Respiratory Medicine (2005) 99, 14771484

EVIDENCE-BASED REVIEW

Non-antibiotictract infections

B. Arroll

Department of General PracticAuckland, New Zealand

Received 15 May 2005; accepted 7

Vitamin C; were clear about the lack of efficacy except in the high-quality studies in which aglobal improvement in symptoms was noted. Some studies were statistically

may be effective in children but is not available in the UK. Bisolvon (a mucolytic) was

ARTICLE IN PRESS

$Please see reference list for Cochrane reviews cited in this evidence-based review.Tel.: +64 9 3737599x86978; fax: +64 9 3737624.0954-6111/$ - see front matter & 2005 Published by Elsevier Ltd.doi:10.1016/j.rmed.2005.09.039

E-mail address: b.arroll@auckland.ac.nz.found to be effective for cough in only one study. For older children and adults,dextromethorphan may be effective (two out of three studies showed benefit), andguiafenesin (an expectorant) showed mixed benefit in two trials. Dexbromphenir-amine (a sedating antihistamine)/pseudoephedrine (6mg/120mg twice daily for 1week) was significantly more effective than placebo for severity of cough, whereas,in another study, loratadine (a non-sedating antihistamine)/pseudoephedrine (5mg/120mg twice daily for 4 days) did not show any difference between the study groups.Vitamin C may have a small role in preventing the common cold, with possibly agreater role in high-intensity physical activity and sub-arctic conditions. ZincZincsignificant, but the Cochrane reviewers were guarded about how clinically significantthey were. For Echinacea, problems were found with the quality of the studies andthe wide range of different forms of this substance. Heated humidified air seemed tobe effective in the UK and Israel, but not the USA, making definitive statementsabout efficacy difficult. Over-the-counter medication for cough seemed to have nodocumented benefit in children under the age of 5 years. Letosteine (a mucolytic)Nasal decongestant;Heated humidified air;KEYWORDSNon-antibiotic;Over-the-counter;Common cold;Cough;Antihistamine;Echinacea;

SummaryObjectives: To review the seven Cochrane reviews of non-antibiotic treatment forthe common cold.Methods: Each Cochrane review was read and summarized, and results presented asrelative risks and, where possible, numbers needed to treat.Results: The main theme that runs through these Cochrane reviews is the variablequality of the primary studies. In general, the reviewers are fairly cautious about thebenefits of any of the treatments other than first-dose decongestants andtreatments for upper-respiratory(common cold)$

e and Primary Health Care, University of Auckland, Private Bag 92019,

September 2005

antihistaminedecongestant combinations. For antihistamines alone, the reviewers

eff ceedpentimisecoleaevi

reference to the baseline risk. Both the ARR and

into consideration. The samefor harm. The NNTB and threported in this review if theycally significant studies or stapooled relative risks. The lby obtaining a pooled relameta-analysis and applyingexpected event rate (PEER). Tof events that occur in the grou

medication in a randomized trial comparing drug

Antihistamines for the common cold

ARTICLE IN PRESS

B. Arroll1478applies to the NNTHe NNTH were onlywere from statisti-tistically significantatter is calculatedtive risk from theit to the patienthe PEER is the ratep taking the placebo

The systematic review on antihistamines for thecommon cold3 was undertaken because the use ofantihistamines for the common cold is widespreadyet there was concern that they may not beeffective. The review contained 32 papers with 35comparisons; 22 trials were of monotherapy and 13were of a combination of antihistamines with othermedication. The number of participants totalled8930. The authors reported large differences inthe NNTB (the reciprocal of ARR) take these factorslozenges seemed to bemethodological issue ocalculated as numbers nneeded to treat for oneConclusion: Most non-aeffective. The most procough, antihistaminednasal decongestants (at& 2005 Published by Els

Introduction

This paper evaluates the seven Cochrane systema-tic reviews that deal with non-antibiotic treat-ments for upper-respiratory tract infection. Thiswas originally intended to include the review ofanti-viral treatment for the common cold, but thisreview has been removed from the library as it wasnot updated. The treatments in this review arethose that may be self-administered in manyjurisdictions (e.g. heated humidified air), as wellas being recommended in others by clinicians andeven subsidized by some health funders. Thus, theterm over-the-counter (OTC) is not strictlycorrect. Most of the clinical syndromes in thisreview fit the definition of the common cold, but,as mentioned in the earlier review in this series onantibiotics for upper-respiratory tract infections:a review of Cochrane reviews,1 there is the issueof microbiological aetiology to consider. It is notalways clear if an infection is of a viral, bacterial ormixed nature. In this paper, infections are assumedto be mainly viral, but the microbiology is notusually known or sought by clinicians.In this overview, effects may be reported as the

