Effects of Prior Treatment with Salmeterol and Formoterol on Airway

Thorax 1996;51:585-589 585 Effects of prior treatment with salmeterol and

formoterol on airway and systemic 2 responses to fenoterol Thorax: first published as 10.1136/thx.51.6.585 on 1 June 1996. Downloaded from

Alison Grove, Brian J Lipworth

Abstract response curves to repeated doses of Background - Previous studies have fenoterol in stable asthmatic patients. shown that both salmeterol and formoterol (Thorax 1996;51:585-589) act as partial P2 receptor agonists in terms of antagonising the extrapulmonary re- Keywords: salmeterol, formoterol, fenoterol, partial sponses to fenoterol in normal subjects. agonist, bronchodilation, asthma. The aim ofthe present study was to extend previous observations in evaluating the Salmeterol and formoterol are both long acting effect of prior treatment with salmeterol P2 receptor agonists used in the treatment of and formoterol on bronchodilator re- asthma. It has been shown in vitro that sal- sponses to fenoterol, a full P2 receptor meterol is a partial 32 receptor agonist'2 and, agonist, in patients with asthma. even at high concentrations, it is unable to Methods - Ten stable asthmatic patients produce the maximal bronchorelaxant response of mean (SE) age 37 (3.7) years and forced elicited by a full agonist such as isoprenaline. expiratory volume in one second (FEV,) This difference can be expressed in terms of 59-5 (4-1)% of predicted completed the the intrinsic efficacy of the agonist, with a full study. One hour after inhaling single doses agonist such as isoprenaline nominally being ofplacebo, salmeterol 25 jg, or formoterol given an intrinsic activity of 1. Other agonists 12 jtg, dose-response curves to repeated can then be compared with this standard. In doses of inhaled fenoterol were con- vitro studies have shown that the intrinsic ac- structed (cumulative doses of100-3200 pg). tivity of salmeterol is of the order of 0 71 Measurements of airway and systemic P2 compared with isoprenaline.3 The intrinsic ac- receptor mediated responses were made tivity of formoterol is also lower than that at baseline, after inhalation of placebo, of isoprenaline with a value of 096,3 thus http://thorax.bmj.com/ salmeterol, or formoterol, and after each formoterol is also a partial agonist - albeit a dose of fenoterol. strong one. Results - Salmeterol and formoterol pro- From first principles it can be predicted that duced significant bronchodilation com- the presence of a partial agonist may inhibit pared with placebo (mean difference and the effects of an agonist with greater intrinsic 95% CI compared with placebo): FEV,, activity.4 In this respect we have previously salmeterol 0-41 (95% CI 0-13 to 0.69) 1, for- shown that oral salbutamol inhibits the effects moterol 0-47 (95% CI 0-19 to 0-75) 1. Sal- of endogenous adrenaline at extrapulmonary on October 2, 2021 by guest. Protected copyright. meterol and formoterol had no significant 12 receptors in normal subjects.5 In addition effect on systemic responses compared we have demonstrated that inhaled salmeterol with placebo. There were no significant and formoterol antagonise the extrapulmonary differences in peak airway responses to P2 receptor responses to endogenous adrenaline fenoterol after treatment with salmeterol and to exogenous inhaled fenoterol, also in or formoterol compared with placebo normal subjects.6 (mean (pooled SE)): FEV,, placebo The aim of the present study was to extend 2.84(0-03)1, salmeterol 2.87 (0.03)1, and the findings of these previous studies in order formoterol 2-88 (0.03)1. There were no sig- to assess whether prior treatment with sal- nificant differences in the area under the meterol or formoterol affects the airway and dose-response curve for any of the para- systemic 12 responses to repeated doses of in- Department of meters during the dose-response curve fol- haled fenoterol in asthmatic subjects. Fenoterol Clinical lowing treatment with salmeterol or was chosen to construct the dose-response Pharmacology, formoterol compared with There curves as it is known to be a full agonist with University of Dundee, placebo. intrinsic Ninewells Hospital and was no difference in the slope of the dose- greater activity than either salmeterol Medical School, response curves to fenoterol for FEVy or or formoterol. Low doses of salmeterol and Dundee DD1 9SY, UK forced expiratory flow after formoterol were used so as not to produce A Grove (FEF2575) maximal bronchodilator activity prior to ad- B J Lipworth treatment with salmeterol or formoterol ministering fenoterol. Correspondence to: compared with placebo, although there Dr A Grove. was a significant (p<005) attenuation of Received 15 August 1995 the slope in the dose-response curve for Methods Returned to authors 4 December 1995 the peak expiratory flow rate (PEFR). SUBJECTS Revised version received Conclusions - Prior treatment with low Ten stable asthmatic patients (five women) 10 January 1996 Accepted for publication doses of salmeterol or formoterol does not of mean (SE) age 37 (3 7) years and forced 18 January 1996 significantly alter bronchodilator dose- expiratory volume in one second (FEV,) 2'04 586 Grove, Lipworth

