INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Obesity is epidemic in North America, affecting a third of the adult population (1). Consequently, dieting and the use of anorexigens (such as the modified amphetamines, fenfluramine, and dexfenfluramine) is common (2). A recent epidemiologic study in Europe found 6% of normal control subjects had been exposed to fenfluramine (3). Although well tolerated by many subjects, fenfluramine and dexfenfluramine have resulted in an outbreak of pulmonary hypertension (PHT) (3). The use of fenfluramine-related anorexigens for more than 3 mo increases the risk of developing PHT 23-fold (3), similar to an earlier "epidemic" caused by aminorex fumarate (2-amino-5-phenyl-2-oxazoline), a related anorexigen (4). The pathology in both reported outbreaks is reminiscent of primary PHT (P-PHT) (4, 5). This is potentially a major public health problem since it is estimated that 18,000,000 prescriptions for fenfluramine were filled in the United States in 1996 alone (2). Despite the recent withdrawal of fenfluramine and dexfenfluramine from the marketplace, some anorexigens are available clinically (phentermine and fluoxetine) and others are under development.

P-PHT itself is rare (annual incidence ~ 1/500,000 inhabitants [3]) and anorexigen-associated PHT (AA-PHT) also occurs in a small minority of the exposed population (3). Another factor must be present to lead to this disease. In experimental studies, dexfenfluramine, fenfluramine, and aminorex caused pulmonary vasoconstriction by inhibiting voltage-gated potassium channels in the smooth muscle cells of resistance level pulmonary arteries (PA) (6). Although pulmonary vasoconstriction to these anorexigens is seen in all normal rats, the magnitude of the response is small and tends to occur at high doses of the drug. However, if one inhibits nitric oxide (NO) synthase (NOS), dexfenfluramine causes marked pulmonary vasoconstriction at concentrations similar to those seen in humans receiving the drugs for weight reduction (7). Because the role of NO in the lung seems to be largely homeostatic, increasing to counterbalance acute vasoconstriction or chronic PHT (8), we hypothesized that persons with impaired NO production are predisposed to AA-PHT.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

A two-phase study of patients with PHT and normal control subjects was performed.

Phase 1: measurement of NO. A prospective study was made of lung NO production (VNO) in nine consecutive patients with AA-PHT and a similar number of patients with P-PHT (n = 8) seen concurrently at a tertiary care hospital for management of PHT between November 1996 and March 1997 (Table 1 and Figure 1). The investigators (K.D., S.L.A.) were blinded as to the cause of PHT. All subjects provided informed consent. The patients with AA-PHT and those with P-PHT had transthoracic echocardiograms, ventilation-perfusion scans, and lung function tests (FEV₁, FVC, and DL_CO) as part of their clinical evaluation. None had evidence of intracardiac shunts, significant thromboembolism, chronic obstructive lung disease, pulmonary fibrosis, or connective tissue disease. Only one patient (a patient with AA-PHT) had a history of cigarette use. Control subjects (n = 12) were sex-matched hospital staff and all were nonsmokers without cardiopulmonary disease. NO (in the nose, lung, and venous blood), minute ventilation, and room air O₂ saturation were measured. The subjects performed a symptom-limited, self-paced, 6-min walk and then measurements were repeated.

View larger version (31K): [in this window] [in a new window]   Figure 1.   Protocol for measuring breath NO. This is an actual trace from the resting portion of the NO measurement protocol in a normal volunteer demonstrating the ability to measure NO levels breath by breath (shaded insert). Breath NO was measured while the subject breathed first ambient air and then medical air (containing < 3 ppb NO). A marked reduction in expired lung NO occurs when the subject breathes medical air. Nasal NO concentrations approach ppm levels (end of trace) because relatively high nasal NO production is not diluted during apnea by passage of lower airway air (containing less NO).

Phase 2: chart review. The hemodynamics of all patients with PHT, both at the time of diagnosis and during therapy, the duration of exposure to anorexigen, and the subsequent clinical course were recorded. The response to acute, inhaled NO, functional class, and chronic therapeutic disposition were noted.

