Management of suspected acute viral upper respiratory tract infection in children with intranasal sodium cromoglicate: a randomised controlled trial

ARTICLES

Management of suspected acute viral upper respiratory tract infection in children with intranasal sodium cromoglicate: a randomised controlled trial Chris C Butler, Mike Robling, Hayley Prout, Kerenza Hood, Paul Kinnersley

Summary Background Upper respiratory tract infection in children is one of the most frequent reasons for visiting a family doctor, and antibiotics are often prescribed inappropriately. Sodium cromoglicate inhibits the ICAM-1 molecule, which is the receptor for human rhinoviruses. We aimed to investigate whether intranasal cromoglicate shortens duration of infection of the upper respiratory tract. Methods We randomly assigned 290 children diagnosed with suspected acute viral upper respiratory tract infection by their family doctor (137 boys, 153 girls; mean age 5·2 years [SD 3·39]) either intranasal 4% sodium cromoglicate spray or intranasal normal saline spray. Follow-up was by daily symptom diary for 2 weeks and by telephone. Canadian Acute Respiratory Illness and Flu Scale (CARIFS) score was the primary outcome measure. Findings 195 patients returned symptom diaries, and 20 of these could not be included in the main analysis. 246 patients completed the telephone interview at week 1. There was no difference in recovery rate over the first week between the two groups, with the estimated difference in slope of log (CARIFS) being –0·01 (95% CI –0·05 to 0·03). There were no differences between the two groups in sideeffects or re-consultation rates. 43 (17%) of 246 children with suspected acute viral upper respiratory tract infection went back to see their family doctor, and 220 (89%) of 246 were managed without prescription of antibiotics. Interpretation Intranasal sodium cromoglicate is not a useful additional treatment for this infection. Our results further clarify the role of prescribed drugs for children with these frequent illnesses. Lancet 2002; 359: 2153–58

Department of General Practice, University of Wales College of Medicine, Cardiff, UK (C C Butler MD, M Robling BSc, H Prout BSc, K Hood PhD, P Kinnersley MD) Correspondence to: Dr Chris C Butler, Department of Family Medicine and Centre for the Evaluation of Medicines, McMaster University, Faculty of Health Sciences, 1200 Main Street West, HSC 2V14, Hamilton, Ontario L8N 3Z5, Canada (e-mail: [email protected])

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

Introduction Infection of the upper respiratory tract in children is one of the most frequent reasons people go to see their family doctor, and antibiotics are regularly prescribed inappropriately.1 This behaviour has led to concerns about bacterial resistance. These infections in children account for 30% of absences from school, and family routines are disrupted.2 Over 90% of infections of this type are caused by viruses, most frequently rhinoviruses.3 On average, a child 4 years old or younger has six to ten infections of the upper respiratory tract a year, and this frequency could double for children of this age in day care or at nursery school.4 Up to 29% of infections of this type in the first 3 years of life lead to otitis media, probably by Eustachian tube inflammation.4 At least 80% of reported exacerbations of asthma in children aged 4–11 years are associated with upper respiratory tract infection, caused mainly by human rhinovirus.5 An ideal treatment for these infections should reduce morbidity and prevent complications, be safe for the individual and community, and be obtainable without consultation of nurses or doctors, thus adding to options for self-care and reducing pressure on health services. The intercellular adhesion molecule-1 (ICAM-1) is the receptor for 90% of human rhinoviruses, and has a central role in recruitment of inflammatory cells and release of soluble mediators of inflammation, including tumour necrosis factor.6,7 These cells and molecules in turn induce ICAM-1 expression.8 The host’s response to rhinovirus might therefore facilitate the spread of viruses to uninfected cells by the inflammatory response, increasing expression of virus receptors on host cells.9 The release of tumour necrosis factor from mast cells is inhibited by sodium cromoglicate.10 There is in-vivo evidence that this drug blocks ICAM-1 expression.11 Furthermore, in vitro, sodium cromoglicate inhibits viral spread between cells.12 Children who regularly inhale nedocromil sodium for asthma have viral illnesses of shorter duration than children who inhale placebo.13 Intranasal nedocromil sodium given before experimental infection with rhinovirus,14 and intranasal sodium cromoglicate started within 24 h of onset of symptoms of upper respiratory tract infection,15 shortened duration of symptoms in adults. On the basis of these efficacy studies, intranasal cromolyn sodium has been identified as a potential treatment for infections of this type.16–20 However, we have not identified studies assessing effectiveness of this treatment in clinical practice. The potential for treatment of infections of the upper respiratory tract by inhibition of ICAM-1 expression warrants pragmatic study.21 Intranasal sodium cromoglicate is safe22 and available without prescription in the UK. We therefore aimed to test the hypothesis that intranasal sodium cromoglicate shortens the duration of suspected acute viral upper respiratory tract infection in children presenting to UK family doctors.

