop- phentermine may further enhance the extracellular 5- ment of PPH after exposure to fen¯uramine. HT increase induced by fen¯uramine. A study by Balcioglu and Wurtman61 examined the impact of phentermine on striatal dopamine and serotonin levels in conscious rats using the microdialysis technique. Serotonergic actions of Their ®ndings demonstrated that 2 and 5 mg=kg phentermine phentermine did not alter serotonin within the stria- tum but did signi®cantly increase striatal dopamine over baseline levels. A second study by Shoaib et al 62 Inhibition of MAO by phentermine noted that phentermine alone, at doses of 1.0 and In the 1970's, two independent research groups 2.0 mg=kg, slightly increased the extracellular levels reported that a variety of anorexic drugs had inhibi- of 5-HT in rat nucleus accumbens, but that the tory actions on monoamine oxidase (MAO), an intra- serotonergic action of phentermine was much less cellular enzyme that inactivates serotonin.57,58 Seiler marked than its impact on brain dopamine. More et al 34 suggested that the MAO-inhibitory (MAOI) importantly, the Shoaib et al 62 study noted that the action of phentermine might account for the ®nding mixture of phentermine with fen¯uramine had an that phentermine prolonged the action of serotonin additive effect on increasing brain serotonin levels, infusion on perfusion pressure in the isolated rat lung an observation consistent with the MAOI effect of preparation. MAO inhibition by phentermine would phentermine. A third microdialysis study by Balcio- also account for the ®ndings of Morita and Mehen- glu and Wurtman63 reported that whereas 2 mg=kg dale59 in which phentermine delayed the metabolism phentermine alone did not alter striatal serotonin Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 730 levels, a combination of 2 mg=kg phentermine with fen¯uramine on the reduction of appetite and body 1mg=kg fen¯uramine produced a 330% increase in weight in rats.25,26 Animal studies have also demon- striatal serotonin in rat in contrast to the 182% strated synergism between phentermine and fen¯ur- increase in striatal serotonin induced by 1 mg=kg amine on reduction of brain serotonin levels,51,52 and fen¯uramine only. These studies document that the between dexfen¯uramine and phentermine on vaso- acute effects of fen¯uramine on brain serotonin are constriction of pulmonary arterial vessels.39 Recent enhanced by the addition of phentermine. Given that studies indicate the coadministration of phentermine phentermine alone has minimal impact on brain increases the incidence of valve disease associated serotonin level, this enhancement would be viewed with fen¯uramine or dexfen¯uramine. Further animal as synergistic. studies are required to con®rm the magnitude of These studies are limited by the small range of synergism between phentermine and the fen¯ura- doses tested, by the use of mostly acute exposure to mines. Further epidemiological studies are needed to phentermine (acute exposure to fen produces assess the extent to which synergism is evident in the increases in extracellular serotonin levels while impact of fen¯uramine in the cardiovascular system. chronic high dose fen exposure can produce declines In each instance, investigators will need to consider in brain serotonin content), and by the sampling of the use of appropriate designs that allow for the extracellular serotonin in a limited number of brain determination of synergism between phentermine regions. Hypothalamic serotonin, for