The Influence of Ipratropium Bromide in the Recovery Phase of Methacholine Induced-Bronchospasm

Home , Methacholine

European Review for Medical and Pharmacological Sciences 2005; 9: 117-123 The influence of ipratropium bromide in the recovery phase of methacholine induced-bronchospasm

B. SPOSATO1,2, S. MARIOTTA1,2, A. RICCI1,2, P. BRUNO1,2, C. TERZANO1, F. MANNINO1,2

1Dipartimento di Scienze Cardiovascolari e Respiratorie, Università “La Sapienza” and 2U.O. Pneumologia, Azienda Ospedaliera Sant’Andrea, Rome – (Italy)

Abstract. – Background: Introduction The protective effect of Ipratropium Bromide (IB) in the methacholine-induced bronchospasm is well Ipratropium Bromide (IB) is an inhaled known from some time. The objective of the anticholinergic drug that exerts a bronchodi- present study was to assess whether a pre- lating action by antagonizing competitively treatment with IB may influence the subse- and non-selectively the acetyl-choline action quent phase of methacholine-induced bron- (and of other agonists) on muscarinic recep- chospasmMethods: relief. tors1. Its main indication is the treatment of Sixteen patients with bronchial COPD; however, it has been proved effective hyper-reactivity (PD20 FEV1 < 200 µg) were randomly assigned to three methacholine chal- in chronic asthma and in asthma exacerba- lenge tests at a 48 to 72 hours interval apart. tions, particularly in combination with beta2- In the first test IB was inhaled before the chal- agonists1-4. lenge (pre-IB), in the second IB was adminis- IB acts by blocking both the broncho-con- tered soon after the PD20 FEV1 (post-IB), and in stricting action of acetyl-choline (Ach) at the third no treatment was given (control). Results: the level of M3 muscarinic receptors of the The pre-IB PD20 FEVp 1 (695 ± 587.6 µg) was significantly greater ( < 0.0001) than bronchial smooth muscle, and the facilitat- that of post-IB (82.2 ± 49.18 µg) and of control ing action on the cholinergic neuro-trans- (73.9 ± 41.8 µg). The dose response slope mission exerted by M1 receptors localised at (DRS) (decline percentage of FEV1/cumulative the parasymphatetic ganglia level. More- methacholinep dose), in pre-IB was greatly low- over, IB is able to block (being a non-selec- er ( < 0.0001) than that in post-IB and in con- tive antagonist) the effect of the stimulation trol. During the bronchospasm relief phase, of pre-junctional M self-receptors localised the increase of FEV1 measured after 5, 10, 15, 2 at the post-gangliar vagal ending level, thus 30 and 60 minutes from the PD20p FEV1 was significantly greater in post-IB ( < 0.05) com- causing a greater Ach delivery and limiting pared with the other two conditions. Converse- the bronchodilating action of the anticholin- ly, the recovery slope (RS) (increase percent- ergic drug1,5. age of FEV1 at 60 minutes after PD20 FEV1 x cu- IB exerts a powerful bronchodilating ac- mulative methacholinep dose) was significantly tion in the methacholine-induced bron- more efficient ( < 0.001) in pre-IB than in chospasm and, when administered before the post-IBConclusion: and in control. In conclusion, ipratropium bro- challenge, a relevant protective effect due to mide confirmed to have a good protective ac- the competitive antagonism of the drug with tivity against methacholine-induced bron- methacholine at the muscarinic receptors’ chospasm; the pre-administration of ipratropi- level1,2,6-9. In fact, based on several trials, a 6 um bromide showed also a positive effect on to 50 fold increase of methacoline dose in IB the recovery phase. Key Words: pre-treated subjects is required to obtain a 6 20% fall of FEV1 . Bronchial hyperreactivity, Methacholine challenge 10 test, Ipratropium bromide, Recovery, FEV . Recently, Mannino et al have pointed in- 1 to evidence that the resolution of the metha-