NO Measurement

The measurement of NO in expired air was performed as previously described (9, 10) using a NO analyzer with a detection threshold of 1 part per billion (ppb) (NOA 280; Sievers Instruments, Boulder, CO). Quantification of NO is based on the gas-phase interaction between NO and ozone (O₃). Some of the NO₂ produced by this reaction is in the excited state (NO₂*) and emits light as electrons return to ground state. Light is measured by a cooled, photomultiplier tube:

(1)

Acquisition of Breath NO

All measurements were performed after the patient had been seated comfortably in a quiet room for several minutes. Because the concentration of NO in Paris' air varies considerably (from > 150 to < 10 ppb), depending on the time of day and automobile traffic, all measurements of lung NO were performed while the patient breathed medical air containing < 3 ppb NO (Figure 1). Breath from the lower airways was sampled selectively using a face mask equipped with a two-way valve, to exclude room air, and a nasal compartment that compressed the nares and excluded nasal air ("non rebreathing mask" no. ABK0020; Sievers). Nasal air was drawn into the NO analyzer through Tygon tubing placed ~ 1 cm inside the left or right nostril (10). Measurements were taken from the lung during ~ 60 s of normal, quiet breathing. Nasal NO was measured from each nostril sequentially during normal respiration (opposite nostril occluded) and during apnea (30 s) after full exhalation. Measurement of NO during apnea eliminated variability caused by ambient NO levels. NO production (VNO_stpd) was measured as:

(2)

where E is the minute ventilation, 0.826 is a conversion factor to adjust volumes to standard temperature and pressure (11), and [NO]_exh is the exhaled NO concentration.

Measurement of Plasma NO_x

NO does not exist per se in the blood but is rapidly inactivated by interaction with hemoglobin or oxidized to form several nitrogen oxides (NO_x), in particular, nitrate (NO₃). NO_x can be measured in small volumes of plasma (0.1 ml) by converting it to NO using a strong reducing environment: vanadium (iii), 2 N HCl at 90° C (EQUATION 3). An antifoaming agent is added (FG-10; Dow Corning, Midland MI):

(3)

Statistics

Values are expressed as the mean ± standard error of the mean. Intergroup differences were assessed by a factorial ANOVA (p < 0.05 is considered statistically significant). Fisher's probable least significant difference test was used to evaluate differences within a group. Simple regression was performed to evaluate relationships between variables.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The AA-PHT group was older (p < 0.01) and tended to have a higher body mass index (p = 0.08) than the P-PHT group (Table 1). The patients had been exposed to fenfluramine or dexfenfluramine for periods ranging from 2 to 18 mo, although only one subject had received the medication more than 8 mo, (Table 2). No improvement was noted after withdrawal of the anorexigen. Furthermore, the PHT was fixed, with no vasodilator response to inhaled NO (Table 2). Systemic vascular conditions were noted in 56% of the patients with AA-PHT (systemic hypertension, transient ischemic attacks, or angina) but not in the P-PHT group (Table 2). One patient with AA-PHT died of PHT during the 3 mo after this study. Although all patients with PHT were judged to be New York Heart Association's Class III-IV, those with AA-PHT were particularly limited, with marked arterial desaturation and a trend toward shorter walking time (Figure 2 and Table 1). VNO was elevated in patients with P-PHT, but not in those with AA-PHT, relative to normal control subjects (Figure 3). Nasal NO concentration (Figure 4) was also less in AA-PHT versus the other groups. Paradoxically, venous NO_x levels were higher in patients with AA-PHT than control subjects or patients with P-PHT (Figure 5A).

View larger version (24K): [in this window] [in a new window]   Figure 2.   PHT and hypoxemia worse in AA-PHT than in P-PHT. Patients with AA-PHT had higher PVR, lower room air, arterial PO2, and desaturated more with exercise (O2 Sat) than did patients with P-PHT. Values are the mean ± SEM.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Lung VNO was reduced in patients with AA-PHT as compared with patients with P-PHT. Both expired NO concentration and NO production are diminished in patients with AA-PHT compared with patients with P-PHT. Values obtained at rest, breathing medical air. Values are the mean ± SEM. *p < 0.05 value differs from all other groups. p < 0.05 value differs from the P-PHT group.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Patients with AA-PHT have low nasal NO production during apnea. NO measured during a brief apnea and expressed as the integral of NO concentration over 30 s (AUC, area under curve). Values are the mean ± SEM. *p < 0.01 value differs from other groups.

