- Email: [email protected]
Treatment of glue ear in general practice
THE LANCET 1
2
3
4
5
Sandham JD, Hull RD, Brant RF. Pulmonary artery flow directed catheter: the evidence. Lancet 1996; 348: 1324. Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of right heart cathterization in the initial care of critically ill patients. JAMA 1996; 276: 889–97. Gore JM, Golberg RJ, Spodick DH, Alpert S, Dalen JE. A community-wide assessment of the use of pulmonary artery catheters in patients with acute myocardial infarction. Chest 1987; 92: 721–31. Naylor, CD, Sibbald WJ, Sprung CL, Pinfold SP, Calvin JE, Cerra FB. Pulmonary artery catheterization: can there be an integrated strategy for guideline development and research promotion? JAMA 1993; 269: 2407–11. Guyatt G. A randomized controlled trial of right-heart catheterization in critically ill patients. J Intens Care Med 1991; 6: 91–95.
SIR—Sandham and colleagues1 state that “the use of pulmonary artery [flowdirected] catheters has been an integral part of the practice in intensive-care units since shortly after the introduction of the technique in the early 1970s by Swan and Ganz”. It is a popular misconception that Swan and Ganz were the originators of the notion of flow-directed pulmonary artery catheterisation to measure pressures and cardiac output. They publicised the value of the technique but their intellectual contribution was minor. They combined in one catheter the ideas of two other groups. Bradley2 was first to use light and flexible flow-directed catheters in man, through which right heart pressures could be measured and used to position the catheter-tip without the need for fluoroscopy. Thus he did bedside monitoring of pulmonary artery pressure in the critically ill for diagnosis and for monitoring therapeutic manoeuvres. With Branthwaite,3 he was first to perform thermodilution cardiac output measurements in healthy volunteers and critically ill patients with a flow-directed thermistor-tipped catheter positioned in the pulmonary artery by observation of lumenal pressure. In the same report they first described internal jugular vein seldinger puncture to position such catheters, for continuous central venous pressure monitoring and for long-term intravenous infusion to overcome difficulties experienced with peripheral lines. Recognised at the time,4 their work is now frequently unacknowledged as the origin of the haemodynamic monitoring that is common in intensive care units. Swan visited Bradley and Branthwaite when they were doing this research. With colleagues, Swan added to the system devised by Bradley and Branthwaite a balloon to aid flotation and enable wedging of the catheter, as previously described in dogs by Lategola and Rahn.5 Bradley was awarded the first chair of
Vol 349 • January 11, 1997
intensive care in England. He went on to devise an algorithm for non-invasive calculation of cardiac output and assessment of right and left ventricular function. His technique has never caught on; presumably because clinicians would rather rely on a blackbox assessment of output based on one of his techniques than learn and trust a new clinical skill based on another. He taught that pulmonary artery catheters should be used sparingly to manage specific problems and not as a substitute for proper clinical assessment. He scorned goaldirected therapy in which the circulation was adjusted to produce supranormal oxygen delivery, and has been proved correct. The call for a moratorium on use of pulmonary artery catheters would be misdirected if high mortality associated with the use of such catheters is the result of application of a flawed therapeutic strategy rather than complications from the catheters. If a similar study looking at the safety of 24-h ambulatory electrocardiogram monitoring had been done in the USA at a time when it was the fashion to treat non-sustained ventricular ectopic arrhythmias in postinfarct patients with anti-arrhythmic drugs would there be calls for a moratorium on that test too? Peter Wilmshurst Royal Shrewsbury Hospital, Shrewsbury SY3 8X Q, UK
1
2
3
4
5
Sandham JD, Hull RD, Brant RF. Pulmonary artery flow directed catheter: the evidence. Lancet 1996; 348: 1324. Bradley RD. Diagnostic right-heart catheterisation with miniature catheters in severely ill patients. Lancet 1964; ii: 941–42. Branthwaite MA, Bradley RD. Measurement of cardiac output by thermal dilution in man. J Appl Physiol 1968; 24: 434–38. Scheinman MM, Abbott JA, Rapaport E. Clinical uses of a flow-directed right heart catheter. Ann Intern Med 1969; 124: 19–24. Lategola M, Rahn H. A self-guiding catheter for cardiac and pulmonary arterial catheterization and occlusion. Proc Soc Exp Biol Med 1953; 84: 667–68.