117 B. Sposato, S. Mariotta, A. Ricci, P. Bruno, C. Terzano, F. Mannino choline-induced bronchospasm is more evi- Methacholine Challenge Test dent in subjects with moderate bronchial hy- The challenge was performed using a stan- per-reactivity compared with those having a dard procedure11,12. Methacholine sulphate more severe degree, due probably to a differ- 1% and 4% was provided by Lofarma (Milan, ent activity of the methacholine degradation Italy) and it was aerosolised by a MEFAR process. MB3 dosimeter (MEFAR elettromedicali On the basis of the above study, we investi- Brescia, Italy) using an MB2 ampoule. The gated whether the pre-treatment with IB may device had the following characteristics: influence not only the phase of methacholine- aerosol was obtained by compressed air at provoked bronchospasm, but also the subse- 1.75 atmosphere, airflow was 9 l/min, nebuli- quent recovery phase. sation time 1 second, output 10 ± 0.2 ml, vol- ume of the atomiser’s content 2 ml. The procedure of the challenge was firstly described10. Briefly, buffer and methacholine were diluted with distilled water and then four different Methods progressive methacholine solutions were pre- pared for the test (ampoule 1:1,25 mg/ml; am- Patients and Study Design poule 2: 2,5 mg/ml; ampoule 3:10 mg/ml; am- Sixteen non-smoker patients (mean age poule 4:40 mg/ml). The buffer solution was 26.3 ± 9.7 years, 9 males and 7 females), with administered first, followed by the administra- intermittent bronchial asthma (defined action of methacholine 12.5 µg, doubling the cording with GINA 2002 guidelines) and doses, step by step, according to the Chai 13 bronchial hyper-reactivity with PD20 FEV1 < method until PD20 FEV1 was obtained. 200 µg as measured in the methacholine challenge test, were randomly assigned to Lung Function Testing three methacholine provocation tests. The A heated Fleisch-type pneumotacograph challenge tests were performed using the with the Cosmed processing system (Cosmed same method at 48 to 72 hours apart. In two Quark 4, Pavona di Albano, Italy) was used tests IB (40 µg) was respectively adminis- for monitoring. FEV1 before and after inhala- tered prior to the methacholine challenge tion of the buffer solution and, step by step, (pre-IB) and soon after the PD20 FEV1 was after administration of the methacholine dos- reached (post-IB); the other test was per- es until PD20 FEV1 was obtained. FEV1 was formed with no drug intake (control). All measured between 30 and 90 seconds follow- participating patients underwent a skin ing each inhalation and the manoeuvre was prick-tests prior to study entry with routine repeated no more than two times; the highest allergens (Graminae, Parietaria, Olea euro- value was selected for the evaluation. The peae, Dermatophagoides pteronissinus and FEV1 measured after inhaling the standard Farinae, dog and cat dander, Alternaria saline solution was considered as the baseline tenuis, Aspergillus fumigatus). All patients value. In the three tests, for all patients the were asymptomatic at entry, were not taking differences in the baseline value of FEV1 be- β2-agonist bronchodilators from at least 24 fore starting inhalation of the buffer solution hours, inhaled or oral corticosteroids from were less than 10%. The recovery time at least 3 weeks, and were not suffering from course of the induced bronchospasm was airways infections or asthma exacerbations studied by measuring FEV1 at 5, 10, 15, 30 in the previous 4 weeks. IB was adminis- and 60 minutes after the 20% fall in FEV1. tered by using pressurised metered dose in- For each subject the methacholine cumula- halers (pMDIs) at the 40 µg dose (two 20 µg tive dose (CD), PD20 FEV1, the decline after puffs). The canister was placed at a 2-cm dis- PD20 and percentage increase of FEV1 at dif- tance from the mouth and was primed dur- ferent times were recorded. PD20 FEV1 was ing a slow and deep inspiration, holding the calculated from the cumulative dose response breath in the following ten seconds; the sec- curve. The effect of methacholine administra- ond dose was administered 30 seconds after tion was studied performing the dose re- the first puff. Informed consent was ob- sponse slope (DRS) on the basis of the O’- tained from all patients. Connor’s method14: decline percentage of