(0 23) 1,59 5 (4 1)% ofpredicted normal, com- acic Society criteria' using a Vitalograph completed the study. All patients were diagnosed as pact spirometer with pneumotachograph head having asthma according to American Thoracic and pressure transducer, and on-line computer Society criteria.8 At an initial screening visit assisted determination of FEVL, FEF2, ,-, and

patients were required to have an FEV, of PEFR. Forced expiratory manoeuvres were Thorax: first published as 10.1136/thx.51.6.585 on 1 June 1996. Downloaded from less than 80% of predicted normal, and to performed from total lung capacity to residual demonstrate at least 15% reversibility to in- volume with measures being taken according haled fenoterol 200 jg (Berotec 200 metered to best test criteria. dose inhaler; Boehringer Ingelheim, Bracknell, Berkshire, UK). In addition, all subjects were required to have a normal physical ex- Extrapulmonary responses amination, 12 lead ECG, haematology and A standard lead II electrocardiogram was mon- biochemical screen. All subjects gave written itored and recorded with a Hewlett-Packard informed consent to participate in the study ECG monitor and printer (Palo Alto, California, which had been approved by the Tayside com- USA) with paper speed set at 25 mm/s. Heart mittee for medical ethics. At the time of the rate was calculated from the mean of five con- study all 10 patients were using inhaled cortico- secutive R-R intervals. steroids in doses of400-2400,g daily, together Finger tremor was recorded by a previously with inhaled short acting bronchodilators on validated method10 using an accelerometer an as required basis. In addition, four patients transducer (Entran Ltd, Ealing, UK). Four were oral recordings were made at each measurement taking theophylline preparations. and the results were stored on computer for subsequent analysis of total tremor power >2 Hz (mg2/s) using computer assisted auto- PROTOCOL Patients attended the laboratory between 08.00 covariance. The mean of three consistent re- and 09.00 hours on three occasions separated cordings was subsequently used in the analysis. by at least one week. Before each visit broncho- Serum potassium levels were measured by dilators were withheld for an appropriate period flame photometry (IL943 analyser, Instru- of time (that is, 48 hours for theophylline pre- mentation Laboratory Ltd, Warrington, UK) parations and eight hours for short acting 32 with analysis being performed in batches at the agonists). A cannula was inserted into a forearm end of the study, and samples being assayed vein to facilitate venous blood sampling. After in duplicate. The coefficients of variation for 30 minutes supine rest, baseline measurements analytical imprecision within and between as- of heart rate, postural finger tremor, serum says were 0 41% and 1-04%, respectively. The potassium levels, and spirometric parameters normal reference range for serum potassium were performed. was required to be levels in our laboratory is 3 5-555 mmol/l.