2153

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

Patients and methods Patients Children were eligible for the study if they were between 1 and 12 years of age and if they presented to their family doctor with suspected acute viral infection of the upper respiratory tract, beginning within the previous 7 days. We defined this infection as an acute illness affecting the upper respiratory tract and which, in the clinician’s opinion, was probably caused by a virus. The illness should predominantly affect the upper rather than the lower respiratory tract. We excluded the following children: those with serious presentations (eg, pneumonia or suspected bacterial infections including otitis media, follicular tonsillitis, and sinusitis); those with chronic rhinitis and acute allergic rhinitis (symptoms lasting longer than a week and whose illness did not fit the pattern of an acute, infectious, febrile illness); those judged to need immediate admittance to hospital or antibiotics; those who had used sodium cromoglicate or steroid treatment within the previous week; those whose carers were unable to comply with the study protocol; and those previously recruited to our trial or who had a concurrently participating sibling. Clinicians were allowed to prescribe paracetamol. Study design We asked family doctors in Bro Taf, Lechyd Morgannwg, and Gwent health authorities in south Wales to recruit patients to the study. We gave those who agreed a supply of nose-spray bottles, to be issued to patients who gave consent in numerical order. Each indistinguishable nosespray bottle contained either 4% sodium cromoglicate or normal saline solution. Saline nasal spray has been reported to have no effect on severity or duration of symptoms of the common cold or rhinosinusitis.23 One spray from the sodium cromoglicate bottle delivered about 5·2 mg of study drug. We assigned every spray bottle a unique number by block randomisation with a randomly changing block size. Clinicians, patients, caregivers, and study team were unaware of study randomisation. An on-call study pharmacist was available to break the randomisation code if a serious adverse event arose. We asked family doctors to recruit the first eligible patient who presented in every consulting session to keep selection bias to a minimum. The family doctor explained the study to the caregiver and provided an information sheet. Caregivers who agreed to participate signed a consent form. An ageappropriate information sheet and consent form were available for children judged by the clinician to be able to read and sign it, in addition to the standard consent form. Family doctors gave every caregiver one bottle of nose spray and a study pack, which contained a symptom diary, a second copy of the information sheet, and return envelopes. Family doctors recorded the patient’s contact details, the unique nose-spray number, and presenting symptoms on an encounter sheet, which was faxed to the study team on the same day. Details of presenting symptoms, sex, and reason for non-recruitment were recorded for patients who were eligible but not recruited. On receipt of the encounter sheet, the study nurse telephoned every caregiver to allow further opportunity for questions and to facilitate daily completion of the symptom diary. Caregivers were asked to administer one spray in each nostril every 2 h during the time that the child was awake for the first 2 study days (including day of recruitment) and every 4 h during waking hours for the next 3 days. They were also asked to complete the symptom diary at the end of every study day until the