118 The influence of ipratropium bromide in the recovery phase of methacholine induced-bronchospasm

FEV1/cumulative methacholine dose (µmol), closed in “Primer of Biostatistics” (Stanton where the decline of the FEV1 was the differ- A. Glantz, McGraw Hill Inc., New York, ence between the percentage of this parame- 1997) was used for the statistical evaluation. ter after saline inhalation and the percentage PD20 FEV1 (µg), DRS and RS values were obtained after PD20 FEV1. The DRS obtained logarithmically (base 10) transformed for in the three conditions was statistically evalu- graphic presentation and the statistical analy- ated. The recovery of bronchoconstriction af- sis. The dose of methacholine delivered to ter PD20 FEV1 was studied by measuring: the mouth of subjects was calculated in terms of micromoles for the assessment of DRS and a) increase FEV1 percentage at 5, 10, 15, 30 RS. ANOVA for repeated measures was used and 60 minutes after PD20 FEV1. These for the comparison of FEV1, PD20 FEV1, CD, parameters were determined by the dif- DRS and RS changes within each condition. ference between the percentage of FEV1 Linear regression was applied for evaluating at PD20 FEV1 and the FEV1 percentage and comparing FEV1 increase in the recovery measured during the steps of the recov- phase (FEV1% increase) after induced bron- ery phase (5, 10, 15, 30, 60 minutes); choconstriction between the three protocols. b) the recovery slope (RS). Since the re- A p value < 0.05 was considered as statistical- covery is the opposite phenomenon to ly significant. the induced bronchoconstriction, the RS was obtained applying the following concept: increase percentage of FEV1 at 60 minutes after PD20 FEV1 × cumula- Results tive methacholine dose (µmol) used in the challenge. The increase FEV1% and Age, sex and functional findings relative the RS in the three protocols were sta- to the patients who underwent the metha- tistically compared. choline challenge tests, in three different conditions, are shown in Table I. The Statistical Analysis FEV1% baseline value before the metha- Data are shown as arithmetic mean and choline challenge test was similar in all the standard deviation. The software package en- three protocols of the study (102.56, 101.3

Table I. Features of 16 asthmatic patients and functional data obtained on the basis of three methacholine challenge tests.

Control Post-IB Pre-IB

Pts number 16 Age (year) 26.3 ± 9.7 Sex M/F 9/7 Prick-tests positive pts 14

FEV1% baseline 102.5 ± 10.15 101.3 ±7.12 102.56 ± 5.48

FEV1% after PD20 74.68 ± 4.25 75 ± 2.58 76.2 ± 2.56

FEV1% decline 27.8 ± 9.52 26.38 ± 7.55 26.37 ± 6.12

FEV1% increase at T60’ 20 ± 5.21 26.13 ± 5.0° 21± 8.34

Log10 PD20 FEV1 (µg) 1.79 ± 0.28 1.83 ± 0.32 2.62 ± 0.54*

Log10 Cumulative Dose (µg) 1.92 ± 0.30 1.96 ± 0.33 2.68 ± 0.53§

PD20 FEV1 (µg) 73.9 ± 41,8 82.2 ± 49.18 695 ± 587.6^

ANOVA for repeated measures was applied for statistical analysis.

° p < 0.0001, post-IB vs. pre-IB and Control; *, § and ^ p < 0.0001 pre-IB vs. post-IB and Control; FEV1% baseline: FEV1 before the test; FEV1% increase at T60: FEV1% increase obtained 60’ after PD20 FEV1; FEV1% decline: FEV1 decline percentage after administration of cumulative dose; Log10 PD20 FEV1 (µg): provocative methacholine dose (µg) administered to obtain a 20% fall in FEV1 logarithmically (base 10) transformed; Log10 Dose Cumulative (µg): methacholine dose administered for challenge test logarithmically (base 10) transformed; PD20 FEV1 (µg): provocative methacholine dose (µg) administered to obtain a 20% fall in FEV1.