FEV, http://thorax.bmj.com/ within 15% of that recorded at the initial screening visit. Subjects were then randomised STATISTICAL ANALYSIS to receive inhaled placebo, salmeterol 25 jtg Data for finger tremor were transformed using (Serevent metered dose inhaler, 25 jtg per actu- logarithm to base 10 to achieve conformation ation; Allen and Hanburys, Uxbridge, Middle- with a normal distribution. Data were then sex, UK), or formoterol 12,ug (Foradil metered analysed using a Statgraphics software package dose inhaler, 12 gg per actuation, Ciba Geigy (STSC Software Publishing Group, Rockville, AG Basel, Switzerland) in a single blind cross- USA). Baseline values, and values following over fashion. All drugs were administered via inhalation of placebo, salmeterol, and for- on October 2, 2021 by guest. Protected copyright. a large volume spacer device in an attempt to moterol, were compared by multifactorial anal- maximise lung delivery. After one hour further ysis ofvariance (MANOVA). Where the overall measurements of airway and systemic para- MANOVA was significant, Duncan's multiple meters were made. A dose-response curve to range testing was used to establish where inhaled fenoterol (Berotec metered dose in- differences were significant. Peak values halers, 100 jig and 200 jig per actuation) was achieved during the dose-response curve by constructed using doses of 100 jig, 100 pg, each individual were obtained, regardless of 200 jg, 400 jig, 800 jg, and 1600 jg- that the doses at which they occurred, and were is, a total cumulative dose of 3200,g. The compared by MANOVA and Duncan's mul- fenoterol was also administered via the large tiple range testing. Analysis of the responses at volume spacer device. Dose increments were individual doses from the dose-response curve given at 20 minute intervals with measurements was not performed in order not to confound being made 15 minutes after each dose. On the error. The area under the curve for each completion of the dose-response curve subjects parameter was also obtained using the tra- received potassium supplements in the form pezoidal rule and compared by MANOVA. In of effervescent potassium 32 mmol (Sando-K, order to compare the overall dose-response Sandoz Pharmaceuticals, Camberley, Surrey, curves least squares regression analysis was UK). performed on the linear part of the curve for each individual and the resulting gradients were compared by analysis of variance. MEASUREMENTS A probability ofp<0 05 (two tailed) was taken Airway responses to be of significance for all tests. Values are given Measurement of FEV,, forced expiratory flow in the text as means (pooled SE). Differences (FEF25-75), and peak expiratory flow (PEF) from placebo, where significant, were calculated were performed according to American Thor- as means and 95% confidence intervals. Airway and systemic f2 responses to fenoterol 587

Table 1 Mean (SE) baseline values before treatment with inhaled placebo, salmeterol 2.9 25 jig, orformoterol 12 jig A Placebo Salmeterol Formoterol FEV,(l) 1-93(0-09) 2-14(0-09) 2-10(0-09) FEF25 1-26(0 14) 1-37(0-14) 1 40(0-14) 75(0/s) Thorax: first published as 10.1136/thx.51.6.585 on 1 June 1996. Downloaded from PEFR (l/min) 333(11) 358(11) 338(11) HR (beats/min) 73(2) 71(2) 76(2) >-2.6 Tremor (log units) 2 30(0 12) 2-31(0- 15) 2 25(0- 16) mr Potassium (mmol/l) 3 88(0 06) 3 86(0 05) 4 02(0 06) LL FEV, = forced expiratory volume in one second; FEF2,75 = forced expiratory flow; PEFR peak expiratory flow rate; HR = heart rate.

2.3 Table 2 Mean (SE) airway and systemic parameters one hour after receiving inhaled placebo, salmeterol 25 pg, orformoterol 12 pg 100 1000 10 000 Placebo Salnmeterol Fornmoterol 2.6 FEV,(l) 2 03(0 07) 2 44(0.07)* 2 50(007)* FEF5,75 (l/s) 1-38(0 13) 1-96(0 13)* 1-83(0 13)* PEFR (l/min) 347(11) 408(11)* 415(11)* 2.4 HR (beats/min) 71(2) 70(2) 72(2) Tremor (log units) 2 23(0 09) 2 09(0 09) 2-27(0 10) Potassium (mmol/l) 4-01(0 06) 3 93(0 06) 3-97(0 06) r- 2.2- Abbreviations as in table 1. * LLu 2 p<0 05 compared with placebo. U- 2UnL Results BASELINE VALUES (table 1) 1.8 There were no significant differences between 100 1000 10 000 baseline values at each visit for any of the parameters measured. 490