Pathophysiology of itching SIR—Your Nov 30 correspondent’s1 comment on our report2 is correct that the activation of opioid-dependent pathways may evoke itching, but suppress pain. Our allusion to both activation of inhibitory circuits and regulation of quality and intensity of perceived itch was intended (but evidently failed) to convey this paradox. The study referred to by Shuster3 emphasised the importance of the sedative action of classic H1antihistamines in allaying the itch of atopic eczema, but several drugs were
tested in only a small group of eczematous dermatoses, and others,4 in placebo-controlled experiments, have reported low-sedation antihistamines to be effective. Perhaps patients with more acute inflammatory eczema involving histamine release5 respond better to the newer minimally sedative antihistamines. We agree with Gillman and colleagues that ␦-receptors may also be important in the role of the opioid system in itching. We are now studying the peripheral and central actions of ␦opioid agonists. We are also grateful to McIntosh for pointing out our wanton use of the term tachykinin. *Malcolm W Greaves, P D Wall United Medical and Dental Schools, Guy’s and St Thomas’s Hospitals, London SE1 7EH, UK
1 2 3
4
5
Bernhard JD. Pathophysiology of itching. Lancet 1996; 348: 1513–14. Greaves MW, Wall PD. Pathophysiology of itching. Lancet 1996; 348: 938–40. Krause L, Shuster S. Mechanism of action of antipruritic drugs. BMJ 1983; 287: 1199–1200. Doherty V, Sylvester DGH, Kennedy CTC, et al. Double blind trial of acrivastine, terfenadine or placebo in 49 patients with atopic dermatitis showed improvements with both but no difference between the active agents. BMJ 1989; 298: 96. Sampson HA, Broadbent KR, BernhiselBroadbent J. Spontaneous release of histamine from basophils and histaminereleasing factor in patients with atopic dermatitis and food hypersensitivity. N Engl J Med 1989; 321: 228–32.
Treatment of glue ear in general practice SIR—Van Balen and colleagues (Sept 14, p 713)1 report the results of a placebo-controlled randomised trial in 162 children with bilateral otitis media with effusion (OME, or glue ear). They conclude that 2 weeks of antibiotic therapy with co-amoxiclav is effective. On an intent-to-treat analysis, and, for the primary endpoint at 2 weeks, the numbers of children with OME were 65 of 80 (81%) in the placebo group and 61 of 82 (74%) in the drug group. The cure-rate difference of 6·8% is neither clinically nor statistically significant and is close to the 4·8% we reported 5 years ago among 518 children treated with 2 weeks of amoxycillin or placebo for OME.2 In van Balen’s trial, ten children in the placebo group had no primary endpoint, by contrast with only three in the drug group (p=0·025), a condition not explainable by chance. Assumptions about missing data therefore are critical to the interpretation. Unless all those children for whom the data are missing had uncured OME the outcome of this
133
THE LANCET
trial is negative (no difference). In advocating antibiotic therapy Van Balen et al do not cite two metaanalyses which showed that antibiotics were ineffective in OME,3,4 and also fail to mention a randomised trial in 310 children in which there were no differences in cure rates between the placebo and the three antibiotictreated groups (cure rates at 4 weeks were 27% and 25–33%, respectively).5 Also not mentioned is the significant increased recurrence rate associated with antibiotic treatment.2,5
the clinical benefits, which would justify this intervention. Ann Bisset Grampian Health Board, Summerfield House, Aberdeen AB15 6RE, UK
1
2
Van Balen FAM, De Melker RA, Touww-Otten FWMM. Double-blinded randomised trial of co-amoxiclav versus placebo for persistent otitis media with effusion in general practice. Lancet 1996; 348: 713–16. Maw AR. Using tympanometry to detect glue ear in general practice. BMJ 1992; 304: 67–68.