119 B. Sposato, S. Mariotta, A. Ricci, P. Bruno, C. Terzano, F. Mannino and 102.4% in control, post-IB and pre-IB, the competitive antagonism between IB and 1,2,6-9 respectively), as the FEV1% decline was up- methacholine on the muscarinic receptors . on achievement of the PD20 FEV1 (27.8, Due to a greater receptor affinity of IB com- 26.3 and 26.3% in control, post-IB and pre- pared with muscarinic agonist, it is necessary IB, respectively). The increase of FEV1 to administer a higher dose of methacholine, measured 60 minutes after the PD20 FEV1 in comparison with the same subjects not was significantly greater in the post-IB pro- treated with IB, to remove the anticholiner- tocol compared with the other two ones (20, gic drug from receptors and, therefore, to 26.13 and 21% in control, post-IB and pre- cause bronchospasm. In fact, when the anti- IB, respectively; p < 0.0001). Both the PD20 cholinergic drug is administered after the FEV1 and, obviously, the total administered challenge (post-IB), an increase of the IB dose were greatly higher in the pre-IB pro- dose was not necessary to completely resolve tocol compared with the other two ones (p the bronchospasm. The IB protecting effect < 0.0001). Figure 1 shows the PD20 FEV1 as may also depend on the greater airways di- measured in the same subjects in the differ- ameter at the time of challenge, due to a de- ent experimental conditions. creased bronchomotor tone caused by the an- The comparison of the dose-response slope ticholinergic drug. Therefore, the induction (Log10 DRS: 1.79, 1.74 and 1.02 in control, of bronchospasm might take place from a post-IB and pre-IB, respectively) of the in- higher level of bronchodilatation, so that it is duced bronchoconstriction show slow signifi- necessary to administer a supplementary cant value for the protocol pre-IB in compar- dose of the agonist6. Furthermore, the IB ison with control and post-IB (Figure 2: p < could inhibit the release of inflammation me- 0.0001). The difference in the slope was relat- diators; it could block the cholinergic recep- ed to the higher dose of methacholine admin- tors of the mast-cells surface by inhibiting the istered to obtain PD20 FEV1 in pre-IB proto- intracellular release of cyclic guanosine 3’,5’- col compared with control and post-IB. monophosphate with a subsequent inhibition The increase of FEV1, as measured at the of the release of bronchoconstrictor media- 15 scheduled times after the PD20 FEV1 was tors, such as histamine . Compared with reached, was greatly higher in the post-IB beta2-agonists, anticholinergics exert their protocol than in the other two ones (Figure bronchodilating action in a later phase. In 3; p < 0.05). But, as mentioned above, the fact, the IB bronchodilating peak typically amount of methacholine administered in takes place 30 to 60 minutes post-dosing, pre-IB, in order to obtain a 20% fall in whereas the peak effect of beta2-agonists oc- FEV1, was higher. Using these doses as cor- curs just after 5 minutes from administra- 1,3,16,17 rective factor (FEV1% increase x cumula- tion . In our study, when the drug has tive dose) a higher recovery slope (log10 RS: been inhaled after the PD20 FEV1 has been 0.95, 1.08 and 1.68 in control, post-IB and reached, the increase of FEV1 measured at pre-IB, respectively) was found in pre-IB in the scheduled time was greatly in excess of comparison with control and post-IB (Fig- that measured in the other two sessions (Fig- ure 4; p < 0.01), suggesting (in these proto- ure 3), and the baseline value was reached af- col) a more effective recovery. ter 60 minutes. This effect demonstrates and confirms that IB is an effective bronchodilator in the methacholine-induced bronchocon- striction1,2,6. Discussion In subjects pre-treated with IB, the recovery of FEV1 at the scheduled times was simi- Our study confirms the protective effect of lar to that obtained with no treatment. Actu- the pre-medication with IB in the metha- ally, the difference in the two situations is choline-induced bronchospasm. In fact, ac- made by the amount of methacholine used to cording with previous findings1,2,6-9, it was nec- induce the bronchospasm (8 times higher in essary to administer a methacholine dose that pre-IB). In fact, by evaluating the relief of was higher (8 times in our study) in IB pre- bronchospasm in relation to the amount of treated subject compared to controls, to ob- used muscarinic agonist drug, the recovery tain a 20% fall of FEV1. This might be due to phase was more efficient in pre-IB (greater

120 The influence of ipratropium bromide in the recovery phase of methacholine induced-bronchospasm

Figure 3. Increase of FEV1 after PD20 FEV1 (at 5, 10,

Figure 1. PD20 FEV1 (Log10 µg) in the 16 asthmatic pa- 15, 30, 60 minutes) in 16 patients on the basis of three tients performed on the basis of three different proto- different protocols. cols (see Table I). Control: no drug was administered during methacholine Control: no drug was administered during methacholine chal- challenge test (MCH-test); pre-IB, post-IB: ipratropium (40 g) was administered before or soon after MCH-test, respec- lenge test (MCH-test); pre-IB, post-IB: ipratropium (40 µg) µ was administered before or soon after MCH-test, respectively. tively. For mean values and statistical analysis see Table I. slope) than in the other two experimental per-reactivity, differed not only during the conditions (Figure 4). Hence, pre-treatment methacholine-provoked bronchospasm in- with IB has caused in the same subjects the duction phase, but also in the following re- change from a high to a mild degree of hyper- covery phase, probably due to a different de- reactivity, not only by increasing the stimula- gree of activity of the enzymatic system that tion threshold of induced bronchospasm, but eliminates methacholine10. also by enhancing the following FEV1 recov- The explanation of the effects that take ery phase. This is consistent with the findings place during the methacholine-induced bron- of a previous work, in which we pointed into chospasm relief in subjects pre-treated with evidence that two groups of subjects, one IB is unknown. At present, no experimental with severe and another with moderate hy- evidences are available in literature. It is like-