2154

child was fully recovered (but for no longer than 2 weeks). Symptom diaries and used medication bottles were returned with stamped addressed envelopes that were provided in the study pack. The 18-item Canadian Acute Respiratory Illness and Flu Scale (CARIFS) score was the primary outcome measure and was incorporated into the symptom diary.24 This scale was developed to measure severity of symptoms and effect on daily living for children up to age 12 years with acute respiratory infections. Four of the items in CARIFS relate directly to the upper respiratory tract (eg, nasal congestion and sore throat) and the remainder assess general signs and symptoms of acute infection (eg, irritability and poor appetite). Three items reliant on selfreport (headache, sore throat, and muscle aches and pains) were not applicable to children younger than 2 years, and scoring was adjusted appropriately.24 Minor wording changes were made to three of the items (3, 13, and 15) to be more culturally applicable for a UK setting. Caregivers also used the symptom diary to record frequency of actual use of the nose spray for each of the first 5 study days and overall assessment of child wellbeing (visual analogue scale) for each day the diary was completed. On day 7, the study nurse telephoned caregivers to ask about adverse effects and to complete verbally the CARIFS for that day. Caregivers were reminded to return the nose-spray bottles. After 2 weeks, caregivers were telephoned again and asked about drug use, adverse effects, complications, new but related diagnoses, hospital admissions and further consultations, and their view about which drug their child had been randomly allocated. The local research ethics committees approved the trial. Statistical analysis Although from the outset we planned to use multilevel modelling for our analysis, absence of sufficient published data meant that we based our original power calculation on the comparison of means for two independent groups. From Canadian data,24 we estimated that a target sample size of 250 would be needed to give 90% power to detect an estimated clinically significant difference in mean CARIFS score of 5 (SD 12·1) on day 4 when testing was done at the 5% level of significance. We aimed to recruit 300 children to allow for loss to follow-up and missing data. However, we did a planned, masked interim analysis of the first 85 diaries returned, and compared the curves of the CARIFS scores over the first 7 days for the two groups. From this analysis (a more powerful technique than comparison of scores for 1 day), we estimated that data for 200 people would be needed for comparison of curves of CARIFS scores over 1 week. Furthermore, in this sample we recorded the SD for CARIFS scores on day 4 to be 10·5, suggesting we needed 140 patients for analysis to detect a difference of five in CARIFS scores on this day. Symptom-diary data were entered twice into two separate data files with SPSS version 6. We compared these files with SPSS Data Entry version 2, and discrepancies were resolved. The main analysis was done by intention to treat with the first 7 days of CARIFS data. For patients with partial data, interpolation was used where possible. A two-level model was fitted to the log of the CARIFS score, with repeated measurements of an individual at level one. The model took the form in the panel (A), and it was fitted with serial correlations (B). The model was fitted with iterative generalised least squares in MLwiN with

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

Sodium cromoglicate Normal saline (n=153) (n=137)

Two-level model A

Age (years, mean [SD])

log(CARIFS)=logA–t(␭+␦I0)

Male sex

B

Days ill at presentation (mean [SD])

advanced macros.25 The primary outcome of the study was based on whether ␦ was significantly different from zero, showing a different recovery rate between the two groups.26 Subgroup analysis involved the same model being fitted for each level of these factors: age (<5 years and ⭓5 years); sex; history of asthma and eczema; history of infection of the upper respiratory tract over past 6 months (⭐2 episodes and ⭓3 episodes); duration of symptoms at index consultation (⭐2 days and ⭓3 days); weight of returned spray bottle (<30 g and ⭓30 g); and clinical features at presentation (coryza, cough, enlarged cervical lymph nodes, malaise, pharyngitis, and temperature). We also compared study groups for adverse events and reconsultation with ␹2 test. Recruited and non-recruited children were compared for presenting symptoms and age with the ␹2 and t tests, respectively. Role of the funding source The UK Medical Research Council was represented at meetings of the trial steering committee to ensure that the trial was done in accordance with good clinical practice. However, the sponsor had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results We recruited 125 family doctors from 54 practices to act as clinician investigators, and 55 of these from 33 practices recruited at least one patient into the study. Patients were recruited during a 1-year period beginning in November, 1999. Family doctors faxed 345 encounter sheets of patients to the study team, including 49 for patients who were eligible but not recruited into the trial (figure 1). The most frequent reasons for non-recruitment were patient’s refusal (n=36) and not enough time in the

345 eligible

296 recruited

49 not recruited 36 refused 4 not enough time in consultation 9 other 6 excluded because consent form not received

290 randomly allocated

Symptom Coryza Cough Raised temperature Pharyngitis Enlarged lymph nodes Malaise

137 intranasal normal saline

Figure 1: Trial profile

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

3·5 (2·4) 119 (78%) 126 (82%) 79 (52%) 70 (46%) 71 (46%) 65 (42%)

3·0 (1·9) 114 (83%) 104 (76%) 77 (56%) 73 (53%) 62 (45%) 65 (47%)