NNTB (number needed to treat for one person tobenefit) and the NNTH (number needed to treat forone person to harm). The NNTB is the inverse of theabsolute risk reduction (ARR) resulting from aparticular treatment in a particular group ofpatients.2 It is felt that it is better to report theARR than the relative risk reduction, as this termrefers to the benefit of a treatment without anywith placebo. Therefore, if the relative risk is lessthan 1 then the NNTB 1/{(1RR) PEER}, and ifthe relative risk is greater than 1 then theNNTB 1/{(RR1) PEER}.2A number of issues affect the quality of the

reviews covered in this paper. These include a widerange of design quality in the studies reviewed. Thedose and quality of some of the medicinal compo-nents and combination medications are also issuesof concern. The latter two of these apply particu-larly to OTC medications. Issues are also raised bydifferent methods of measuring end points, such ascontinuous scales compared with dichotomous out-comes (i.e. feeling well or absence of cough). Assome of the medications are designed for specificpurposes (e.g. cough suppressant), then absence ofcough will be appropriate. For decongestants, itwill be nasal discharge and stuffiness. This raisesthe question of what would be the appropriateoutcome for an antihistamine for the common cold.Ideally, the primary outcomes are specified

beforehand. However, it is not always possible totell if the analysis was done on the outcomes ofprimary interest or on the secondary outcomes. Inthis particular field, underpowering because ofsmall study size is also an issue. NNTB can only becalculated if dichotomous outcomes are used, butthey can be calculated for different symptoms,providing some practical information to guidetreatment decisions.fective, but the issue of unblinding due to taste was aoncern to the reviewers. The benefits and harms areed to treat for one person to benefit (NNTB) and numbersrson to harm (NNTH), and were calculated by the author.biotic treatments for the common cold are probably noting are dextromethorphan, bisolvon and guiaphenesin forngestant combinations for a wide range of symptoms,st for the first dose) and possibly zinc lozenges.er Ltd.

A benefit for global improvement was found in a

ARTICLE IN PRESS

Non-antibiotic treatments for upper-respiratory tract infections 1479subgroup analysis of the higher quality studies offirst-generation sedating antihistamines. The rela-tive risk (RR) for benefit was 0.9; 95% confidenceinterval (CI) 0.820.98. On the basis of the range ofPEERs in individual studies in this review, the NNTBfor a benefit was 12 for a PEER of 0.81 (this meansthat 81% of patients still had symptoms at the timeof analysis) and 17 for a PEER of 0.6. The RR foradverse effects for the sedating antihistamine was1.20 (95% CI 1.031.40). The NNTH for a PEER of0.05 was 93 and for 0.36 the NNTH was 14. For thenon-sedating antihistamines, no sedating adverseeffects were found (RR 1.10; 95% CI 0.552.18).For monotherapy in small children aged 25 years,there was little evidence of benefit. One large trial(6 months to 5 years) found no benefit, and onesmaller trial found improved rhinorrhea.

Combination antihistaminedecongestant

Similar lack of benefit was seen with combinationof antihistamines and decongestants for youngchildren. In older children and adults, evidence ofbenefit was found, and this was seen in five out ofthe six included trials. The RR for benefit in termsof global evaluation at 24 days was 0.31 (95% CI0.150.65). This is an NNTB of 2.5 for a PEER of 0.58and an NNTB of 4 for a PEER rate of 0.39. For theonly significant harm (dry mouth), the RR was 1.93(95% CI 1.352.75). The NNTH ranged from 246 forthe PEER rates of 0.023 and 0.54, respectively. Thissuggests that more patients may experience anadverse event (dry mouth) than will benefit fromcombination antihistamines and decongestants.The following interpretation can be made on thebasis that (1) combination therapies seem to beeffective; (2) monotherapy with antihistamines isnot effective; and (3) the systematic review ondecongestants (below)4 shows some benefit: it maybe the decongestant component of the combinationmedications that is providing the effectiveness, andthe antihistamine is contributing little or nothing toefficacy.

