The Association of Acetaminophen and Asthma Prevalence and Severity

John McBride argues that the evidence for a causal link between acetaminophen exposure and asthma in children warrants a policy of avoiding acetaminophen use for children at risk of asthma.[1] He makes this recommendation despite admitting that the evidence is not sufficient to warrant inference of a causal link, and calls for a randomized controlled trial. His claim is that the evidence as it currently exists warrants precautionary action despite its scientifically inconclusive nature, and his argument is thus a philosophical one, based on an ethical principle: "first, do no harm".

However, the principle of non-maleficence is wrongly invoked in this context, because to avoid doing harm, or even risking harm, one must know must know whether ones actions will cause, or risk causing, harm. The principle cannot be invoked to plug a gap in this knowledge, since this knowledge is required to invoke it. The principle is a moral one and adds nothing to situations of epistemic uncertainty. Moreover, doing nothing may itself result in harm. It is a hot philosophical potato whether harm caused by doing nothing brings moral culpability. However, the hot potato can be passed on in the present context, because the relation of a care- giver to a patient is one which (it is generally agreed) brings special duties. When a clinician has a patient under her care it not plausible that she can escape doing wrong by doing nothing.

A better way to make decisions under uncertainty of this kind is to evaluate both the probability and the seriousness that harm will occur due to the various courses of action contemplated, including doing nothing.

However, McBride's recommendation to refrain from using acetaminophen among children at risk of asthma remains unwarranted on this improved way of making decisions under uncertainty. This is because McBride does not review evidence concerning the impact of alternative courses of action, especially ibuprofen (McBride's recommended alternative), either on asthma or on other health outcomes.

A review of evidence concerning a causal link between acetaminophen and asthma does not suffice for assessing the likely effect of a policy restricting acetaminophen use, because it does not tell us about the likely effects of that policy. Those effects will depend on what is done instead of prescribing acetaminophen: doing nothing, prescribing something else, or adopting some non-pharmaceutical course of action.

This point is not merely theoretical. It has sharp practical relevance given the plausible links between non-steroidal anti- inflammatories and asthma. Existing evidence concerning these links must be reviewed before it can be said that existing evidence warrants avoiding acetaminophen use among children at risk of asthma.

McBride's recommendation is therefore not warranted by either the ethical argument he employs, nor the evidence he adduces to support it. No recommendation on acetaminophen use can be made by considering only the evidence concerning a link with asthma. Evidence concerning the likely effect of alternative courses of action, especially the effect of prescribing ibuprofen or aspirin, on both asthma and other health outcomes, must also be reviewed, before McBride's recommendation can be said to be justified.

[1] McBride JT. The Association of Acetaminophen and Asthma Prevalence and Severity. Pediatrics. 2011;128:1181-5.

Conflict of Interest:

None declared

McBride concludes that "Until future studies document the safety of this drug, children with asthma or at risk for asthma should avoid the use of acetaminophen."[1] The alternate explanation for this association is that it is due to confounding by indication. That is, acetaminophen (paracetamol) is used to treat the symptoms of the true risk factor - respiratory tract infections.

Viral respiratory tract infections are amongst the most common childhood infections and have been established as a strong risk factor for the development of and exacerbation of asthma in children[2 3]. Discomfort and fever associated with respiratory tract infections are also commonly the reason for administration of acetaminophen in young children. Given this, any analysis seeking to elucidate the relationship between acetaminophen use and asthma must also take into account the association with respiratory tract infections, either be adjusting for respiratory tract infections or limiting the study to acetaminophen use for non- respiratory indications[4]. Unfortunately, many of the studies cited by McBride"[1] as supporting a causal link between acetaminophen and asthma have not undertaken either approach[5]. Although the possibility of confounding by indication is acknowledge by McBride[1], there are a number of important pieces of evidence not considered within this review that suggest the association found between acetaminophen and asthma is likely to be due to confounding. We summarise these findings below to bring balance to this discussion.

