INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cystic fibrosis is the most common lethal autosomal-recessive disease in the white population (1). Patients with a forced expiratory volume (FEV₁) less than 30% of the predicted value have a 50% chance of dying within 2 yr (2). Moreover, after adjustment for age and sex, the relative risk of death within 2 yr was 2.0 for each decrement in FEV₁ of 10% below the predicted value (2). Cystic fibrosis is caused by mutations in a single gene located on the long arm of chromosome 7 that encodes the cystic fibrosis transmembrane conductance regulator (CFTR) (3). CFTR has multiple functions regulating fluid balance across epithelial cells in the respiratory, hepatobiliary, gastro-intestinal, and reproductive tracts and the pancreas (4).

The inflammatory-immune process of infectious origin in the lungs of cystic fibrosis patients is associated with severe oxidative stress and amplified by deficiencies in lipophilic antioxidants due to exocrine pancreatic insufficiency (5, 6). This is particularly true for vitamin E, as a relative deficit in this antioxidant seems to predispose to oxidant lung injury (5). Several studies have reported altered ex vivo indices of oxidant stress in this setting (5, 6). However, these measurements suffer from major limitations due to lack of sensitivity or specificity in assessing the actual rate of lipid peroxidation in vivo (reviewed in Patrono and FitzGerald [7]).

Recently, a novel series of bioactive prostaglandin (PG) F₂-like compounds (isoprostanes) have been characterized. These are formed from nonenzymatic free radical-catalyzed peroxidation of arachidonic acid as the result of free radical attack of cell membrane phospholipids and low-density lipoprotein (LDL) oxidation (8, 9). Thus, F₂-isoprostanes are potential indexes of oxidative stress and lipid peroxidation in vivo as well as biochemical end points for antioxidant dose-finding studies (7). Moreover, the biological effects of nanomolar concentrations of the F₂-isoprostane, 8-iso-PGF₂, renders it a candidate molecule to transduce, at least in part, the effects of lipid peroxidation on platelet activation and pulmonary dysfunction (7).

Previous studies in patients with cystic fibrosis have suggested altered platelet function, as reflected by ex vivo measurements of platelet aggregation (10), eicosanoid production (11, 12), and response to inhibitors (12). However, the relevance of these capacity measurements to the actual occurrence of platelet activation in vivo is largely unknown (13). Platelet-derived thromboxane (TX) A₂ might exert local contractile effects on pulmonary blood vessels and bronchial smooth muscle (14), possibly contributing to the pathophysiology of cystic fibrosis. These effects might be shared by other contractile agon-ists, including other eicosanoids (e.g., sulfidopeptide leukotrienes [LT]) and isoeicosanoids (e.g., iso-PG, iso-TX, and iso-LT).

In the present study, we examined whether formation of F₂-isoprostanes is altered in vivo through measurements of urinary 8-iso-PGF₂ excretion in patients with cystic fibrosis. Moreover, we attempted to correlate this index of lipid peroxidation with the degree of pulmonary dysfunction and with thromboxane metabolite excretion, a noninvasive index of platelet activation in vivo (13). Finally, we tested the hypothesis that persistently enhanced lipid peroxidation is, at least in part, a consequence of inadequate vitamin E supplementation and can be modulated by increasing vitamin E levels. Our results are consistent with the hypothesis that enhanced lipid peroxidation is an important feature of cystic fibrosis and may contribute to persistent platelet activation and pulmonary dysfunction in this setting. Furthermore, the present results provide a rationale for reassessing the adequacy of vitamin E supplementation in patients with cystic fibrosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

From 1993 to 1998, 36 patients with cystic fibrosis (16 females and 20 males, aged 19 ± 8 yr; range 6 to 44) were recruited for the study. The height and weight of these patients averaged 156 ± 14 cm (range 117 to 175) and 47 ± 15 kg (range 17 to 77), respectively. Thirty-six healthy subjects (16 females and 20 males) with a similar age distribution (range 6-44 yr) were also studied during the same period.

The diagnosis of cystic fibrosis was based on characteristic clinical manifestations together with an abnormal sweat chloride concentration (15), i.e., 60 mmol/L or higher, by the standard Gibson-Cooke method (16). We excluded patients with comorbid conditions known to be associated with enhanced F₂-isoprostane formation (7). These included diabetes mellitus, hypercholesterolemia, and cigarette smoking. Moreover, patients with renal insufficiency or proteinuria (by serum creatinine levels and urinalysis) and altered hepatic function (by liver enzymes and echography) were also excluded.