Echinacea

The systematic review on Echinacea5 included 16studies (eight were of prevention and eight ofstudy design, participants, interventions and out-comes. For monotherapy, no evidence was found ofa clinically significant effect on general recoverywith antihistamines as monotherapy for eitherchildren or adults.treatment for upper-respiratory tract infections),with a total of 3396 participants. There was suchvariation in the preparations used and the qualityof the study methods, that the authors felt thatthey could not pool the results.

Prevention in placebo trials

Two of the five prevention trials with a placebofound a statistically significant lower incidence ofinfection in the treatment group (RR 0.51 and0.67), whereas, in the other trials, there were onlytrends in favour of the treatment groups. Thepooled RR was 0.86 (95% CI 0.721.02). In assessingthe severity and duration of infections, three trialsfound no clear trends favouring the treatment orthe placebo group in the severity of recurringinfections.

Prevention trials with no placebo

In all three prevention trials with no treatmentcontrols, the number of children with infectionwas significantly lower in the group receivingthe Echinacea combination compared with theno treatment group. The pooled RR for thesestudies was 0.56 (95% CI 0.480.65). The concernwith these trials is that there was no placebo, socaution is needed in interpreting this positiveresult.

Treatment versus placebo

In the six trials comparing treatment with placebo,the outcomes (symptom scores, running nose andduration of illness) was a statistically significantpooled result. They were continuous outcomes, andhence an NNTB could not be calculated. In the fourtrials of treatment comparing placebo with runningnose as the outcome, the pooled result wasstatistically significant (weighted mean difference[WMD] 0.65; 95% CI 0.93 to 0.37). In two trialscomparing treatment with placebo for outcomes interms of duration of symptoms, the pooled resultwas statistically significant (WMD 1.0; 95% CI1.98 to 0.02). In a recent, large, placebo-controlled trial, Echinacea purpurea, as dosed inthis study, was not effective in treating upper-respiratory infection symptoms in children agedbetween 2 and 11 years, and its use was associatedwith an increased risk of rash.6 The authors of theCochrane review made the following comments:While overall there is some positive evidence, fewrecommendations can be made regarding the use ofEchinacea products in practice. The heterogeneity

cold

cold

placebo in reducing cough symptoms in doses of

ARTICLE IN PRESS

B. Arroll1480Five studies were suitable for inclusion in thereview on nasal decongestants for the commoncold,4 which contained only single component orsingle-dose studies. For the single-dose studies,four studies were included, and all were statisti-cally significant for a reduction in the nasalsymptom of congestion after treatment comparedwith placebo. The WMD between active andplacebo was 0.13 (95% CI 0.19 to 0.06).Discrete outcomes were not reported, hence anThis systematic review on heated, humidified airfor the common cold7 assessed the use of inhaledheated water vapour with the help of a rhinotherm(a machine designed to deliver heated watervapour to a persons nasal cavity) in the treatmentof the common cold. The control group usuallyreceived unheated humidified air. Efficacy of thisintervention in the studies carried out in UK andIsrael differed from those carried out in the USAwith similar equipment and methodology. Thestudies from Israel and UK reported beneficialeffect of rhinothermy in individuals with rhinophar-yngitis, whereas three randomized-controlled trialsfrom the USA failed to replicate the findings ofthese previous investigators. Equipment used toadminister the warm vapour differed slightly in theUK (anaesthetic mask) and USA (nozzle). Findingsfor symptom score, nasal culture washings or asubjective perception of benefit were not signifi-cant. The reviewers concluded by stating: sincethe studies reporting use of rhinothermy haveshown only subjective benefit in the symptoms ofthe common cold in the UK and Israel, this therapycannot be recommended universally because theresults of trials from the US are equivocal.

Nasal decongestants for the commonof the available preparations and the limitedquality and consistency of the evidence do notallow clear conclusions about which product mightbe effective in what dose and in what circum-stances. Patients and healthcare providers whowant to use preparations containing extracts ofEchinacea should be aware of the possible extremedifferences in the chemical composition and thatthere is no solid base of evidence concerning theirefficacy.5

Heated, humidified air for the common120mg per day or as a single dose of 50mg. Threestudies were included, all of which examineddextromethorphan in a single dose of 30mg. Twostudies favoured dextromethorphan over placebo,whereas a third study showed no effect. Moguis-teine in doses of 600mg/day was no more effectivethan placebo, apart from reducing cough in asubgroup of participants with more severe nightcough. Significant adverse effects were found inthe intervention group, which consisted mainly ofnausea, vomiting and abdominal pain (NNTH 7).