Table 1: Characteristics of patients at presentation

consultation (4). 296 patients were recruited, but six of these were excluded because written consent forms were not received. Thus 290 patients were randomly allocated. There was no difference between recruited and nonrecruited children by sex or number of days ill. However, compared with the non-recruited group, the recruited sample reported more symptoms in total (3·5 vs 3·0 respectively, p=0·006), malaise more often (45% vs 29%; p=0·04), and were older (mean age 5·2 [SD 3·39] vs 4·0 [2·57] years; p=0·028). 153 children were randomly allocated intranasal sodium cromoglicate and 137 intranasal normal saline (figure 1). Baseline characteristics of the two study groups were closely similar (table 1). Symptom diaries were returned for 195 (67%) children, week 1 telephone interviews were completed for 246 (85%), and week 2 interviews completed for 247 (85%). Either a diary was returned or week 1 interview completed for 264 (91%) children. Of the 195 children who returned symptom diaries, 20 could not be included in the main multilevel analysis because large amounts of data were missing. Another 15 children had some missing data, but we were able to interpolate sufficiently from the data provided to include them in the main analysis. Therefore, data on 175 children were included in the main analysis. Presenting symptoms, age, duration of illness, and weight of returned nose-spray bottle for children 60

50

40

30

20

10

0

153 intranasal sodium cromoglicate

5·1 (3·4) 62 (45%)

Data are number (%) unless otherwise stated.

CARIFS score

cov(etet–s)=␴e2 exp(–␣s) A=constant; t=time; I=binary variable indicating group; ␭=effect over time; ␦=difference between treatment groups; ␣=estimated as part of the modelling process; s=absolute distance (in time) between measurements.

5·3 (3·4) 75 (49%)

Sodium cromoglicate Figure 2: CARIFS score across week 1

Normal saline

The boxplot shows the median and the IQR (black line and grey box, respectively). Error bars show the largest and smallest values that were not outliers. Circles represent values that are more than 1·5 box-lengths from the 25th and 75th percentiles (outliers). Asterixes show values that are more than 3·0 box-lengths from the 25th and 75th percentiles (extreme values). Circles and asterixes represent individual patients.

2155

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

included in the main analysis (n=175) were compared with variables for those omitted because of nonresponse or missing data (n=115). Children included in the main analysis were older than those excluded (mean age 5·6 [SD 3·53] vs 4·5 [3·05] years; p=0·004) and had been ill for longer (mean duration 3·5 [2·46] vs 2·9 [1·61] days; p=0·026). There were no differences between these two groups in terms of presenting symptoms, weight of returned nose-spray bottle, or caregiver-reported nose-spray use. Figure 2 shows a boxplot of the median (IQR) CARIFS scores for each study group by day of study. The multilevel model was fitted to the log of the CARIFS data (panel). The basic two-level model, with only the constant fitted, resulted in a –2 exp(log [likelihood]) of 3655·15. Adding time—ie, day of CARIFS measurement—with a serial correlation structure greatly improved the fit of the model, with a –2 exp(log [likelihood]) of 2093·18, showing improvement in CARIFS score over the first 7 study days (p<0·0001). Adding treatment effect (sodium cromoglicate or normal saline group) did not improve the fit of the model, with a –2 exp(log [likelihood]) of 2093·12. The estimated difference between the two groups over time was –0·01 (95% CI –0·05 to 0·03), which is not significantly different from zero. This finding suggests that the recovery rates of the children in the two groups did not differ significantly, with the model suggesting a difference between the two groups of 0·2 units of CARIFS on day 4. Furthermore, the two study groups were compared by the day 7 CARIFS score obtained from the week 1 telephone interview. The mean (SD) of log CARIFS score for sodium cromoglicate group was 1·2 (1·10) and 1·4 (1·11) for the normal saline group. These figures were not significantly different (p=0·166). Results of the planned subgroup analyses for age, sex, history of atopy, history of infections of the upper respiratory tract over the past 6 months, duration of symptoms at index consultation, weight of returned medication bottle, and clinical features at presentation showed closely similar non-significant estimates of the treatment effect to the main model. Table 2 provides data for adverse effects obtained from the telephone interviews at 1 and 2 weeks. If the caregiver did not know whether there had been an adverse effect, we treated this response as no report of adverse effect. At week 1, patients given sodium cromoglicate were less likely to report burning and irritation in the nose. For both groups, the most frequently cited problem was that the child did not like the spray. There were no significant differences between study groups in frequency of reported adverse effects at 2 weeks. Sodium cromoglicate* Normal saline† p Adverse effect Week 1 Burning in the nose Irritation Tightness of the chest Nose-bleed Other Week 2 Burning in the nose Irritation Tightness of the chest Nose-bleed Other

11 (8%) 21 (15%) 11 (8%) 6 (4%) 33 (24%)

19 (17%) 31 (28%) 10 (9%) 3 (3%) 33 (30%)

0·04 0·02 0·85 0·45 0·34

3 (2%) 18 (14%) 5 (4%) 6 (5%) 16 (12%)

7 (6%) 13 (11%) 9 (8%) 4 (3%) 7 (6%)

0·12 0·62 0·15 0·72 0·09

*At week 1, n=136; at week 2, 132. †At week 1, n=110; at week 2, 115.