View larger version (22K): [in this window] [in a new window]   Figure 5.   Venous NOx levels are elevated in patients with AA-PHT (panel A). Plasma NOx is elevated in AA-PHT, consistent with increased oxidation of NO to biologically inactive NOx species. Values are the mean ± SEM. p < 0.01 value differs from all other groups (panel B). NOx levels are inversely related to the observed VNO.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The primary finding of this study was that patients with AA-PHT have a deficiency of basal airway NO production (Figure 3) relative to patients with P-PHT. Conversely, in P-PHT, VNO was elevated compared with that in normal control subjects, consistent with experimental studies indicating that NO is a compensatory product of PA endothelium, which increases in PHT to counteract acute or chronic vasoconstriction (8). A secondary finding is that AA-PHT presents as a more severe syndrome than does P-PHT, with worse hypoxemia (Table 1 and Figure 2) and a higher incidence of systemic vascular conditions (i.e., systemic hypertension and angina; Table 2). A third new finding is that these subjects, most of whom were not morbidly obese, contracted the illness after taking fenfluramine or dexfenfluramine for 3 to 6 mo, and none had a spontaneous remission of illness. During several years of clinical follow-up, many remained symptomatic and required continuous infusion of prostacyclin.

Although anorexigens have led to an increased incidence of PHT, the great majority of people exposed do not develop the syndrome. Similarly, in normal animals, the anorexigens cause a consistent but minor elevation of PA pressure but do not, in the short term, lead to PHT (6, 12, 13). There appears to be some individual susceptibility factor necessary for the development of PHT. In the pulmonary circulation, NOS inhibition exaggerates the effects of the anorexigens, suggesting endothelial injury or NOS deficiency might predispose to a hypertensive response to anorexigens (6). Consistent with this, VNO is decreased in patients with AA-PHT compared with those with P-PHT. Furthermore, the severity of PHT, both at diagnosis and on vasodilator therapy, was inversely related to VNO in patients with AA-PHT (Figure 6). Although we do not know whether the NO deficiency preceded or followed the ingestion of anorexigen, it was not due to PHT itself (as PHT was associated with increased VNO in P-PHT) nor was it due to any mechanism requiring the presence of the anorexigen itself such as enhanced serotonin release (since all anorexigens had long been discontinued). Furthermore, there was no relationship between VNO and PVR (while receiving therapy or at initial diagnosis) in the P-PHT group (not shown).

View larger version (15K): [in this window] [in a new window]   Figure 6.   PVR is inversely proportional to VNO in AA-PHT. Those subjects with the lowest NO production had the highest PVR. This relationship did not exist in the P-PHT group.

Recent studies suggest there is a deficiency of basal NO synthesis in patients with untreated, essential systemic hypertension (14). This defect in EDRF activity seems to predate the development of hypertension, as defective acetylcholine-dependent vasodilatation can be demonstrated in the normotensive children of hypertensive parents (14). Thus, we postulate there is a genetic pool of EDRF-deficient subjects within the normal population. The finding of increased NO synthesis in P-PHT is consistent with many animal studies. In experimental PHT, EDRF activity and NO synthesis are usually preserved (8) and NOS protein expression is upregulated (15), which moderates the rise in PVR (8). However, in a recent study of NOS immunohistochemistry in P-PHT, a reduction of NOS expression was reported (16). No information was available as to possible anorexigen use, and NOS activity was not measured.