470 RESPONSES TO PLACEBO, SALMETEROL OR C FORMOTEROL (table 2) : 450 Both salmeterol and formoterol produced a in and significant increase FEVI, FEF2515, X 430 PEFR compared with placebo (mean difference U] and 95% CI): FEV,, salmeterol 0-41 (95%CI CLl 0 13 to 0 69) 1, formoterol 0-47 (95% CI 0 19 410 http://thorax.bmj.com/ to 0 75) 1; FEF2575, salmeterol 0-58 (95% CI 390 0 07 to 1 09) l/s, formoterol 0-46 (95% CI 0 04 11 0 to 0 87) l/s; PEFR, salmeterol 61 (95% CI 20 1000 1 0 000 to 102) 1/min, formoterol: 68 (95% CI 27 to Dose of fenoterol (,ug) 109) 1/min. There were no significant differences between the bronchodilator responses to Figure 1 Dose-response curves to inhaledfenoterol (cumulative doses of 100-3200 jg) constructed one hour salmeterol 25 .g and formoterol 12,ig. after inhaling placebo salmeterol 25 pig (V), or

(0), on October 2, 2021 by guest. Protected copyright. Neither salmeterol nor formoterol signi- formoterol 12 pg (A) for (A) FEV,, (B) FEF2, -,, and ficantly increased heart rate or finger tremor, (C) PEE or lowered serum potassium levels compared with placebo. pared with placebo. There were no significant DOSE-RESPONSE CURVES TO FENOTEROL differences in the area under the dose-response Because of the significant confounding effect curve for airway or systemic responses following ofsalmeterol and formoterol on baseline airway treatment with salmeterol or formoterol com- parameters compared with placebo, the dose- pared with placebo. response curves are shown as absolute values, Compared with placebo there were no sig- rather than as changes from baseline (figs 1 nificant differences in the peak values for FEV,, and 2). Dose-dependent increases in FEVI, FEF2515, or PEFR obtained during the dose- FEF2575, and PEFR were seen, together with response curves following treatment with sal- dose-dependent increases in heart rate and fin- meterol or formoterol. Likewise, there were ger tremor and a fall in serum potassium levels. no significant differences in the peak systemic Regression analysis revealed no significant responses after treatment with salmeterol or differences in FEV, or FEF2575 responses after formoterol compared with placebo (table 3). treatment with salmeterol or formoterol compared with placebo, although there was a stat- istically significant (p<005) attenuation in the Discussion slope for PEFR responses after treatment with The results of the present study show that prior both salmeterol and formoterol compared with treatment with low doses of salmeterol and placebo. Regression analysis revealed no sig- formoterol have no significant effects on the nificant differences in systemic responses after bronchodilator or systemic P32 receptor me- treatment with salmeterol or formoterol com- diated responses to fenoterol either in terms of 588 Grove, Lipworth

explained by the difference in doses of long acting 12 agonist used in the two studies. In the present study it is likely that a relatively Z- 3.8 low degree of receptor occupancy occurred, E and this did not therefore antagonise the effects Thorax: first published as 10.1136/thx.51.6.585 on 1 June 1996. Downloaded from E of fenoterol. The dose-response curves for systemic 12 effects may, however, be suggestive of n 3..4 a non-significant trend towards a reduction in a)E heart rate and finger tremor after U) treatment with salmeterol compared with placebo. An- other factor worth considering is that receptor 3.0 occupancy may be greater at steady state after 1100 1000 10 000 chronic dosing as compared with single dosing. 95 However, as 12 receptor downregulation and associated subsensitivity to salbutamol occurs after chronic dosing with salmeterol," it would not be possible to separate out this phe- Ec, 85 nomenon from that of 12 receptor antagonism. co U) The primary aim of the present study was to -0 evaluate the effect of ihese long acting 12 agon- a) ists on bronchodilator responses to fenoterol. X. 75 If higher doses of salmeterol or formoterol co a) had been administered it is likely that they themselves would have produced the maximum bronchodilation in an individual, making it to assess 100 1000 10000 impossible interactions with fenoterol because of confounding effects on airway geo- 3.6 metry. The effect on airway geometry is clearly c evident from the shape of the dose-response curves, in that a plateau response for FEV, and en 3-2 PEFR was achieved at a lower dose offenoterol following active treatment. It should be L-CM pointed 0)c 2.8 out that, irrespective of prior treatment, the final airway response achieved was comparable. 0 This is not, however, an explanation for failure E a) 2.4 to demonstrate antagonism of systemic effects where a ceiling in response did not occur. http://thorax.bmj.com/ In vitro studies are not subject to such con- 2.0 iLE straints. In a study using precontracted guinea 100 1000 10000 pig trachea and human bronchus, high con- Dose of fenoterol (jig) centrations of salmeterol (O 1-1l O,mol/l) inhibited relaxant responses to a variety of other Figure 2 Dose-response curves to inhaledfenoterol with intrinsic (cumulative doses of 100-3200 ,g) constructed one hour 132 agonists higher efficacy.'2 after inhaling placebo (0), salmeterol 25 ,g or Interestingly, it was noted that the degree of