Table 2: Reported adverse effects

2156

During follow-up, an antibiotic was prescribed for 17 of 136 children in the sodium cromoglicate group and for nine of 110 children in the normal saline group; 24 in the sodium cromoglicate group and 19 in the normal saline group went to see their family doctor again. Three children attended a hospital casualty department for possible complication of suspected acute viral upper respiratory tract infection. Of these, two children, both in the sodium cromoglicate group, were admitted and both subsequently made a full recovery. Of the 243 caregivers who responded about whether they had guessed which study drug they had been randomly allocated to, 188 said they did not know, 36 thought they had used normal saline, and 19 thought they had used the experimental nose spray. There was no association between respondents’ guesses and study randomisation.

Discussion The aim of our study was to establish if intranasal sodium cromoglicate would be a useful treatment in children in the UK who go to see their family doctor; children were eligible if they had been ill for up to 1 week. We showed that intranasal sodium cromoglicate provides no additional benefits for children with suspected acute viral infection of the upper respiratory tract. We could not find reports of clinical assessments of the effect of intranasal cromolyn sodium in people with infection for longer than 24 h at the time of starting treatment. A diagnosis of suspected acute viral infection of the upper respiratory tract is likely to include several causes in various stages of the illness. Barrow and colleagues14 started nedocromil sodium before experimentally infecting patients with rhinovirus. Aberg and colleagues15 excluded patients if they had been ill for longer than 24 h. Results of a meta-analysis27 showed that sodium cromoglicate was ineffective for asthma in preschool children in whom viral episodic wheeze is likely to be an important factor. Our trial was done in children because viral upper respiratory tract infection and prescription of antibiotics for this indication are most frequent in this age group.3,28 Some children might have presented when it was too late for sodium cromoglicate to be effective or excessive nasal secretions prevented contact with target molecules. Similarly, caregivers might have struggled to give the nose spray to young children. However, results of subgroup analysis showed that sodium cromoglicate was equally ineffective for children who went to see their family doctor early on in their illness, for children who did not have a runny nose, and for children aged 5 years and older. Some clinicians might have been reluctant to take part in the study because they did not want to further encourage people to seek help for a self-limiting condition. 70 family doctors who did agree to take part did not recruit any patients and the remaining 55 varied considerably in the number of children they recruited. Although information was obtained for a small number of children who were not recruited, there were probably other eligible children who went to see their family doctor during the study period for whom no information was gathered. However, since the randomisation was at the level of the individual child, these considerations are unlikely to have produced a systematic bias between treatment groups.