All but one patients were found to be colonized with Pseudomonas aeruginosa on the basis of three consecutive sputum cultures obtained over a period of at least 6 mo.

Twenty-five patients were considered to have pancreatic insufficiency based on fecal fat excretion > 10% during a 3-d fat balance study.

All patients were taking pancreatic enzyme supplements orally in the form of Pancrease (Cilag spa, Milan, Italy) (three to six capsules per meal). All patients were receiving vitamin supplements that included vitamin B complex (B₁ 2 mg, B₂ 1 mg, B₆ 1 mg, vitamin H 0.1 mg, nicotinamide 10 mg), vitamin A (5,000 IU/d) and D (1,000 IU/d), vitamin C (200-600 mg/d), and E (d,l--tocopherol acetate, 200 mg/d). Some patients were taking oral antibiotics at the time of study: amoxycillin (Zimox, Pharmacia/Upjohn, Milan, Italy) (four patients), erythromycin (Eritrocina, Abbott, Campoverde, Italy) (two patients), ciprofloxacin (Ciproxin, Bayer, Milan, Italy) (eight patients), ofloxacin (Oflocin, Glaxo, Verona, Italy) (five patients), ceftazidime (Glazidim, Glaxo) (four patients), and tobramycin (Nebicina, Lilly, Sesto Fiorentino, Italy) (five patients). Other drugs included famotidine (Famodil, Sigma-Tau, Pomezia, Italy) (four patients), oral or inhaled salbutamol (Ventolin, Glaxo) (nine patients), and prednisolone (Urbason, Hoechst-Marion-Roussel, Milan, Italy) (six patients). None of the subjects had ingested any nonsteroidal antiinflammatory drug during the 2 wk prior to sampling and all were nonsmokers. All subjects were asked to avoid cold remedies during the study. Patients were instructed to follow a normal diet pattern with no specific restrictions throughout the studies. None of the patients was acutely ill at the time of study. The clinical and functional characteristics of the 36 patients are detailed in Tables 1 and 2.

Patients were followed-up on average for 36 mo (range 21 to 54) after measurement of prostanoid metabolite excretion. Nine died during follow-up.

Design of the Studies

Informed consent was obtained from all patients and volunteers after approval of the protocols by the Institutional Review Board.

In the first study, a cross-sectional comparison of urinary 8-iso-PGF₂ and 11-dehydro-TXB₂, a major enzymatic metabolite of TXB₂ (17), was performed between patients and controls. Overnight urine samples were added with 1 mmol/L 4-hydroxy-TEMPO (Sigma Chemicals, St. Louis, MO) as an antioxidant, frozen immediately, and stored at 20° C until extraction. To assess the reproducibility of urinary 8-iso-PGF₂ excretion, three consecutive 8-h urine samples were obtained from four patients with cystic fibrosis (1 M, 3 F; aged 14 to 20 yr).

Results from the first phase of the study indicated abnormally high 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion in patients with cystic fibrosis. Therefore, a second study was performed to examine the relative contribution of platelets to enhanced thromboxane metabolite excretion (18). For this purpose, four patients with cystic fibrosis (3 M, 1 F; aged 10 to 31 yr) were given aspirin 50 mg once daily for 1 wk with the evening meal. Overnight urine samples were collected before, 24 h after the last aspirin administration, and on the fourth and eighth day after withdrawing aspirin.

Because small amounts of 8-iso-PGF₂ can be formed by human platelets and monocytes through a cyclooxygenase (COX)-dependent mechanism (19), an additional study was performed to evaluate whether the inhibition of COX-1 and -2 activity had any influence on 8-iso-PGF₂ excretion in patients with cystic fibrosis. For this purpose 11 of the 36 patients with cystic fibrosis (6 F and 5 M; aged 6 to 41 yr) were randomized to receive ibuprofen, a nonselective COX-1/COX-2 inhibitor (22) (30 mg/kg/d) or nimesulide, a preferential COX-2 inhibitor (23) (3 mg/kg/d), each drug for 7 d. These patients collected overnight urine samples before dosing, on the last day of the randomized treatment, and 7 d after drug withdrawal, for measurement of 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion.