Expectorants in adults

Two trials compared guaifenesin with placebo. Inthe larger study, participants taking guaifenesinNNTB was not able to be calculated. No objectivedata were available for these studies. Overall, thereduction in nasal airway resistance for activemedication compared with placebo was significant.The forms of medication included oxymetazolinenasal drops, oral phenylpropranolamine, oral nor-phedrine and oral pseudoephedrine.For the multiple-dose studies, only one study

used oral psuedoephedrine 60mg.8 No significantbenefit was found for the symptom of congestionafter repeated doses of decongestant over a 5-dayperiod. Again, no discrete data were reported,hence NNTB could not be calculated. Because ofthe lack of other studies to support this result, noconclusion can be drawn on the efficacy ofrepeated doses of decongestant.

Over-the-counter medications for acutecough in children and adults inambulatory settings

Acute cough is a common symptom, and is oftenassociated with the common cold. Many peopleself-prescribe OTC cough medicines for themselvesor their children, and many primary-care cliniciansrecommend them to their patients as first-linetreatment. OTCs are available to the public frompharmacies and shops often without medical pre-scription, although, in some countries, they can beprescribed (Table 1).

Antitussives in adults

The Cochrane review of OTC medications for acutecough in children and adults in ambulatory settings9

assessed six trials comparing antitussives withplacebo. Codeine was no more effective than

200mg four times a day stated that the medicinewas helpful compared with placebo in reducing thecough frequency and intensity, with an NNTB of 2.Four participants (two in each group) reportedside-effects, including nausea and hives in theactive treatment group, and headaches, drowsinessand excessive perspiration in the placebo group. Inanother study, sputum thickness was significantlyreduced for guaifenesin 480mg 6 hourly with anNNTB of 2. Adverse effects were not reported inthis study.

Mucolytics in adults

One trial compared a mucolytic (bisolvon linctus)5mg three times daily for an average of 4 days withplacebo. Active treatment reduced cough fre-quency and symptom scores on days 4 and 8. TheNNTB was 15 and adverse effects were not

Other drug combinations in adults

Three studies compared combinations of drugsother than antihistaminedecongestant with place-bo. They contained multiple compounds, such asVicks Mednite, which contains dextromethorphan,doxylamine, ephedrine and paracetamol. WhenVicks Mednite was given as a single dose at bedtimefor 2 days, 57.6% of participants in the activetreatment group rated the formulation as goodor better in relieving cough compared with 32.2% inthe placebo group, with an NNTB of 4. Sevenparticipants in the active treatment group reportedgiddiness or drowsiness compared with four parti-cipants in the placebo group. Another medica-tion known as EM-VIER was more effective thanplacebo for reducing cough in adults (reducedcoughing fits 25% vs. 11%, NNTB 7 and reducedurge to cough with NNTB of 7). No adverse effects

ARTICLE IN PRESS

CoTBrte

Non-antibiotic treatments for upper-respiratory tract infections 1481reported.

Antihistaminedecongestant combinations inadults

Two studies compared antihistaminedecongestantcombinations with placebo. Dexbrompheniramine(a sedating antihistamine)pseudoephedrine (6mg/120mg twice daily for 1 week) was significantlymore effective than placebo for severity of cough.Adverse effects increased, including dizziness anddry mouth. In the other study, loratadine (a non-sedating antihistamine)pseudoephedrine (5mg/120mg twice daily for 4 days) did not show anydifference between the study groups. Thirty percent of the intervention reported dry mouth,headache and insomnia compared with 21% in theplacebo group. This translated to an NNTH of 11.

Table 1 Potential treatments for the common cold.

Type Symptom NNTB

Antihistamine Global improvement 1217Antihistaminedecongestant

General improvementand nasal symptoms

2.54

Dextromethorphan Cough NR

Nasal decongestant Congestion NRGuaifenesin Cough, sputum

thickness2

Bisolvon Cough 15Zinc lozenges Symptoms of cold 48

Some of these results differ from the conclusions of specificgive some direction to patients keen on taking a remedy. NNnumber needed to treat for one person to harm; NR, not repowere observed in both groups. Another studycompared a dextromethorphansalbutamol combi-nation, and dextromethorphan alone with placebo.Dextromethorphansalbutamol was superior to pla-cebo or dextromethorphan alone in relieving coughat night (mean symptom score 0.19 vs. 0.67 and0.44, respectively, on day 4. The dextromethor-phansalbutamol combination led to more tremorthan placebo, and no serious adverse effects werereported.