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

The return rate of the diaries was disappointing at 67%, with caregivers of young children with short duration of illness most likely not to respond. However, there were no differences between responders and nonresponders in presenting symptoms, study group, or weight of returned nose spray bottle. A total of 175 children were included in the analysis of CARIFS score curves over 1 week, a number that is reasonably close to the target of 200 from our interim analysis. The actual difference in CARIFS score on day 4 estimated from the multilevel model was 0·2 CARIFS score units, which is well short of the five units we estimated would be clinically important. The negligible differences between groups in this sample suggest that 175 children were sufficient for the analysis. The absence of difference between the groups in the primary analysis, based on the diary data over 7 days, accorded with the comparison with CARIFS score data from 246 children obtained by telephone on day 7. We successfully followed up 91% of children either by diary or telephone interview. From a statistical point of view, the multilevel model (panel) allowed us to use the information about the course of the illness, incorporating all the data provided by the repeated measures design. This approach is more powerful than use of a simple summary statistic, and has enabled us to model the within-child correlation structure, rather than assuming equal correlation across time points (compound symmetry). Taken together, these findings suggest that the chances of a type II error are remote and therefore that our study is highly powered. Although we have shown that sodium cromoglicate is not useful for children going to see their family doctor with suspected acute viral infection of the upper respiratory tract, we have also shown that this illness was managed without prescription of antibiotics for 220 (89%) children. These infections can be managed safely in primary care, since during 2-week follow-up after the initial consultation, only three children attended a hospital emergency service for symptoms possibly related to the illness. Of these, two were admitted and all made a full recovery. It is noteworthy that 43 (17%) children went back to see their family doctor. This proportion is similar to that in a study of winter upper respiratory tract infection in south Wales.29 This finding suggests that clinicians are still not identifying effectively those children at the initial consultation who are likely to have a lengthy illness. These children might have bacterial rather than viral infections, and thus could benefit from antibiotic treatment or alternatively more effective communication interventions about the self-limiting, but possibly extended, nature of the illness. Contributors C Butler was responsible for study design and protocol development, acted as principal investigator, and contributed to analysis, appraisal, and report writing. M Robling was the trial manager and oversaw data entry, data cleaning, and contributed to implementation, analysis, appraisal, and report writing. H Prout was the study nurse who liaised with practices, was responsible for follow-up of patients, and contributed to report writing. K Hood was the trial statistician and contributed to implementation, appraisal, and report writing. P Kinnersley was co-investigator and contributed to study design, implementation, analysis, appraisal, and report writing. I Russell contributed to protocol development and statistical analysis. N Stott contributed to protocol development, appraisal, and report writing. H Houston contributed to study implementation, interpretation, and report writing.

Acknowledgments The UK Medical Research Council funded the study (grant number MRC G9900236). C Butler was supported by a fellowship from NHS Wales Research and Development for Health and Social Care. Sheila Morris was the study secretary. Independent members of the trial steering committee were Tom Fahey (chair), Paul Little, and Prof Max Bachmann, and all made an invaluable contribution to the successful implementation of the study. Prof Julian Hopkin, Norman Vetter, and Robert Newcombe formed the data monitoring and ethics committee. We thank the Department of General Practice of the University of Wales College of Medicine for support; the family doctors who recruited children into this study; and the caregivers and children for taking part.

References 1

2 3

4

5

6

7

8

9

10

11

12 13

14

15

16 17 18 19 20

21 22 23

Conflict of interest statement None declared.

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

Butler CC, Rollnick S, Kinnersley P, Jones A, Stott NCH. Reducing antibiotics for respiratory tract symptoms in primary care; consolidating ‘why’ and considering ‘how’. Br J Gen Pract 1998; 48: 1865–70. Lowenstein SR, Parrino TA. Management of the common cold. Adv Intern Med 1987; 32: 207–34. Breese Hall C, McBride JT. Upper respiratory tract infections: the common cold, pharyngitis, croup, bacterial tracheitis and epiglottitis. In: Pennington JE, ed. Respiratory infections: diagnosis and management. New York: Raven Press, 1994: 101–24. Wald ER, Guerra N, Byers C. Upper respiratory tract infections in young children: duration of and frequency of complications. Pediatrics 1991; 87: 129–33. Johnston SL, Pattemore PK, Sanderson G, et al. Community study of role of viral infections in exacerbations of asthma in 9 to 11 year old children. BMJ 1995; 310: 1225–28. Wegner CD, Gundel RH, Reilly P, Hynes N, Letts G, Rothelein R. Intercellular adhesion molecule (ICAM-1) in the pathogenesis of asthma. Science 1990; 247: 456–59. Dustin ML, Singer KH, Tuck DT, Springer TA. Adhesion of T lymphocytes to epidermal keratinocytes is regulated by IFN-gamma and ICAM-alpha. J Exp Med 1988; 167: 1323–40. Balfour-Lynn IM, Valman HB, Wellings R. Tumour necrosis factoralpha and leukotriene E4 production in wheezy infants. Clin Exp Allergy 1994; 24: 121–26. Staunton DE, Merluzze VJ, Rohelein R. A cell adhesion molecule, ICAM-1, is the major surface receptor for rhinoviruses. Cell 1989; 56: 848–53. Bissonette EY, Befus AD. Modulation of mast cell function in the gastrointestinal tract. In: Immunopharmacology of the gastrointestinal system. London: Academic Press, 1993: 95–103. Hoshino M, Nakamura Y. The effect on inhaled sodium cromoglycate on cellular infiltration into the bronchial mucosa and the expression of adhesion molecules in asthmatics. Eur Respir J 1997; 10: 858–65. Penttinen K, Aarnio A, Hovi T. Disodium cromoglycate can inhibit virus-induced cytopathic effects in vitro. BMJ 1977; 1: 82. Konig P, Eigen H, Ellis MH. The effect of nedocromil sodium on childhood asthma during the viral season. Respir Crit Care 1995; 152: 1879–86. Barrow GI, Higgins W, Al-Nakib W, Smith AP, Wenham RBM, Tyrrell DAJ. The effect of intranasal nedocromil sodium on viral upper respiratory tract infections in human volunteers. Clin Exp Allergy 1990; 20: 45–51. Aberg N, Aberg B, Alestig K. The effect of inhaled and intransal sodium cromoglycate on symptoms of upper respiratory tract infections. Clin Exp Allergy 1996; 26: 1045–50. Mossad SB. Treatment of the common cold. BMJ 1998; 317: 33–36. Johnson SL. Cromolyns: treatment for the common cold? Clin Exp Allergy 1996; 26: 989–94. Johnston SL. Problems and prospects of developing effective therapy for common cold viruses. Trends Microbiol 1997; 5: 58–63. Friedman ND, Sexton DJ. The common cold. In: UpToDate, Online version 9.3. Wellesley: UPTODATE, 2001: 1–9. Hay CM. Treatment and prevention of rhinovirus infections. UpToDate, Online version 9.3. Wellesley: UPTODATE, 2001: 1–7. Stanciu LA, Djukanovic R. The role of ICAM-1 on T-cells in the pathogenesis of asthma. Eur Respir J 1998; 11: 949–57. Kuzemko J. Twenty years of sodium cromoglycate treatment: a short review. Respir Med 1989; 83: 11–16. Adam P, Stiffman M, Blake R. A clinical trial of hypertonic saline nasal spray in subjects with the common cold or rhinosinusitis. Arch Fam Med 1998; 7: 39–43.