(V) on October 2, 2021 by guest. Protected copyright. formoterol 12 jg (A) for (A) serum potassium, (B) heart inhibition produced appeared to vary de- rate, and (C) finger tremor. pending on the agonist used - for example, responses to fenoterol were inhibited less than dose_ responses to salbutamol. These findings rep- the peak values achieved or in the overall resent a deviation from conventional theories response curves. governing agonist/antagonist interactions In contrast to the present study, we have shown in which predict that an antagonist should pro- previously normal subjects thE7t2i- duce a parallel shift to the right of the dose- haled salmeterol 300,ug and formoterol 72ptg response curve for a given agonist. In other antagonise the extrapulmonary 12 receptor words, the magnitude of the shift should be mediated responses to both endogenou h ald determined by properties of the antagonist renaline released during exercise and in] a e rather than the agonist. The inference is that fenoterol.6 This apparent discrepancy m:ay be antagonism may have been observed had salbutamol rather than fenoterol been used to Table 3 Mean (SE) peak values for airway and systemic parameters obtained durring a construct the dose-response curve. dose-response curve to inhaled fenoterol (total dose 3200 Mg) constructed one hour aJfter In this respect, Smyth and co-workers ex- receiving inhaled placebo, salmeterol 25 ,g, orformoterol 12 jg amined the interaction between salmeterol and Placebo Salmeterol Formoterol salbutamol in vivo. " They found that sal- FEV,(I) 2 84(0 03) 2 87(0 03) 2 88(0-03) meterol in doses of 50-200,g had no sig- FEV, (%) 85 4 86-3 86 4 nificant effect on bronchodilator or systemic FEF25-5 (I/s) 2-54(0-09) 2 45(0 09) 2 64(0 09) FEF25,5 (%) 61 9 59 8 64 5 responses to salbutamol in a group of mild PEFR (1/min) 473(6) 474(6) 491(6) asthmatics. Even after receiving the lowest dose PEFR (%) 101-7 102-0 104-2 HR (beats/min) 90(3) 87(3) 92(3) of salmeterol, it was evident that salbutamol in Tremor (log units) 3 42(0 17) 3 16(0-13) 342(017) a cumulative dose of 3600,g produced very Potassium (mmol/l) 3 34(0 06) 3-24(0 06) 3-18(0-06) little additional increase in FEVI. Both the Abbreviations as in table 1. present study and that of Smyth et al assessed Airway and systemic 12 responses to fenoterol 589