Antihistamines in adults

Three trials compared antihistamines with placebo.Antihistamines were no more effective than place-bo in relieving cough symptoms. Adverse eventswere about equal in the intervention and controlgroups.

NNTH Cochrane reviewers assessment

1493 Only for high-quality studies11 Dry mouth. Not effective in under

5-year oldsNR Two out of three studies show

benefitNR Benefit only at 1 dayNR Both studies had significant results

NR Only one study on this medication6 Concerns about unblinding

chrane reviews, but are presented in a comparative format to, number needed to treat for one person to benefit; NNTH,d.

Mucolytics in children

Two studies compared antihistaminedecongestant

Other drug combinations

showed a satisfactory response reported by theirparents compared with 57% of children in the

mega-dose range throughout the year if the only

about 28 days to 24 days per year per child. Such abenefit is not trivial, but is it worth the cost of long-

ARTICLE IN PRESS

B. Arroll1482placebo group, which did not reach statisticalsignificance (P 0:5). Adverse effects were notreported.Another RCT in 51 children compared a combina-

tion of dextromethorphan 1.5mg per ml andsalbutamol 0.2mg per ml 5ml three times dailyfor children under the age of 7 years or 10ml threetimes a day for older children with placebo. Therewas no evidence of efficacy.One trial involving 43 children tested two paedia-tric cough syrups (Triaminicol syrup and Dorcolpaediatric cough syrup). Compared with placebo,69% of children in both active treatment groups(brompheniraminephenylpropanolamine) combi-nations in children aged 6 months to 5 years, andneither was more effective than placebo. In one ofthe studies, a higher proportion of children wereasleep in the active treatment group (46.6%) thanin the placebo group (26.5%, P 0:53).A trial of letosteine (25mg three times daily for 10days with placebo) reported symptom score on afour-point scale favouring active treatment fromday 4 until day 10. This medication is not currentlyavailable in the UK.

Antihistaminedecongestant combinationsResults of studies in children

Antitussives in children

One study of 57 children (mean age 4.7 years; range18 months to 12 years) with night cough compareda single dose for 3 nights of dextromethorphan andcodeine with placebo. No benefit was found foreither medication over placebo.

Expectorants in children

No studies using expectorants met the reviewersinclusion criteria.term prophylaxis, and could an equivalent benefitperhaps be achieved in children through therapyalone? They also make the comment that, in somepopulations, there may be a low intake of vitamin Cand hence explain why some studies find a benefitanticipated benefit is to rather slightly shorten theduration of colds which occur for adults, two orthree times per year. Our pooled estimate suggeststhat long-term supplementation might result in anupper estimate average reduction of annual com-mon cold morbidity from about 12 days (based onDouglas 1979; unpublished Australian data) toabout 11 days per year for adults. For childrenunder 12 years who experience colds more fre-quently, long-term prophylaxis might be associatedwith an average reduction in 4 symptom days fromAntihistamines in children

One trial testing antihistamines in children aged1.55 years found that they were no more effectivethan placebo.

Vitamin C

The role of oral vitamin C (ascorbic acid) inpreventing and treating the common cold has beencontroversial for many years. Public interest ishigh, and vitamin C continues to be widely sold andused as a preventive and therapeutic agent for thiscondition. The systematic review on vitamin C10

only considered studies of vitamin C 200mg or moreper day.

Prevention of the common cold

Twenty-three studies in normal people wereincluded, and the pooled relative risk wasof borderline statistical significance (RR 0.98;95% CI 0.951.00). A subgroup of six papers study-ing marathon runners, skiers and soldiers on sub-arctic exercises reported another pooled relativerisk of borderline statistical significance for pre-vention of common colds (RR 0.96; (95% CI0.921.00).The Cochrane reviewers make this comment:

Although these findings point to a definitephysiological effect by prophylactic vitamin C oncommon cold duration, the practical significance ofthese findings is less convincing. It would not seemreasonable to ingest vitamin C regularly in the

regular vitamin C prophylaxis indicates that vitamin

treated with zinc complained of altered or bad or

the oral mucosa and distortion of taste have also

A large, high-quality factorial, randomized-decongestants and the combination with

Re

2 al

3 F.e

4 or3.