To investigate the short-term effects of increasing the level of vitamin E supplementation on urinary 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion, 10 of the 36 patients (4 F, 6 M; aged 10 to 44 yr) were recruited into an additional study based on their willingness to participate in a protocol requiring multiple blood and urine sampling. They were given 600 mg d,l--tocopherol acetate (Evion; Bracco, Milan, Italy) daily for 2 wk after the baseline evaluation. On Day 14, they were asked to return to the clinic with a 12-h overnight urine sample and had a fasting blood sample drawn for plasma vitamin E measurements. Thereafter, the patients went back to their usual dosage (200 mg/d) of the same formulation of vitamin E for 4 wk and repeated the same blood and urine collections.

Genetic, Biochemical, and Functional Analyses

Genotyping was performed in all patients (24); they were screened for the most frequent mutations in the CFTR gene in the Sicilian population (25).

Pulmonary function tests (spirometric measurements and determinations of arterial blood gases) were performed according to the American Thoracic Society standards (26) at least 3 h after arousal, using the Spiro Analyzer ST-250 (Fukuda Sangyo, Japan). Broncho-dilators were withheld for at least 12 h before testing. FEV₁, forced vital capacity (FVC), and forced expiratory flow at 25-75% of vital capacity (FEF_25-75) were each expressed as a percentage of the predicted normal value for the patient's age, sex, and height (26). The Sa_O2 was measured with an Oxyshuttle Pulse Oximeter (Sensor Medics, Anaheim, CA) (27).

Urinary 8-iso-PGF₂ and 11-dehydro-TXB₂ were measured by previously described and validated radioimmunoassay methods (28, 29). Measurements of urinary immunoreactive 8-iso-PGF₂ and 11-dehydro-TXB₂ have been validated using different antisera and by comparison with gas chromatography/mass spectrometry, as detailed elsewhere (28, 29).

Vitamin E plasma concentrations were determined by high-performance liquid chromatography (30). Urine creatinine was determined with a commercially available kit (Behring Testomar-Creatinina Combipack, Scoppito, Italy). Total cholesterol was measured as previously described (31).

Statistical Analysis

The data were analyzed by nonparametric methods to avoid assumptions about the distribution of the measured variables. An analysis of variance was performed with the Kruskall-Wallis method. Subsequent pairwise comparisons were made with the Mann-Whitney U test. The differences between baseline and posttreatment values were analyzed with the Wilcoxon signed-rank test. Moreover, the association of eicosanoid measurements with other parameters was assessed by the Spearman rank correlation test. All values are reported as means ± 1 SD. Significance was considered to be indicated by a p value of less than 0.05. All calculations were made with the computer program Stat View (Abacus Concepts, Berkeley, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline Measurements

Urinary 8-iso-PGF₂ excretion was significantly (p = 0.0001) higher in patients with cystic fibrosis (618 ± 406 pg/mg creatinine; mean ± SD, n = 36) than in age-matched healthy subjects (168 ± 48 pg/mg creatinine) (Figure 1). Moreover, 31 of the 36 (i.e., 86%) patients had excretion rates in excess of 2 SD above the control mean (Table 1). Urinary 8-iso-PGF₂ excretion was highly reproducible over three consecutive 8-h samples, with an intrasubject coefficient of variation of 15 ± 4 % (Table 3). Neither specific drugs (e.g., antibiotics) nor home oxygen (n = 6) accounted for enhanced F₂-isoprostane formation.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Urinary excretion rates of 8-iso-PGF2 in 36 patients with cystic fibrosis and 36 age-matched healthy subjects. Dots represent individual measurements.

Patients with cystic fibrosis had significantly enhanced 11-dehydro-TXB₂ excretion versus control subjects (2,440 ± 1,453 versus 325 ± 184 pg/mg creatinine; p = 0.0001) with 34 out of 36 (i.e., 94%) having metabolite excretion in excess of 2 SD above the control mean (Figure 2). A statistically significant correlation was found between 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion rates in the group of patients with cystic fibrosis (rho = 0.51; p = 0.0026). As detailed in Table 1, these biochemical abnormalities were not related to any specific genotype.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Urinary excretion rates of 11-dehydro-TXB2 in 36 patients with cystic fibrosis and 36 age-matched healthy subjects. Dots represent individual measurements.

The urinary excretion of 8-iso-PGF₂ was inversely related to FEV₁ (rho = 0.40, p = 0.0195). All but one patient with FEV₁ less than 50% of the predicted value had 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion rates in excess of 2 SD above the normal mean (Table 1).