Figure 2. Differences in dose response slope of three methacholine challenge test in 16 patients on the basis Figure 4. Difference in recovery slope performed in 16 of tree different protocols. patients on the basis of three different protocols. Dose response slope (DRS): decline percentage of FEV1/cu- Recovery slope: FEV1 increase at 60’ after PD20 FEV1 × cumu- mulative methacholine dose (µmol); Control (N): no drug was lative methacholine dose (µmol); Control: no drug was admin- administered during methacholine challenge test (MCH-test); istered during methacholine challenge test (MCH-test); pre- pre-IB (G), post-IB (L): ipratropium (40 µg) was administered IB, post-IB: ipratropium (40 µg) was administered before or before or soon after MCH-test, respectively. soon after MCH-test, respectively.

121 B. Sposato, S. Mariotta, A. Ricci, P. Bruno, C. Terzano, F. Mannino ly that the anticholinergic drug administered A further hypothesis might be that IB di- before the challenge test, by exerting a rectly activates the acetyl-cholinesterase via greater affinity for muscarinic receptors than the muscarinic receptor, thus causing a more that of methacholine, makes it necessary the rapid bronchospasm relief. In fact, as above administration of a higher amount of agonist reported, the subjects pre-treated with IB to remove IB from the receptor. Hence, when undergo a shift in the degree of hyper-re- part of methacholine has been metabolised by sponsiveness, moving from a high to a re- the acetyl-cholinesterase in the recovery duced level of hyper-reactivity both in the phase, IB (having a higher affinity) is reposi- induction and in the following relief phase tioned on muscarinic receptors by removing, of bronchospasm. Therefore, an enhanced in turn, the cholinergic agonist and thus exert- metabolism of methacholine may account ing its bronchodilating action by the competi- for this quantitative difference, which is tive antagonism not only towards the acetyl- probably caused by the pre-challenge IB in- choline, but also towards methacholine. halation. Another hypothesis to explain the greater In conclusion, ipratropium bromide con- efficiency of the relief of bronchospasm in- firmed to have a good bronchodilator and pro- duced in pre-IB may be due to the IB bron- tective activity against methacholine-induced chodilating effect that takes place approxi- bronchospasm; furthermore, the pre-adminis- mately 30 to 60 minutes after dosing. This is tration of ipratropium bromide showed a posi- not confirmed by our data, since the increase tive effect on the recovery phase. of FEV1 no more changed (i.e. increase of slope) after 15 to 30 minutes from the PD20 FEV1 has been reached, compared with the control protocol. Another explanation may be linked to the evidence that probably in IB References pre-treated subjects the great amount of GROSS NJ. methacholine localised at the pre-synaptic 1) Ipratropium bromide. N Engl J Med level may cause, after the PD20, a greater 1988; 319: 486-494. SCHLUETER DP. production and/or activation of acetyl- 2) Ipratropium bromide in asthma. Am cholinesterase (from a functional reserve) to J Med 1986; 81(5A): 55-60. limit (post-receptorial down-regulation) the CHERVINSKY P. 3) Concomitant bronchodilator therapy cholinergic stimulus. This is the typical result and ipratropium bromide. Am J Med 1986; of the acetyl-choline release that is rapidly 81(5A): 67-73. 18 O’DRISCOLL BR, TAYLOR RJ, HORSLEY MG, CHAMBERS hydrolysed by the cholinesterase enzyme . 4) DK, BERNSTEIN A Therefore, it is likely that the bronchospasm . Nebulised salbutamol with and relief would be more efficient with a greater without ipratropium bromide in acute airflow ob- methacholine degradation. In fact, it has struction. Lancet 1989; 1: 1418-1420. BARNES PJ. been demonstrated that in rats with choline- 5) Muscarinic receptor subtypes in air- deprived diet the acetyl-cholinesterase activi- ways. Eur Respir J 1993; 6: 328-331. MORRIS HG ty is reduced compared with normal fed con- 6) . Review of ipratropium bromide in in- trols, with a greater contractile response to duced bronchospasm in patients with asthma. acetyl-choline but not to carbachol or to 5- Am J Med 1986; 81(5A): 36-44. MAGNUSSEN H, NOWAK D, WIEBICKE W. hydroxy-triptamine19. It is likely that the re- 7) Effect of induced acetyl-choline production due to the haled ipratropium bromide on the airway re- choline-deprived diet causes a reduced sponse to methacholine, histamine and exercise in patients with mild bronchial asthma. Respira- acetyl-cholinesterase activity via a self-regu- tion 1992; 59: 42-47. lation process; conversely, a great amount of CRIMI N, PALERMO F, O LIVIERI R, POLOSA R, SETTINIERI I, 8) MISTRETTA A acetyl-choline or methacholine, as in our . Protective effects of inhaled ipratropi- case, may enhance its activity. In addition, um bromide on bronchoconstriction induced by the butyryl-cholinesterase, whose presence adenosine and methacholine in asthma. Eur at the bronchial level has been demon- Respir J 1992; 5: 560-565. AVITAL A, SANCHEZ I, CHERNICK V strated20, may also exert a potential role in 9) . Efficacy of salbuta- this possible self-regulation process of the mol and ipratropium bromide in decreasing cholinergic stimulation. bronchial hiperreactivity in children with cystic fi- brosis. Pediatr Pulmonol 1992; 13: 34-37.