5 ore

6 dctA

7 d.y;

8 nal

antiinflammatory agent in the treatment of experi-

ARTICLE IN PRESS

Non-antibiotic treatments for upper-respiratory tract infections 1483occurred at higher incidences in participantstreated with zinc.unpalatable taste, which suggests zinc lozengeswere distinct from placebo lozenges and, in thisrespect, blinding may have been compromised.In one of the trials, the formulation contained13.3mg zinc, about half that used in most othertrials; the incidence of nausea in participantstreated with zinc (10/49) was significantly in-creased from that reported for participants takingplacebo (2/50). This is an NNTH of 6. Irritation ofC plays some role in respiratory defence mechan-isms.

Zinc for the treatment of the common cold

The systematic review of zinc for the treatment ofthe common cold11 included seven studies examin-ing the use of zinc lozenges for the treatment ofthe common cold. The reviewers were conservativeabout the findings, yet an intention-to-treat analy-sis at 7 days found a statistically significant RR of0.69 (95% CI 0.560.85). The NNTB ranged from 4 to8 for PEERs at 0.62 to 0.34, respectively. The authorswere concerned about the dose of medication andblinding of studies. In four trials, participantsand some do not. Most studies do not take intoaccount regular vitamin C intake.

Treatment of the common cold withvitamin C

Seven trials of treatment of the common cold aftersymptoms have been published, but no significantdifference was seen from placebo. However, onelarge trial reported equivocal benefit from a 4 gtherapeutic dose at onset of symptoms.

Prevention of the common cold withvitamin C

The failure of vitamin C supplementation to reducethe incidence of colds in the normal populationindicates that routine mega-dose prophylaxis is notrationally justified for community use. However,evidence shows that it could be justified in peopleexposed to brief periods of severe physical ex-ercise, cold environments, or both. Also, theconsistent and statistically significant small bene-fits on duration and severity for people using. Guyatt G, Rennie DRE. Users guide to the medicliterature. Chicago: AMA Press; 2002.

. De Sutter AI, Lemiengre M, Campbell H, Mackinnon HAntihistamines for the common cold. In: The CochranLibrary, Issue 2. Chichester, UK: Wiley; 2003.

. Taverner D, Bickford L, Draper M. Nasal decongestants fthe common cold. In: The Cochrane Library, IssueChichester, UK: Wiley; 2004.

. Melchart D, Linde K, Fischer P, Kaesmayr J. Echinacea fpreventing and treating the common cold. In: The CochranLibrary, Issue 1. Chichester, UK: Wiley; 2004.

. Taylor JA, Weber W, Standish L, Quinn H, et al. Efficacy ansafety of echinacea in treating upper respiratory trainfections in children: a randomized controlled trial. JAM2003;290:282430.

. Singh M. Heated, humidified air for the common colIn: The Cochrane Library, Issue 2. Chichester, UK: Wile2004.

. Sperber SJ, Sorrentino JV, Riker DK, Fg H. Evaluation of aalpha agonist alone and in combination with a nonsteroid1. Arroll B. Antibiotics for upper respiratory tract infections: anoverview of Cochrane reviews. Respir Med 2005;99:25581.placebo needs to be undertaken to clarifythe effective component(s)

Large, high-quality, randomized-controlledtrials are needed to clarify the efficacy ofdextromethorphan, bisolvon and guiafenesin

ferencescontrolled trial comparing antihistamines,Practice points

Negotiation about the likely benefits andharms of any of the non-antibiotic treat-ments is warranted, as few are consistentlybeneficial.

First-dose decongestants, either topical ororal, seem to be effective for symptomcontrol

Antihistamine decongestant combinationsseem to be effective for a wide range ofsymptoms

Patients need to be reminded that symp-toms do linger and that, at 10 days from theonset of illness with the common cold, 25%still have symptoms12

Research directions

mental rhinovirus colds. Bull New York Acad Med 1989;65:14560.

9. Schroeder K, Fahey T. Over-the-counter medications for acutecough in children and adults in ambulatory settings. In: TheCochrane Library, Issue 4. Chichester, UK: Wiley; 2004.

10. Hemila H, Chalker E, DSouza RR, Douglas RM, Treacy B.Vitamin C for preventing and treating the common cold.

In: The Cochrane Library, Issue 4. Chichester, UK: Wiley;2004.

11. Marshall I. Zinc for the common cold. In: The CochraneLibrary, Issue 2. Chichester, UK: Wiley; 2004.

12. Butler CC, Kinnersley P, Hood K, et al. Clinical course ofacute infection of the upper respiratory tract in children:cohort study. BMJ 2003;327:10889.

ARTICLE IN PRESS

B. Arroll1484