example, is and the fen¯uramines. likely more important for the understanding of the A secondary issue of the present review is that impact of fen-phen on eating than is serotonin within synergism between phentermine and fen¯uramine can the striatum. be viewed in the context of a drug-drug interaction. As was noted above, few studies have examined Fen¯uramine increases serotonin release and blocks the action of phentermine on serotonin in the periph- reuptake of serotonin whereas phentermine may ery. Fristrom et al 64 noted that phentermine released increase serotonin levels indirectly via inhibition of serotonin from rabbit blood platelets. Platelets are the MAO. These interactions would be expected to major storage pool for 5-HT in blood. A more recent enhance serotonin activity leading to a synergism of study in humans, indicates that phentermine may also clinical effect (for example, suppressed appetite and play a role in enhancing the level of serotonin in body weight) as well as adverse effects associated blood platelets. Maher et al 45 reported that the with excessive serotoninergic activity. The fen-phen administration of either 15 mg or 30 mg phentermine experience may increase our understanding of the to human volunteers increased platelet serotonin mechanisms of action of these appetite suppressants, levels by approximately 70% after one hour but did alone and in combination, and is likely to lead to the not alter plasma levels of serotonin, indicating that future development of appetite suppressant regimens this effect on platelet serotonin activity probably with known ef®cacy and safety. relates to the MAO inhibition effected by phenter- mine. Moreover, few studies have examined the impact of the fen-phen combination on platelet 5- Acknowledgements HT activity. Redmon et al 65 reported that chronic Dr Wellman and Dr Maher have acted, or have been fen-phen exposure did not alter plasma serotonin approached to act as consultants and=or expert wit- levels. This report, however, used only a single nesses for both sides involved in litigation over fen- assessment of plasma serotonin (24 h after last dose) phen in the USA. for subjects taking fen-phen, and this study did not assess platelet serotonin levels in fen-phen treated References subjects. 1 Kuczmarski RJ, Flegal KM, Campbell SM, Johnson CL. Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. JAMA 1994; 272: 205 ± 211. 2 Bray GA. Health hazards of obesity. Endocrinol Metabol Clin Conclusions NAm1996; 25: 907 ± 919. 3 Pingitore R, Dugoni BL, Tindale RS, Spring B. Bias against overweight job applicants in a simulated employment inter- The fen-phen combination enjoyed wide acceptance view. J Appl Psychol 1994; 79: 909 ± 917. because this combination was thought to reduce body 4 Blackburn GL, Miller D, Chan S. Pharmaceutical treatment of obesity. Nursing Clin of N Am 1997; 32: 831 ± 848. weight (and associated risks of obesity) in the absence 5 Bray GA. Use and of appetite-suppressant drugs of signi®cant risk. A remarkable aspect of fen-phen in the treatment of obesity. Ann Int Med 1993; 119=7:Pt 2: use in the treatment of human obesity is the fact that 1 ± 13. this use occurred `off-label' without animal studies to 6 Weintraub M, Hasday JD, Mushlin AI, Lockwood DH. A either understand the mechanism(s) of action of the double-blind clinical trial in weight control. Use of fen¯ur- amine and phentermine alone and in combination. Arch Inter- fen-phen combination or to substantiate the safety of nal Med 1984; 144: 1143 ± 1148. the fen-phen combination. Animal studies clearly 7 Weintraub M. Long-term weight control study: conclusions. indicate a synergism between phentermine and dex- Clin Pharmacol Therap 1994; 51: 642 ± 646. Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 731 8 Kassirer JP, Angell M. Losing weight ± An ill-fated new 30 Backmann R, Dengler H, K Gahl, Greiser R, Jesdinsky HJ, years resolution. New Eng J Med 1998; 338: 52 ± 54. Loogen F. Primary Pulmonary Hypertension. Darmstadt: Dr. 9 Connolly HM, Crary JL, McGoon MD, Hensrud DD, Edwards Dietrich Steinkopff Publishers, 1972. BS, Edwards WD, Schaff HV. Valvular heart disease asso- 31 Heuer L, Benoit W, Hoydrich H, Kummer D, Schick J. ciated with fen¯uramine-phentermine. New Eng J Med 1977; Pulmonale hypertonie durch appetitzughr (Mirapront). Chir 337: 581 ± 588. Praxix 1978; 23: 497 ± 504 (German). 10 Abenhaim L, Moride Y, Brenot F, Rich S, Benichou J, Kurz 32 Schnabel KH, Shatz V, Busch S, Just H. Drug-induced X, Higenbottam T, Oakley C, Wouters E, Aubier M, Simon- primary vascular pulmonary hypertension. Med Welt 1976; neau G, Begaud B. Appetite-suppressant drugs and the risk of 27: 1300 ± 1303. primary pulmonary hypertension. International Primary Pul- 33 Fishman AP. Aminorex to Fen=Phen: An epidemic foretold. monary Hypertension Study Group. New Eng J Med 1996; Circulation 1999; 99: 156 ± 161. 335=9: 609 ± 616. 34 Seiler KU, Wassermann O, Wensky H. On the role of 11 Brenot F, Herve P, Petipretz P, Paren F, Duroux P, Simonneau serotonin in the pathogenesis of pulmonary hypertension G. Primary pulmonary hypertension and fen¯uramine use. Brit induced by drugs; an experimental study in the Hrt J 1993; 70: 537 ± 541. isolated perfused rat lung, II. Fen¯uramine, , mefe- 12 Delcroix M, Kurz X, Walckiers D, Demedts M, Naeije R. norex, phentermine and R 800. Clinical & Experimental High incidence of primary pulmonary hypertension associated Pharmacol Physiol 1976; 3: 323 ± 330. with appetite suppressants in Belgium. Eur Respir J 1998; 12: 35 Herve P, Launay J.M, Scrobohaci ML, Brenot F, Simoneau G, 271 ± 276. Pettipretz P, Poubeau P, Cerrina J, Duroux P, Drouet L. 13 Follath F, Burkart F, Schweize, W. Drug-induced pulmonary Increased plasma serotonin in primary pulmonary hyperten- hypertension? BMJ 1971; 1=5743: 265 ± 266. sion. Am J Med 1995; 99: 249 ± 254. 14 Mark EJ, Patalas ED, Chang HT, Evans RJ, Kessler SC. Fatal 36 Brink C, Cerrina C, Labat C et al. The effects of contractile pulmonary hypertension associated with short-term use of on isolated pulmonary arterial and venous muscle fen¯uramine and phentermine. New Eng J Med 1997; 337: preparations derived from patients with primary pulmonary 602 ± 606. hypertension Am Rev Respir Dis 1988; 137=part 2: A106. 15 Simonneau G, Fartoukh M, Sitbon O, Humbert M, Jagot JL, 37 Hosoda Y. Pathology of pulmonary hypertension: A human Herve P. Primary pulmonary hypertension associated with the and experimental study. Pathol Intl 1994; 44: 241 ± 267. use of fen¯uramine derivatives. Chest 1998; 114=3: Suppl, 38 Kanai Y, Hori S, Tanaka T, Yasuok M, Watanabe K, Aikaa N, S195 ± S199. Hosoda Y. Role of 5-hydroxytryptamine in the progression of 16 Rea WP, Rothman RB, Shippenberg TS. Evaluation of the monocrotaline induced pulmonary hypertension in rats. Car- conditioned reinforcing effects of phentermine and fen¯ura- diovasc Res 1993; 27: 1619 ± 1623. mine in the rat - concordance with clinical studies. Synapse 39 Reeve HL, Nelson DP, Archer SL, Weir KE. The effects of 1998; 30: 107 ± 111. ¯uoxetine, phentermine, and on pulmonary artery 17 Blundell JE. Is there a role for serotonin (5-hydroxytrypta- pressure and electrophysiology. Am J Physiol 1999; 213 ± 219. mine) in feeding? In J Obes 1977; 1: 15 ± 42. 