Effects of Cyclooxygenase Inhibition

A COX-1 inhibitor, low-dose aspirin, was used to investigate the relative contribution of platelets to increased thromboxane metabolite excretion in patients with cystic fibrosis. Enhanced 11-dehydro-TXB₂ excretion was suppressed by > 80% following aspirin administration and recovered slowly on drug withdrawal (Figure 3), consistent with time-dependent return of platelet COX-1 activity to the systemic circulation (18). Urinary 8-iso-PGF₂ excretion was largely unaffected during platelet COX-1 inhibition with low-dose aspirin.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Urinary excretion rates of 8-iso-PGF2 (upper panel ) and 11-dehydro-TXB2 (lower panel ) before and after 1 wk of cyclooxygenase-1 inhibition with low-dose aspirin. Metabolite excretion was measured in four patients with cystic fibrosis before, 24 h after the last aspirin administration, and on the fourth and eighth day after withdrawing aspirin. The solid bars depict mean ± 1 SD. *p < 0.001.

Two structurally unrelated cyclooxygenase inhibitors, i.e., ibuprofen and nimesulide, were used to explore the potential contribution of inflammatory cell COX-2 activity to F₂-isoprostane formation in patients with cystic fibrosis. As shown in Figure 4, urinary 8-iso-PGF₂ excretion was not affected to any significant extent during 1 wk of COX-1/COX-2 inhibition achieved with either agent.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Urinary excretion rates of 11-dehydro-TXB2 (upper panels) and 8-iso-PGF2 (lower panels) before, during, and after cyclooxygenase inhibition with ibuprofen (left panels) or nimesulide (right panels). Eleven patients with cystic fibrosis were randomized to receive ibuprofen (30 mg/kg/d; n = 5) or nimesulide (3 mg/kg/d; n = 6) for 7 d. Metabolite excretion was measured before dosing, on the last day of treatment, and 7 d after withdrawing the drugs. The solid bars depict mean ± 1 SD values of metabolite excretion expressed as percentage of baseline measurements. *p < 0.05 versus baseline; **p = 0.0005.

These findings are consistent with a noncyclooxygenase mechanism of F₂-isoprostane formation in patients with cystic fibrosis, as characterized in other clinical settings (28, 32, 33). Interestingly, the extent of 11-dehydro-TXB₂ suppression following ibuprofen or nimesulide was not as pronounced as with low-dose aspirin, thus excluding an important contribution of extraplatelet sources (e.g., monocyte/macrophage COX-2) to enhanced TXA₂ biosynthesis in cystic fibrosis.

Effects of Vitamin E Supplementation

We also examined the effects of increasing vitamin E supplementation (from 200 to 600 mg/d) on the urinary excretion of 8-iso-PGF₂ and 11-dehydro-TXB₂ to test the hypothesis that inadequate vitamin E availability was, at least in part, responsible for enhanced lipid peroxidation and platelet activation in patients with cystic fibrosis. Urinary 8-iso-PGF₂ excretion measured twice, 6 wk apart, on vitamin E 200 mg/d was highly reproducible in these 10 patients and averaged 594 ± 397 and 550 ± 293 pg/mg creatinine, respectively.

Increased vitamin E supplementation caused significant (p = 0.005) changes in plasma vitamin E levels with a doubling of mean values from 11.6 ± 2.6 to 23.5 ± 9.6 µmol/L. As depicted in Figure 5, these changes were associated with significant (p = 0.005) reductions in urinary 8-iso-PGF₂ and 11-dehydro-TXB₂ excretion by 42% and 29%, respectively. A large proportion of the variability in the percentage reduction of urinary 8-iso-PGF₂ excretion associated with incremental vitamin E supplementation could be accounted for by the individual dose of vitamin E corrected for body weight (Figure 6).

View larger version (23K): [in this window] [in a new window]   Figure 5.   Effects of increasing vitamin E supplementation from 200 to 600 mg daily on 8-iso-PGF2 (upper panel ) and 11-dehydro-TXB2 (lower panel ) in 10 patients with cystic fibrosis. Metabolite excretion was measured at baseline, when patients were on long-term supplementation with 200 mg/d, at the end of 2-wk dosing with vitamin E 600 mg/d, and 1 mo after returning to vitamin E 200 mg/d. Dots and lines connecting dots represent individual repeated measurements in the same patient.

