Portal de Periódicos da CAPES

Fenoterol, Asthma Deaths, and Asthma Severity

1997; Elsevier BV; Volume: 112; Issue: 4 Linguagem: Inglês

10.1378/chest.112.4.1148-b

1931-3543

Neil Pearce, Carl Rurgess, Julian Crane, Richard Reasley,

Pharmacological Effects and Assays

Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar (December 1996) report the findings of a cohort of 655 patients aged 15 to 55 years at a single Auckland hospital who were followed for the occurrence of death from asthma or a severe life-threatening attack. These are the same data that were previously published by Garrett et al,2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar except that Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar have considered all asthma drugs whereas Garrett et al2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar focused on the findings for fenoterol and salbutamol. However, publishing virtually the same data twice does not increase the validity of the study design or of the interpretation of the findings. Garrett et al2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar and Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar found that inhaled fenoterol was associated with severe life-threatening asthma (rate ratio=2.1), but that this risk decreased to 1.5 when the analysis was controlled for the markers of asthma severity used in our previous New Zealand case-control studies,3Crane J Pearce NE Flatt A et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1983: a case-control study.Lancet. 1989; i: 917-922Abstract Scopus (448) Google Scholar–5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar and further declined to 1.0 when the analysis was controlled for additional severity markers. Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar found similar patterns for other asthma drugs. Although we could take issue with many of the statements made by Rea et al1 and Garrett et al,2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar particularly those relating to the (mis)interpretation of our own studies, we will focus on four key issues which severely limit the conclusions that the Green Lane group draw with regard to our own studies of fenoterol and asthma deaths. First, they have studied a different population than we did. We studied persons aged 5 to 45 years in all major hospitals throughout New Zealand during the years 1977 to 1987, whereas they studied persons aged 15 to 55 years at a single hospital during 1987 through 1989 (although 18% were also admitted to other hospitals during the study period). The inclusion of older persons is a highly questionable action because of the difficulties in classifying asthma deaths6Sears MR Rea HH De Boer G et al.Accuracy of certification of deaths due to asthma: a national study.Am J Epidemiol. 1986; 124: 1004-1011Crossref PubMed Scopus (117) Google Scholar and ICU admissions in this age group. The different time period may also be important, since we also found less extreme risks during the later years of the epidemic.5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar However, what is most crucial is that their study was based on patients attending a single hospital in their own area, thus reflecting their own prescribing decisions and those of colleagues in their area. This may explain why they found that fenoterol was selectively prescribed to more severe asthmatics, whereas there is little or no evidence of this in hospital admissions for asthma for New Zealand as a whole.7Beasley R Pearce NE Burgess C et al.Confounding by severity does not explain the association between fenoterol and asthma death.Clin Exp Allergy. 1994; 24: 660-668Crossref PubMed Scopus (31) Google Scholar Second, they studied a different outcome than that addressed in our original studies.3Crane J Pearce NE Flatt A et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1983: a case-control study.Lancet. 1989; i: 917-922Abstract Scopus (448) Google Scholar–5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar Their study involved 90 ICU admissions and only 15 deaths, and the data were not presented separately. (The authors state that the unadjusted effect estimates were similar for deaths and ICU admissions, but the numbers of deaths were too small for the crucial adjusted estimates.) Our original studies found clear-cut evidence of an association of fenoterol with asthma deaths,3Crane J Pearce NE Flatt A et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1983: a case-control study.Lancet. 1989; i: 917-922Abstract Scopus (448) Google Scholar–5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar whereas our findings were more equivocal for ICU admissions.8Burgess C Pearce NE Thiruchelvam R et al.Prescribed drug therapy and near-fatal asthma attacks.Eur Respir J. 1994; 7: 498-503Crossref PubMed Scopus (28) Google Scholar There are considerable difficulties in interpreting studies of ICU admissions because of the increased potential for confounding, and the findings are often inconsistent with those of asthma deaths.9Beasley R Pearce NE The use of near-fatal asthma for investigating asthma deaths.Thorax. 