Erdem I Cantekin
Authors’ reply
Department of Otolaryngology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA
SIR—We agree with Cantekin that information for the primary endpoints are critical to the interpretation in all randomised trials. The way we presented the results of our trial (first we provided the intention-to-treat finding in all participants with available endpoints and subsequently discussed the possible effect of missing values on the outcome measure) is appropriate. We also performed a (worse-case) sensitivity analysis, mentioned in the discussion. There is no reason at all, however, to assume that the nine children who were lost to follow-up (three in the drug group and six in the placebo group) and the four children with uninterpretable tympanograms (all in the placebo group) in our study would all have had uncured OME. The fact that more children in the placebo group had no assessment of the primary endpoint than children in the drug therapy group is coincidence, and is unrelated to therapy preference because treatment allocation was random and our trial was done in a double-blind fashion with no early looks. We further fail to see how the mention of the two meta-analyses and the randomised trial would have altered our results or conclusions. Our study proved that antibiotic treatment has a short-term effect in a general practice population of children with persistent bilateral OME. However, we emphasise that we do not advise antibiotic treatment for all children with persistent bilateral OME. We clearly state that: “the decision whether or not to prescribe antibiotics in children with persistent OME should depend on the individual child and the possible sequelae in that child”. If Cantekin is trying to tell us to be careful in prescribing antibiotics in patients with persistent OME we fully agree. However, our study pointed out that antibiotic therapy should be considered in certain children with persistent OME. With respect to Bisset, our results can only be generalised to children with bilateral OME because only these children were included in our trial. The sequelae of OME include recurrent episodes with acute otitis media,
1
2
3 4
5
Van Balen FAM, de Melker RA, Touw-Otten FWMM. Double-blind randomised trial of co-amoxiclav versus placebo for persistent otitis media with effusion in general practice. Lancet 1996; 348: 713–16. Cantekin EI, McGuire TW, Griffith TL. Antimicrobial therapy for otitis media with effusion (‘secretory’ otitis media). JAMA 1991; 266: 3309–17. Bailar JC. The practice of meta-analysis. J Clin Epidemiol 1995; 48: 149–57. Cantekin EI. Antibiotics to prevent acute otitis media and to treat otitis media with effusion. JAMA 1994; 272: 203–04. Mandel EN, Rockette HE, Paradise JL, Bluestone CD, Nozza RJ. Comparative efficacy of erythromycin-sulfisoxazole, cefaclor, amoxicillin or placebo for otitis media with effusion in children. Pediatr Infect Dis J 1991; 10: 899–906.
SIR—Rigorous randomised control trial evidence in otitis media is very much to be encouraged. However, the results of van Balen and colleagues’ trial1 should be interpreted with caution. First, these investigators use tympanograms as an outcome measure. Although tympanograms are sensitive and specific2 they may lead to over-referral of mild cases. Resolution of symptoms (if any) would clearly be of greater clinical interest. Unilateral OME is likely to be less detrimental to speech and language development than bilateral OME. Is treatment of unilateral OME justified? Second, the follow-up time is short at 2 weeks. To justify widespread use of antibiotics, follow-up at 3 months or even longer would be of interest, especially in view of the fluctuating nature of OME. Third, why did general practitioners continue to treat with antibiotics after the trial period? Did the trial tend to encourage irrational prescribing? For example, decongestant drops were included in the trial although they are known to be ineffective. In population terms, widespread use of antibiotics is likely to be expensive, and runs the risk of increasing antibiotic resistance and treatment of children who do not have a clinical problem. More research is needed on
134
structural damage to the eardrum, and behavioural problems.1–3 These sequelae also occur in unilateral OME and therefore we believe that these children need careful follow-up and could also benefit from antibiotic treatment. Moreover, OME is usually not restricted to one ear, and often occurs alternately in both ears. The fluctuating nature of OME is an important fact, that tends to affect the magnitude of the observed effect in trials. Nevertheless, we did record a relevant difference between the two arms of the study. It would certainly be of interest to extend the follow-up period beyond 3 months. We could not do so mainly because the Dutch guidelines on OME advise referral to a surgeon after 3 months, and the dropout rate in our study would therefore have increased considerably. At the follow-up visit general practitioners were given three possible management decisions: (1) to refer the patient to a surgeon (according to the Dutch guidelines); (2) to extend the follow-up period for 3 more months without treatment; or (3) with antibiotic treatment. These options were included because Thomsen et al4 found a longer lasting effect when antibiotics were given for 30 days. The reason for giving decongestant nosedrops were described in the method section. In our discussion we mentioned that our findings to not imply that antibiotics should be given to all children with persistent OME. We agree that more research is needed to assess which children are likely to benefit most from antibiotic treatment. *F A M van Balen, R A de Melker Department of General Practice/Family Medicine, Universiteitsweg 100, 3584 CG Utrecht, Netherlands
1
2
3 4
Teele DW, Klein JO, Rosner BA. Otitis media with effusion during the first three years of life and development of speech and language. Pediatrics 1984; 74: 282–87. Haggard MP, Birkin JA, Browning GG, Gatehouse S, Lewis S. Behavior problems in otitis media. Pediatr Infect Dis 1994; 13: S43–50. de Melker RA. Treating persistent glue ear in children. BMJ 1993; 306: 5–6. Thomsen J, Sederberg-Olsen J, Balle V, Vejlsgaard R, Stangerup SE, Bondessen G. Antibiotic treatment of children with secretory otitis media. Arch Otolaryngol Head Neck Surg 1989; 115: 447–51.