View larger version (8K): [in this window] [in a new window]   Figure 6.   Correlation between individual weight-adjusted vitamin E dose and percentage reduction in urinary 8-iso-PGF2 excretion in the 10 patients with cystic fibrosis who performed the vitamin E study depicted in Figure 5.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mutations in the CFTR gene result in defective chloride, sodium, and water transport in the epithelial cells of the pancreas and the respiratory, gastrointestinal, hepatobiliary, and reproductive tracts (1, 4). This in turn leads to luminal obstruction of the various exocrine ducts due to dehydrated and viscous secretions. Obstructive respiratory disease is responsible for over 90% of the mortality observed in patients with cystic fibrosis (2). Its pulmonary manifestations include airway obstruction by thick, mucopurulent secretions and chronic endobronchial bacterial infections with an intense inflammatory response eventually perpetuating airway and parenchymal injury, with progressive deterioration in lung function (15).

Although the precise sequence of events leading to bacterial colonization and chronic inflammation in patients with cystic fibrosis remains elusive, there is a clear association between malnutrition and deteriorating lung function (34). Increased oxidative stress is likely to play an important role in the pathophysiology of lung injury in this setting. Evidence of enhanced lipid peroxidation in patients with cystic fibrosis has been reported previously (5, 6, 35) and attributed to an imbalance between increased generation of free radicals and decreased availability of endogenous antioxidants. Free radicals, mainly derived from oxygen, have been implicated in a variety of human diseases, including cystic fibrosis. However, despite a vast interest in this field, the ability to measure reliably this process in vivo has remained elusive (7, 40).

In the present study we have measured the urinary excretion of 8-iso-PGF₂ as a marker of in vivo lipid peroxidation (7). Formation of F₂-isoprostanes and other isoeicosanoids (reviewed in Maclouf and coworkers [41]) reflects a nonenzymatic process of peroxidation of arachidonic acid on cellular membrane phospholipids, that is catalyzed by free radicals (8, 10). In contrast to lipid hydroperoxides, which rapidly decompose, F₂-isoprostanes are chemically stable end-products of lipid peroxidation that are released by phospholipases, circulate in plasma, and are excreted in urine (42). We have developed specific and sensitive immunoassay methods to measure 8-iso-PGF₂ (28), also known as iPF_2-III (43). This analytical approach has been previously used to demonstrate enhanced lipid peroxidation in other clinical settings, including hyper-cholesterolemia (32, 44), diabetes mellitus (33), and interstitial lung diseases (45).

The results of the present study demonstrate that formation and urinary excretion of 8-iso-PGF₂ are abnormally elevated in the vast majority of a group of 36 patients with cystic fibrosis with a wide range of lung disease severity. Patients with lower FEV₁ values had the highest metabolite excretion rates, thus suggesting that progressively deteriorating pulmonary function is associated with increasing rates of lipid peroxidation in this setting. Urinary measurements have the obvious advantage of noninvasiveness allowing repeated measurements over time, but are not informative as to the site(s) of origin and cellular elements involved in isoprostane formation. Because of the ubiquitous nature of the substrate generating F₂-isoprostanes in response to free radical attack, localization of 8-iso-PGF₂ formation is likely to be related to the source of reactive oxygen species. These in turn may reflect neutrophil activation as well as secretory products of Pseudo-monas aeruginosa, the most commonly found pathogenic bacterium in patients with cystic fibrosis (46). Evidence consistent with a pulmonary source of enhanced F₂-isoprostane formation was obtained recently through measurements of immunoreactive 8-iso-PGF₂ in the breath condensate of eight patients with cystic fibrosis, demonstrating approximately 3-fold higher levels than in healthy subjects (47).

Both persistent platelet activation, observed in the vast majority of our patients, and chronic lung inflammation might contribute, at least in part, to enhanced 8-iso-PGF₂ formation through a COX-1- or COX-2-dependent mechanism (19), respectively. However, the results of intervention studies with low-dose aspirin, ibuprofen, and nimesulide argue against this possibility. Thus, an aspirin regimen largely suppressing in vivo TXA₂ biosynthesis had no effects on urinary 8-iso-PGF₂ excretion (Figure 3). Moreover, neither ibuprofen at full antiinflammatory dosage shown effective in attenuating lung inflammation in patients with cystic fibrosis (48) nor nimesulide, a preferential COX-2 inhibitor with a COX-1/COX-2 IC₅₀ ratio of approximately 20 (23), had any dectable effect on 8-iso-PGF₂ excretion (Figure 4). Although we can exclude a role of COX-2 activity in F₂-isoprostane formation, the short-term nature of the ibuprofen/nimesulide study does not allow us to exclude that increased lipid peroxidation is, at least in part, a consequence of chronic lung inflammation.