1993; 48: 1093-1094Crossref PubMed Scopus (22) Google Scholar For example, Garrett et al2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar found no association between psychosocial problems and severe lifethreatening attacks;2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar this is inconsistent not only with our findings,5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar but also with their own previously published studies of asthma deaths.10Rea HH Scragg R Jackson R et al.A case-control study of deaths from asthma.Thorax. 1986; 41: 833-839Crossref PubMed Scopus (279) Google Scholar Third, this choice of a different population and a different outcome may explain why the Green Lane group got different results, even when using the same analytical methods that we used. When they adjusted for the recognized markers of asthma severity,11Crane J Pearce NE Burgess C et al.Markers of risk of asthma death or readmission in the 12 months following a hospital admission for asthma.Int J Epidemiol. 1992; 21: 737-744Crossref PubMed Scopus (159) Google Scholar the fenoterol relative risk decreased from 2.1 to 1.5. We observed no such change, and no evidence of confounding by severity, in our studies. (The one exception is the analysis,5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar which Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar and Garrett et al2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar selectively quote, that involved an additional control group which was more susceptible to confounding and which was included only because of complex methodologic issues raised by Boehringer Ingelheim critics, including one of Rea and Garrett's co-authors.12Poole C Lanes SF Walker AM Fenoterol and fatal asthma [letter].Lancet. 1990; 335: 920Abstract PubMed Scopus (10) Google Scholar.) In fact, the fenoterol relative risk increased markedly when we restricted the analysis to the most severe asthmatics, a phenomenon that indicates that our findings could not be due to confounding by severity.13Hensley MJ Fenoterol and death from asthma [letter].Med J Aust. 1992; 156: 882PubMed Google Scholar The Green Lane group did not find this, even when using the same analytic methods, suggesting that the differences are in the population and outcome being considered, rather than in the analytic methods. Furthermore, they found similar patterns for fenoterol and for other asthma drugs, whereas we found that only fenoterol showed consistently elevated risks in all analyses, and that these elevated risks did not decrease.and in fact increased.when we reduced confounding by restricting the analysis to the most severe asthmatics. Finally, the additional severity markers used by the Green Lane group were, with one exception, all used in one or more of our studies and, did not produce the findings that they report. The one apparent exception to this is that the Green Lane group also considered “severity of the previous attack” as measured by “inability to speak.” However, in our third case-control study, we considered “severity of the previous attack” as indicated by measures of respiratory and metabolic function, and found little or no evidence of confounding by severity.5Grainger J Woodman K Pearce NE et al.Prescribed fenoterol and death from asthma in New Zealand, 1981-1987: a further case-control study.Thorax. 1991; 46: 105-111Crossref PubMed Scopus (233) Google Scholar Once again, the difference appears to be in the population and outcome studied by the Green Lane group, rather than in the analytical methods. It should also be noted that the Green Lane group did not, as it claimed, adjust for baseline risk, since they also adjusted for markers of severity during the follow-up period. These markers, eg, the number of hospital admissions during the follow-up period, may represent intermediate stages of the causal process if the mechanism by which fenoterol increases the risk of fatal asthma involves increases in asthma severity over time. In summary, the studies of Rea et al1Rea HH Garrett JE Lanes SF et al.The association between asthma drugs and severe life-threatening attacks.Chest. 1996; 110: 1446-1451Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar and Garrett et al2Garrett JE Lanes SF Kolbe J et al.Risk of severe lifethreatening asthma and P-agonist type: an example of confounding by severity.Thorax. 1996; 51: 1093-1099Crossref PubMed Scopus (62) Google Scholar provide useful information on causes of near-fatal asthma attacks in their own area during 1987 through 1989. Unfortunately, they tell us little or nothing about the association between fenoterol and asthma deaths in New Zealand as a whole during the peak years of the epidemic. Fenoterol, Asthma Deaths, and Asthma SeverityCHESTVol. 112Issue 4PreviewWe have not published “virtually the same data twice.” As stated in the introduction to our paper, the objective was to look at the safety of drugs other than inhaled β-agonists. No previous study had done this. For all drugs, crude rate ratio estimates fell progressively when adjusted for an increasing number of potential confounding variables; this fact illustrates that controlling for severity is essential in exploring risk or safety in retrospective studies of asthma. Our study also showed that it is only possible to exclude confounding if one has access to a substantial number of well-defined severity markers. Full-Text PDF