activity in vitro and in vivo. Br Jf Pharmacol 1991;104: interactions between long and short acting 665-71. agonists at the peak effect of the long acting 2 Linden A, Bergendal A, Ullman A, Skoogh BE, Lofdahl CG. Salmeterol, formoterol, and salbutamol in the isolated drug.'415 It may also be relevant to assess such guinea-pig trachea: differences in maximum relaxant effect interactions towards the end of a normal dose and potency but not in functional antagonism. Thorax 1993;48:547-53. - interval that is, 12 hours after administering 3 Anderson GP. Formoterol: pharmacology, molecular basis Thorax: first published as 10.1136/thx.51.6.585 on 1 June 1996. Downloaded from the drug. After 12 hours it would be expected of agonism, and mechanism of long duration of a highly potent and selective 02-adrenoceptor agonist broncho- that the long acting drug would be producing dilator. Life Sci 1993;52:2145-60. significant but not maximal bronchodilation. 4 Waller DG. ,B-adrenoceptor partial agonists: a renaissance in cardiovascular therapy. Br Jf Clin Pharmacol 1990;30: The results of the present study, together 157-71. with the work of Smyth et al, are reassuring in 5 Grove A, McFarlane LC, Lipworth BJ. Expression of the 32-adrenoceptor partial agonist/antagonist activity of sal- that they suggest that the presence of a long butamol in states of low and high adrenergic tone. Thorax acting agonist does not attenuate the airway 1995;50: 134-8. P2 6 Grove A, Lipworth BJ. Evaluation of the 02-adrenoceptor response to shorter acting agonists with higher agonist/antagonist activity of formoterol and salmeterol. Thorax 1996;51:54-8. intrinsic efficacies, at least in stable asthmatic 7 O'Donnell SR, Wanstall JC. Evidence that the efficacy subjects under conditions of basal broncho- (intrinsic activity) of fenoterol is higher than that of salbutamol on f,Badrenoceptors in guinea-pig trachea. EurJ motor tone. It is worth noting, however, that Clin Pharnzacol 1989;36:239-45. in vitro studies have suggested that differences 8 American Thoracic Society. Standards for the diagnosis and care ofpatients with chronic obstructive pulmonary disease in intrinsic activities between agonists become and asthma. Am Rev Respir Dis 1987;136:225-44. more pronounced in the presence of increased 9 American Thoracic Society. Standardisation of spirometry - 1987 update. Am Rev Respir Dis 1987;136:1285-98. bronchial tone.'6 In human bronchus pre- 10 Lipworth BJ, McDevitt DG. Beta-adrenoceptor responses contracted with acetylcholine, for example, sal- to salbutamol in normal subjects. Eur Jf Clin Pharmacol 1989;36:239-45. meterol behaves as an even weaker agonist with 11 Grove A, Lipworth BJ. Bronchodilator subsensitivity to an intrinsic activity of 0-62 compared with inhaled salbutamol after regular twice daily treatment with salmeterol in asthmatic patients. Lancet 1995;346:201-6. 0 71 at basal tone.'6 A similar pattern was 12 Kallstrom BL, Sjoberg J, Waldeck B. The interaction be- also demonstrated with formoterol. Thus it is tween salmeterol and f2-adrenoceptor agonists with higher efficacy on guinea-pig trachea and human bronchus in conceivable that in patients with acute severe vitro. BrJPharmacol 1994;113:687-92. 13 Smyth ET, Pavord ID, Wong CS, Wisniewski AFZ, Williams asthma, with a higher degree ofbronchial tone, J, Tattersfield AE. Interaction and dose-equivalence of relatively weak partial agonists such as sal- salbutamol and salmeterol in patients with asthma. BMJ 1993;306:543-5. meterol may antagonise the bronchodilator 14 Derom EY, Pauwels RA. Time course of bronchodilating effects of salbutamol or fenoterol. This issue effect of inhaled formoterol, a potent and long-acting sympathomimetic. Thorax 1992;47:30-3. may merit further investigation, particularly in 15 Ullman A, Svedmyr N. Salmeterol, a new long-acting in- view of concerns raised over a possible increase haled 02-adrenoceptor agonist: comparison with salbutamol in asthmatic patients. Thorax 1988;43:674-8. in asthma deaths in patients prescribed sal- 16 Naline E, Zhang Y, Qian Y, Marion N, Anderson GP, meterol. 7 Grandordy B, et al. Relaxant effects and durations of action offormoterol and salmeterol on the isolated human bronchus. Eur RespirJ1 1994;7:914-20. 17 Castle W, Fuller R, Hall J, Palmer J. Serevent nationwide

1 Ball DI, Brittain RT, Coleman RA, Denyer LH, Jack D, surveillance study: comparison of salmeterol with sal- http://thorax.bmj.com/ Johnson M, et al. Salmeterol, a novel, long-acting 02- butamol in asthmatic patients who require regular adrenoceptor agonist: characterisation of pharmacological bronchodilator treatment. BMJ 1993;306: 1034-7. on October 2, 2021 by guest. Protected copyright.