122 The influence of ipratropium bromide in the recovery phase of methacholine induced-bronchospasm

MANNINO F, S POSATO B, RICCI A, GRASSO D, DE RUFFIN RE, FITZGERALD JD, REBUCK AS 10) 16) . A comparison CLEMENTI F, M ARIOTTA S. Induction and recovery of the bronchodilator activity of SCH 1000 and phases of methacholine-induced broncoconstric- salbutamol. J allergy Clin immunol 1977; 59: 139- tion using FEV1 according to the degree of 141. QUEZADA A, MALLOL J, MORENO J, RODRIGUEZ J bronchial hyperresponsiveness. Lung 2001; 179: 17) . Ef- 137-145. fect of different inhaled bronchodilators on recov- STERK PJ, FABBRI LM, QUANJER PH, 11) et al. Airway re- ery from methacholine-induced bronchoconstric- sponsiveness. Standardized challenge testing tion in asthmatic children. Pediatr Pulmonol 1999; with pharmacological, physical and sensitizing 28: 125-129. BARNES PJ. stimuli in adults. Report Working Party Standard- 18) Modulation of neurotransmission in ization of Lung Function Tests, European Com- airways. Physiological Reviews 1992; 72: 699- munity for Steel and Coal. Official Statement of 729. the European Respiratory Society. Eur Resp J ITABASHI S, OHRUI T, SEKIZAWA K, MATSUZAKI Y, SASAKI 1993; 16: 53-83. 19) H AMERICAN THORACIC SOCIETY. . Peripheral airway hyperresponsiveness in the 12) Guidelines for Metha- choline-deficently fed rat. Respir Physiol 1993; choline and Exercise Challenge Testing-1999. 92: 219-225. Am J Respir Crit Care Med 2000; 161: 309-329. NOREL X, ANGRISANI M, LABAT C, CHAI H, FARR RS, FROEHLICH LA, 20) et al. Degradation 13) et al. Standardiza- of acetylcholine in human airways: role of butyryl- tion of bronchial inhalation procedures. J Allergy cholinesterase. Br J Pharmacol 1993; 108: 914- Clin Immunol 1975; 56: 323-327. 919. O’CONNOR G, SPARROW D,TAYLOR D, SEGAL M, WEISS S. 14) Analysis of dose-response curves to methacholine. Am Rev Respir Dis 1987; 136: 1412-1417. CUGELL DW Corresponding author: 15) . Clinical pharmacology and toxicology of e-mail address: Bruno Sposato, MD, ipratropium bromide. Am J Med 1986; 81(5A): 18-22. [email protected]

123