Thus, a more likely explanation for enhanced F₂-isoprostane formation in patients with cystic fibrosis is that it reflects increased lipid peroxidation possibly due to altered prooxidant/antioxidant balance (5, 36). A major determinant of the latter seems to be related to inadequate vitamin E bioavailability, as previously suggested (37), and confirmed by the present findings.

Besides reflecting an ongoing process of enhanced lipid peroxidation in the lungs of patients with cystic fibrosis, increased local levels of F₂-isoprostanes might contribute to the progressive nature of lung function impairment in this setting. Thus, 8-iso-PGF₂ is a potent vasoconstrictor and induces DNA synthesis in vascular smooth muscle cells, through interaction with receptors that are distinct from but closely related to PGH₂/TXA₂ receptors (49). Moreover, 8-iso-PGF₂ displays smooth muscle constrictor activity in human and guinea pig bronchi in vitro (50) and induces airflow obstruction and airway plasma exudation in vivo in guinea pigs (51).

That enhanced formation of 8-iso-PGF₂ may exert clinically detectable biological effects in patients with cystic fibrosis is suggested by the linear correlation between its excretion rate and that of 11-dehydro-TXB₂, a noninvasive index of in vivo platelet activation (17, 29). This finding is consistent with the reported capacity of 8-iso-PGF₂, at nanomolar concentrations, to induce platelet adhesion (52) and amplify platelet aggregation in response to other agonists (53). Moreover, the present findings confirm and extend similar observations linking enhanced oxidant stress to platelet activation in other clinical settings such as type IIa hypercholesterolemia and diabetes mellitus (32, 33). The present results are novel inasmuch as they establish a relationship between F₂-isoprostane and TXA₂ biosynthesis in vivo outside the potential confounding effect of these complex metabolic abnormalities that per se may affect platelet function (31, 54). The results of the short-term vitamin E intervention study are consistent with the causal nature of such a relationship and with an important role played by 8-iso-PGF₂ in transducing the effects of oxidant stress on platelet activation. One potential limitation of the present study is represented by the absence of a functional assessment of platelet reactivity in response to various agonists in vitro. However, it should be pointed out that although platelet aggregation studies may be mechanistically informative, they do not reflect the extent of platelet activation in vivo, inasmuch as the biosynthetic capacity of human platelets to produce TXA₂ when challenged in vitro exceeds the actual rate of TXA₂ biosynthesis in vivo by several orders of magnitude (55). We are not aware of any classical platelet-mediated clinical syndrome affecting patients with cystic fibrosis, perhaps because of inadequate sample size of the studies to allow a reliable assessment of cardiovascular risk as well as because of early death of such patients due to pulmonary infections. Whether the platelet release of vasoactive and bronchoactive autacoids may contribute to potentially reversible changes in lung function remains unanswered by the present investigation because of the short duration and inadequate sample size of the low-dose aspirin protocol.

The results of the present study have clinical implications related to the adequacy of current levels of vitamin E supplementation in patients with cystic fibrosis. Deficiencies in vitamins A and E occur commonly in this setting (34). Current recommendation for supplemental vitamin E is 100-200 IU/d for patients aged 4 to 10 yr and 200-400 IU/d for patients older than 10 yr (34). Despite this level of supplementation, the vast majority of our patients with cystic fibrosis had biochemical evidence of abnormal lipid peroxidation associated with lower than normal plasma levels of vitamin E. Increasing the dose of supplemental vitamin E to 600 mg/d doubled the plasma vitamin E levels and produced a substantial reduction in F₂-isoprostane formation, without actually normalizing 8-iso-PGF₂ excretion in all patients. Larger than the average 42% reduction in F₂-isoprostane formation might be obtained by further increasing the level of vitamin E supplementation as suggested by the apparent weight-adjusted dose-response relationship (Figure 6).

Thus, our results provide a rationale and allow a formal sample size calculation for a dose-finding study of vitamin E supplementation in patients with cystic fibrosis, using urinary 8-iso-PGF₂ as a biochemical end point. The latter could also be used to assess the adequacy of vitamin E supplementation, which is likely to change over time as a function of disease progression. Finally, the pathophysiologic implications of enhanced formation of F₂-isoprostanes and other biologically active isoeicosanoids should lead to the design of new clinical trials of vitamin E in patients with cystic fibrosis with the rate of disease progression as the primary end point.