Sudden unexplained death SIR—Sudden unexplained death in epilepsy, upon which Nashef and Brown comment (Nov 16, p 1324),1 may well have a link to the excessive sympathomimetic activity seen in acute stroke and subarachnoid haemorrhage.2 We have found that the increased plasma catecholamine concentrations in
Vol 349 • January 11, 1997
2
3
4
5
Sandham JD, Hull RD, Brant RF. Pulmonary artery flow directed catheter: the evidence. Lancet 1996; 348: 1324. Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of right heart cathterization in the initial care of critically ill patients. JAMA 1996; 276: 889–97. Gore JM, Golberg RJ, Spodick DH, Alpert S, Dalen JE. A community-wide assessment of the use of pulmonary artery catheters in patients with acute myocardial infarction. Chest 1987; 92: 721–31. Naylor, CD, Sibbald WJ, Sprung CL, Pinfold SP, Calvin JE, Cerra FB. Pulmonary artery catheterization: can there be an integrated strategy for guideline development and research promotion? JAMA 1993; 269: 2407–11. Guyatt G. A randomized controlled trial of right-heart catheterization in critically ill patients. J Intens Care Med 1991; 6: 91–95.
SIR—Sandham and colleagues1 state that “the use of pulmonary artery [flowdirected] catheters has been an integral part of the practice in intensive-care units since shortly after the introduction of the technique in the early 1970s by Swan and Ganz”. It is a popular misconception that Swan and Ganz were the originators of the notion of flow-directed pulmonary artery catheterisation to measure pressures and cardiac output. They publicised the value of the technique but their intellectual contribution was minor. They combined in one catheter the ideas of two other groups. Bradley2 was first to use light and flexible flow-directed catheters in man, through which right heart pressures could be measured and used to position the catheter-tip without the need for fluoroscopy. Thus he did bedside monitoring of pulmonary artery pressure in the critically ill for diagnosis and for monitoring therapeutic manoeuvres. With Branthwaite,3 he was first to perform thermodilution cardiac output measurements in healthy volunteers and critically ill patients with a flow-directed thermistor-tipped catheter positioned in the pulmonary artery by observation of lumenal pressure. In the same report they first described internal jugular vein seldinger puncture to position such catheters, for continuous central venous pressure monitoring and for long-term intravenous infusion to overcome difficulties experienced with peripheral lines. Recognised at the time,4 their work is now frequently unacknowledged as the origin of the haemodynamic monitoring that is common in intensive care units. Swan visited Bradley and Branthwaite when they were doing this research. With colleagues, Swan added to the system devised by Bradley and Branthwaite a balloon to aid flotation and enable wedging of the catheter, as previously described in dogs by Lategola and Rahn.5 Bradley was awarded the first chair of
Vol 349 • January 11, 1997
intensive care in England. He went on to devise an algorithm for non-invasive calculation of cardiac output and assessment of right and left ventricular function. His technique has never caught on; presumably because clinicians would rather rely on a blackbox assessment of output based on one of his techniques than learn and trust a new clinical skill based on another. He taught that pulmonary artery catheters should be used sparingly to manage specific problems and not as a substitute for proper clinical assessment. He scorned goaldirected therapy in which the circulation was adjusted to produce supranormal oxygen delivery, and has been proved correct. The call for a moratorium on use of pulmonary artery catheters would be misdirected if high mortality associated with the use of such catheters is the result of application of a flawed therapeutic strategy rather than complications from the catheters. If a similar study looking at the safety of 24-h ambulatory electrocardiogram monitoring had been done in the USA at a time when it was the fashion to treat non-sustained ventricular ectopic arrhythmias in postinfarct patients with anti-arrhythmic drugs would there be calls for a moratorium on that test too? Peter Wilmshurst Royal Shrewsbury Hospital, Shrewsbury SY3 8X Q, UK
1
2
3
4
5
Sandham JD, Hull RD, Brant RF. Pulmonary artery flow directed catheter: the evidence. Lancet 1996; 348: 1324. Bradley RD. Diagnostic right-heart catheterisation with miniature catheters in severely ill patients. Lancet 1964; ii: 941–42. Branthwaite MA, Bradley RD. Measurement of cardiac output by thermal dilution in man. J Appl Physiol 1968; 24: 434–38. Scheinman MM, Abbott JA, Rapaport E. Clinical uses of a flow-directed right heart catheter. Ann Intern Med 1969; 124: 19–24. Lategola M, Rahn H. A self-guiding catheter for cardiac and pulmonary arterial catheterization and occlusion. Proc Soc Exp Biol Med 1953; 84: 667–68.