40 Khan MA, Herzog CA, Stpeter JV, Hartley GG, Madlonkay R, 18 Garattini S. Biological actions of drugs affecting serotonin and Dick CD, Asinger RW, Vessey JT. The prevalence of cardiac eating. Obesity Res 1995; 3: (Suppl): S470. valvular insuf®ciency assessed by transthoracic echocardio- 19 Garattini S, Bizzi A, Caccia S, Mennini T, Samanin R. graphy in obese patients treated with appetite-suppressant Progress in assessing the role of serotonin in the control of drugs. New Eng J Med 1998; 339: 713 ± 718. food intake. Clin Neuropharmacol 1988; 11: Suppl, S32. 41 Wee CC, Phillips RS, Aurigemma G, Erban S, Kriegel G, 20 Heal DJ, Cheetam SC, Prow MR, Martin KF, Buckett WR. A Riley M, Douglas PS. Risk for valvular heart disease among comparison of the effects on central 5-HT function of sibu- users of fen¯uramine and dexfen¯uramine who underwent tramine hydrochloride and other weight modifying agents. Br. echocardiography before use of medication. Ann Internal J. Pharmacol 1998; 125: 301 ± 308. Med 1998; 129: 870 ± 874. 21 Gundlach C, Martin KF, Heal DJ, Auerbach SB. In vivo 42 Gross SD, Leper NB, Daley WL, Buchbinder NA, Hicky A, criteria to differentiate monoamine reuptake inhibitors from Samuels BA. Dose and duration of appetite suppressant releasing agents: is a reuptake inhibitor. J Phar- activity impacts incidence of valvular heart disease. Circula- macol Exp Therap 1997; 283: 581 ± 591. tion 1998; 138. 22 Curzon G, Gibson EL, Oluyomi AO. Appetite suppression by 43 Gardin JM. An echocardiographic evaluation of patients commonly used drugs depends on 5-HT receptors but not on 5- exposed to anorexigens as compared to unmatched controls. HT availability. Trends Pharmacol Sci 1977; 18: 21 ± 25. Presentation at the 1988 meeting of the AHA, Dallas, TX. 23 Bray GA. Evaluation of drugs for treating obesity. Obesity Res 44 Shively BK. Echocardiographic diagnosis. Presentation at the 1995; 3: Suppl, S434. 1998 AHA meeting Dallas, Texas. 24 Yelnosky J, Panasevich RE, Borrelli AR, Lawlor RB. Phar- 45 Maher TJ, Ulus IH, Wurtman RJ. Phentermine and other macology of phentermine. Archives Internationales de Phar- monoamine oxidase inhibitors may increase plasma serotonin macodynamie et de Therape 1969; 178: 62 ± 76. when given with fen¯uramines. Lancet 1999; 353: 38. 25 Roth JD, Rowland NE. Ef®cacy of administration of dexfen- 46 Lundin, L. Carcinoid heart disease: A cardiologist's view- ¯uramine and phentermine, alone and in combination, on point. Acta Oncologica 1991; 30: 499 ± 502. ingestive behavior and body weight in rats. Psychopharmacol 47 Red®eld MM, Nicholson WJ, Edwards WD, Tajk, AJ Valve 1998; 137: 99 ± 106. disease associated with ergot alkaloid use: Echocardiographic 26 Roth JD, Rowland NE. Anorectic ef®cacy of the fen¯urami- and pathologic correlations. AnnInternMed,1992; 117: 50 ± 52. ne=phentermine combination in rats: additivity or synergy? 48 McCann U, Hatzidimitriou G, Ridenour A, Fischer C, Eur J Pharmacol 1999; in press. Yuan J, Katz J, Ricaurte G. Dexfen¯uramine and serotonin 27 Wellman PJ, Tow S, McMahon L. Isobolographic assessment : Further preclinical evidence that clinical cau- of the effects of combinations of and tion is indicated. J Pharmacol Exptl Therapeut 1994; 269: fen¯uramine on food intake in rats. Pharmacol Biochem 792 ± 798. Behav 1995; 50: 287 ± 291. 