2157

For personal use. Only reproduce with permission from The Lancet Publishing Group.

ARTICLES

24 Jacobs B, Young NL, Dick PY, et al. Canadian Acute Respiratory Illness and Flu Scale (CARIFS): development of a valid measure for childhood respiratory infections. J Clin Epidemiol 2000; 53: 793–99. 25 Yang M, Rasbash J, Goldstein H, Barbosa M. MLwiN macros for advanced multilevel modelling, version 2.0. London: Multilevels Models Project, Institute of Education, University of London, 1999. 26 Goldstein H, Healy M, Rasbash J. Multilevel time series models with applications to repeated measures data. Stat Med 1994; 13: 1643–55.

27 Tasche MJA, Uijen JHJM, Bernsen RMD, Jongste JC, van der Wouden JC. Inhaled disodium cromoglycate (DSCG) as maintenance therapy in children with asthma: a systematic review. Thorax 2000; 55: 913–20. 28 Majeed A, Moser K. Age- and sex-specific antibiotic prescribing in general practice in England and Wales in 1996. Br J Gen Pract 1999; 49: 735–36. 29 Stott NCH. Management and outcome of winter upper respiratory tract infections in children aged 0-9 years. BMJ 1979; 1: 29–31.

Clinical picture Pulmonary tumour embolism Jong-Won Ha, Se-Kyu Kim, Byung-Chul Chang A 55-year-old woman presented with an 11-month history of dyspnoea on exertion that had been recently aggravated. 4 months earlier, she had been diagnosed as having a pulmonary embolism, for which she was on anticoagulation with warfarin. Computed tomography of the thorax show a large embolus occupying the lumen of the left main and right main pulmonary arteries (arrows in figure A). Digital subtraction angiography showed total occlusion of the left pulmonary artery and luminal narrowing of the right upper

lobar pulmonary artery due to embolus (arrows in figure B). The patient underwent surgery to remove the pulmonary emboli. At operation, large amounts of thrombus-like material were removed from both pulmonary arteries (figure C). Microscopy showed fragments of poorly differentiated squamous carcinoma mixed with necrotic debri and coagulative fibrin (figure D). Extensive search for the primary focus of malignancy was unrevealing. The patient died 16 months after surgery.

Department of Internal Medicine (S K Kim MD, B C Chang MD) and Cardiothoracic Surgery (J W Han MD) Yonsei University College of Medicine, Seoul, Korea

2158

THE LANCET • Vol 359 • June 22, 2002 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.