The mechanism by which NO production is impaired in patients with AA-PHT is uncertain. These patients were quite hypoxemic, and severe hypoxia can inhibit endothelial NOS in vitro (17). Severe hypoxia can inhibit EDRF activity in certain patients (18). A recent report by Dweik and colleagues (19) suggests airway NO production in normal humans is regulated by oxygen concentration throughout the physiological range. However, there was no correlation between PO₂ and VNO in either PHT group in this study. Alternatively, the low apparent VNO in patients with AA-PHT might reflect accelerated oxidative destruction of NO in the AA-PHT group. In support of this hypothesis, plasma NO_x levels are markedly elevated in patients with AA-PHT. Furthermore, there is a strong inverse relationship between VNO and plasma NO_x (Figure 5B). This is consistent with accelerated oxidation of NO, such that NO does not survive to escape into breath for measurement. Increased NO oxidation with diminished EDRF activity has been reported in conditions with elevated glucose levels or excess production of O₂ radicals (20, 21). Alternatively, the patients with AA-PHT may have, for unrelated reasons (genetic or acquired), a pre-existing NOS deficiency that was unimportant clinically until they were exposed to fenfluramine. The AA-PHT group was older than the P-PHT group (Table 1), and in the aging process many factors can lead to impaired EDRF activity (22). It must be conceded that high plasma NO_x levels could also be due to undetected factors other than enhanced oxidation-increased intravascular VNO such as intergroup differences in diet or renal function. However, there were no intergroup differences in kidney function noted (as indicated by blood urea nitrogen and creatinine). In addition, the inverse relationship between VNO and plasma NO_x only occurred in the AA-PHT group and was not found in the P-PHT group (data not shown). It appears unlikely that dietary nitrate intake would explain the elevated plasma NO_x in the patients with AA-PHT, as the patients in the P-PHT group, who were also very ill, had concentrations of plasma NO_x similar to those in the healthy volunteers. Thus, it is unproved that NO is being more rapidly oxidized in patients with AA-PHT, but this remains a possible explanation for low NO excretion by the lung in the face of high plasma NO_x.

There were three technical factors that we considered in designing this study. First, since the majority of breath NO comes from the nose (10, 23) or the paranasal sinuses (24), an occlusive mask was used to eliminate nasal contamination and measure NO exclusively from the airways. We also directly measured nasal NO to see if there was any difference in nasal NO between control subjects and patients with PHT. The depressed nasal NO levels in AA-PHT (Figure 4) supports the idea of a generalized lung NO deficiency. Second, we controlled for variable levels of NO in ambient air (the result of automobile engines and environmental pollution). VNO was measured with the subjects breathing medical air (containing < 3 ppb NO). In Figure 1 it is evident that on a day with high ambient NO levels, the lung was rapidly extracting ambient NO. Levels of NO and calculated VNO are much lower when the subject breaths medical versus ambient air, and only the former reflects endogenous production. The use of medical air is essential for measurements to be compared between centers or over time. Third, to calculate VNO one must correct for the E. The results can be quite discordant between NO production and concentration, particularly during exercise. Indeed, while [NO]_exh concentrations fall with exercise or hyperventilation, VNO rises as a consequence of a disproportionate increase in E (25). Our results are largely consistent with a recent report by Riley and colleagues (11) who noted that resting VNO and [NO]_exh were similar in patients with P-PHT and normal control subjects (11). The smaller rise in VNO in our normal control subjects compared with those in other studies (11) reflects the much milder protocol for exercise we used, necessitated by the inclusion of debilitated patients with PHT.

Patients with AA-PHT had a higher incidence of systemic vascular conditions (systemic hypertension, transient cerebral ischemic attack, or angina) than did those with P-PHT (56 versus 0%) (Table 2). Although we do not know the systemic blood pressure in the two groups prior to their first use of anorexigens, the association is supported by reports of systemic vascular complications stemming from use of anorexigens, including but not limited to myocardial infarction (26) and unstable angina (27). The precise mechanism for these effects is uncertain, but they may relate to anorexigen-induced systemic vasoconstriction, as occurs in the pulmonary circulation. Recently, anorexigens have been associated with cardiac valvular insufficiency (28). Although none of the patients with AA-PHT had cardiac valvular insufficiency, none had recent or prolonged exposure to the anorexigen (Table 2).

Limitations

This study lacked prospective measurements of the NO system prior to initiating anorexigen therapy. The low incidence of AA-PHT makes it difficult to acquire prospective NO measurements in a population of obese subjects and await the sporadic development of PHT. Although the sample size appears small, this study of 21 patients with AA-PHT and P-PHT constitutes roughly half the cases expected in France for a year. This report remains highly relevant despite the recent removal of dexfenfluramine and fenfluramine from the marketplace. The demand for drug treatments for obesity remains high, and many new anorexigens are under development. In addition, fluoxetine and phentermine remain available for treatment of obesity. Finally, mechanistic human studies such as this, which might be important in avoiding future outbreaks of PHT, are no longer possible since withdrawal of these drugs.