49 Baumann MH, Schuster CR, Rothman RB. Effects of phen- 28 Gurtner HP, Aminorex and pulmonary hypertension. A termine and on fen¯uramine-induced depletion of review. Cor et Vasa 1985; 27: 160 ± 171. serotonin in mouse brain. Drug Alc Dep 1996; 41: 71 ± 74. 29 Douglas JG, Munro JF, Kitchin AH, Muir AL, Proudfoot AT. 50 Halladay AK, Fisher H, Wagner GC. Interaction of phenter- Pulmonary hypertension and fen¯uramine. Br Med J 1981; mine plus fen¯uramine: neurochemical and neurotoxic effects. 6296: 881 ± 883. Neurotoxicol 1998; 19: 177 ± 183. Fen¯uramine ± Phentermine synergism PJ Wellman and TJ Maher 732 51 Lew R, Weisenberg B, Vosmer G, Seiden LS. Combined 59 Morita T, Mehendale HM. Effects of and phentermine=fen¯uramine administration enhances depletion phentermine on the pulmonary disposition of 5-hydroxy- of serotonin from central terminal ®elds. Synapse 1997; 26: in the rat in vivo. Am Rev Respir Dis 1983; 127: 36 ± 45. 747 ± 750. 52 McCann U, Jie Y, Ricaurte GA. Neurotoxic effects of 60 Aulakh CS, Hill JL, Wozniak KM, Murphy DL. Fen¯uramine- (‡=7)fen¯uramine and phentermine, alone and in combina- induced suppression of food intake and locomotor activity is tion, on monoamine neurons in the mouse brain. Synapse differentially altered by the selective type A monoamine 1998; 30: 239 ± 246. oxidase inhibitor clorgyline. Psychopharmacol 1988; 95: 53 Hunsinger RN, Kibbe AH, Wilson MC. The effect of previous 313 ± 317. d-amphetamine treatment on the disposition and lethality of 61 Balcioglu A, Wurtman RJ. Effects of phentermine on striatal fen¯uramine in the rat. Toxicol Appl Pharmacol 1985; 79: dopamine and serotonin release in conscious rats: in vivo 236 ± 245. microdialysis study. Intl J Obesity Rel Metabol Dis 1998; 54 Weir EK, Reeve HL, Huang JM, Michelakis E, Nelson DP, 22: 325 ± 328. Hampl V, Archer SL. Anorexic agents aminorex, fen¯ura- 62 Shoaib M, Baumann MH, Rothman RB, Goldberg SR, Schind- mine, and dexfen¯uramine inhibit potassium current in rat ler CW. Behavioral and neurochemical characteristics of pulmonary vascular smooth muscle and cause pulmonary phentermine and fen¯uramine admininistered separately and vasoconstriction. Circulation 1996; 94: 2216 ± 2220. as a mixture in rats. Psychopharmacol 1996; 131: 296 ± 306. 55 Reeve H, Archer SL, Weir EK. (1998) Dexfen¯uramine 63 Balcioglu A, Wurtman RJ. Effects of fen¯uramine and phen- releases intracellular Ca2 ‡ in rat pulmonary artery smooth termine (fen-phen) on dopamine and serotonin release in rat muscle cells. Circulation, 1790. striatum: in vivo microdialysis study in conscious animals. 56 Weir EK, Reeve HL, Johnson G, Michelakis ED, Nelson DP, Brain Res 1998; 813: 67 ± 72. Archer SL. A role for potassium channels in smooth muscle 64 Fristrom S, Airaksinen MM, Halmekoski J. Release of platelet cells and platelets in the etiology of primary pulmonary 5-hydroxytryptamine by some anorexic and other sympatho- hypertension. Chest 1998; 114=3: Suppl, S200 ± S204. mimetics and their acetyl derivatives. Acta Pharmacol Tox- 57 Moller-Nielsen I, Dubnick B. Pharmacology of chlorphenter- icologicol 1977; 41: 218 ± 224. mine. In: Costa E. and Garattini (eds) and 65 Redmon B, Raatz S, Bantle JP. Valvular heart disease asso- Related Compounds. Raven Press: New York, 1970, 63 ± 73. ciated with fen¯uramine-phentermine. New Eng J Med 1997; 58 Seiler KU, Wasserman O. MAO-inhibitory properties of 337=24: 1773 ± 1774. anorectic drugs. J Pharm Pharmacol 1973; 25: 576 ± 578.