We measured the days of use and indication for use of acetaminophen in the Melbourne Atopy Cohort study (MACS), a birth cohort of 620 children with a family history of allergic disease. In MACS, a crude analysis revealed greater total use of acetaminophen in the first two years of life (OR=1.18, 95%CI=1.00-1.39 per doubling dose), as well as acetaminophen use for respiratory tract infections, were associated with increased risk of childhood asthma[6], a finding that is consistent with many other studies in this area. However, adjustment for the frequency of infections in early life reduced these associations (OR=1.08, 0.91-1.29) and, importantly, acetaminophen use for non-respiratory indications was not associated with any allergic disease outcome, including asthma (OR=0.95, 0.81-1.12)[6]. Similar findings were seen in the LISA study[7], where children with asthma at age 6 had received more courses of acetaminophen in the first year of life for respiratory tract infections (mean=1.02) than children who did not have asthma (mean=0.79, p<0.01)[7]. There was no difference between asthmatic and non-asthmatic children in the early life use of acetaminophen for non respiratory infections (0.06 vs 0.05, p=0.89). Finally, using data from the Italian arm of phase two of the International Study of Asthma and Allergies in Childhood (ISAAC) [8] , Rusconi et al found that early life acetaminophen use was associated with increased risk of wheeze that starts early in life (transient early life wheeze: OR=2.27, 1.98-2.62, and persistent wheeze: OR=1.77, 1.49-2.10), but not wheeze that starts later in childhood (late onset wheeze: OR=1.12, 0.97-1.31). Our study[6] has clearly demonstrated confounding by indication, and other recent studies support this explanation for the observed association. Studies that do not take indication into account will overestimate the strength of the association between acetaminophen use in early life and risk of childhood asthma.

We do not advocate unnecessary use of any medication in children, including acetaminophen. Clinicians should continue to educate parents about the management of common childhood fevers and pain. However, the current state of evidence, concerning the role of acetaminophen as a cause of asthma, is not sufficiently convincing to avoid the use of acetaminophen in children if clinically indicated.

References

1. McBride JT. The association of acetaminophen and asthma prevalence and severity. Pediatrics 2011; 128: 1181-1185. 2. Jackson DJ, Gangnon RE, Evans MD, Roberg KA, Anderson EL, Pappas TE, Printz MC, Lee WM, Shult PA, Reisdorf E, Carlson-Dakes KT, Salazar LP, DaSilva DF, Tisler CJ, Gern JE, Lemanske RF, Jr. Wheezing rhinovirus illnesses in early life predict asthma development in high-risk children. Am J Respir Crit Care Med 2008; 178: 667-672. 3. Kusel MM, de Klerk NH, Kebadze T, Vohma V, Holt PG, Johnston SL, Sly PD. Early-life respiratory viral infections, atopic sensitization, and risk of subsequent development of persistent asthma. J Allergy Clin Immunol 2007; 119: 1105-1110. 4. Lowe AJ, Dharmage SC, Abramson MJ. Paracetamol use for non-respiratory indications and subsequent asthma: a valuable way to eliminate confounding by respiratory infections. Int J Epidemiol 2011. 5. Lowe A, Abramson M, Dharmage S, Allen K. Paracetamol as a risk factor for allergic disorders. Lancet 2009; 373: 120; author reply 120-121. 6. Lowe AJ, Carlin JB, Bennett CM, Hosking CS, Allen KJ, Robertson CF, Axelrad C, Abramson MJ, Hill DJ, Dharmage SC. Paracetamol use in early life and asthma: prospective birth cohort study. BMJ 2010; 341: c4616. 7. Schnabel E, Heinrich J. Respiratory tract infections and not paracetamol medication during infancy are associated with asthma development in childhood. J Allergy Clin Immunol 2010; 126: 1071-1073. 8. Rusconi F, Gagliardi L, Galassi C, Forastiere F, Brunetti L, La Grutta S, Piffer S, Talassi F. Paracetamol and antibiotics in childhood and subsequent development of wheezing/asthma: association or causation? Int J Epidemiol 2011; 40: 662-667.

Conflict of Interest:

SCD has received a grant from GlaxoSmithKline for unrelated work, MJA received funding from Reckitt Benckiser for an unrelated research project, and is on the GlaxoSmithKline Scientific Advisory Committee for the Australian Asthma Study, and these companies might have an interest in the submitted work. All other authors declare that they do not have any competing interests.