Pathophysiology of itching SIR—Your Nov 30 correspondent’s1 comment on our report2 is correct that the activation of opioid-dependent pathways may evoke itching, but suppress pain. Our allusion to both activation of inhibitory circuits and regulation of quality and intensity of perceived itch was intended (but evidently failed) to convey this paradox. The study referred to by Shuster3 emphasised the importance of the sedative action of classic H1antihistamines in allaying the itch of atopic eczema, but several drugs were
tested in only a small group of eczematous dermatoses, and others,4 in placebo-controlled experiments, have reported low-sedation antihistamines to be effective. Perhaps patients with more acute inflammatory eczema involving histamine release5 respond better to the newer minimally sedative antihistamines. We agree with Gillman and colleagues that ␦-receptors may also be important in the role of the opioid system in itching. We are now studying the peripheral and central actions of ␦opioid agonists. We are also grateful to McIntosh for pointing out our wanton use of the term tachykinin. *Malcolm W Greaves, P D Wall United Medical and Dental Schools, Guy’s and St Thomas’s Hospitals, London SE1 7EH, UK
1 2 3
4
5
Bernhard JD. Pathophysiology of itching. Lancet 1996; 348: 1513–14. Greaves MW, Wall PD. Pathophysiology of itching. Lancet 1996; 348: 938–40. Krause L, Shuster S. Mechanism of action of antipruritic drugs. BMJ 1983; 287: 1199–1200. Doherty V, Sylvester DGH, Kennedy CTC, et al. Double blind trial of acrivastine, terfenadine or placebo in 49 patients with atopic dermatitis showed improvements with both but no difference between the active agents. BMJ 1989; 298: 96. Sampson HA, Broadbent KR, BernhiselBroadbent J. Spontaneous release of histamine from basophils and histaminereleasing factor in patients with atopic dermatitis and food hypersensitivity. N Engl J Med 1989; 321: 228–32.
Treatment of glue ear in general practice SIR—Van Balen and colleagues (Sept 14, p 713)1 report the results of a placebo-controlled randomised trial in 162 children with bilateral otitis media with effusion (OME, or glue ear). They conclude that 2 weeks of antibiotic therapy with co-amoxiclav is effective. On an intent-to-treat analysis, and, for the primary endpoint at 2 weeks, the numbers of children with OME were 65 of 80 (81%) in the placebo group and 61 of 82 (74%) in the drug group. The cure-rate difference of 6·8% is neither clinically nor statistically significant and is close to the 4·8% we reported 5 years ago among 518 children treated with 2 weeks of amoxycillin or placebo for OME.2 In van Balen’s trial, ten children in the placebo group had no primary endpoint, by contrast with only three in the drug group (p=0·025), a condition not explainable by chance. Assumptions about missing data therefore are critical to the interpretation. Unless all those children for whom the data are missing had uncured OME the outcome of this
133
THE LANCET
trial is negative (no difference). In advocating antibiotic therapy Van Balen et al do not cite two metaanalyses which showed that antibiotics were ineffective in OME,3,4 and also fail to mention a randomised trial in 310 children in which there were no differences in cure rates between the placebo and the three antibiotictreated groups (cure rates at 4 weeks were 27% and 25–33%, respectively).5 Also not mentioned is the significant increased recurrence rate associated with antibiotic treatment.2,5
the clinical benefits, which would justify this intervention. Ann Bisset Grampian Health Board, Summerfield House, Aberdeen AB15 6RE, UK
1
2
Van Balen FAM, De Melker RA, Touww-Otten FWMM. Double-blinded randomised trial of co-amoxiclav versus placebo for persistent otitis media with effusion in general practice. Lancet 1996; 348: 713–16. Maw AR. Using tympanometry to detect glue ear in general practice. BMJ 1992; 304: 67–68.
Erdem I Cantekin
Authors’ reply
Department of Otolaryngology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA
SIR—We agree with Cantekin that information for the primary endpoints are critical to the interpretation in all randomised trials. The way we presented the results of our trial (first we provided the intention-to-treat finding in all participants with available endpoints and subsequently discussed the possible effect of missing values on the outcome measure) is appropriate. We also performed a (worse-case) sensitivity analysis, mentioned in the discussion. There is no reason at all, however, to assume that the nine children who were lost to follow-up (three in the drug group and six in the placebo group) and the four children with uninterpretable tympanograms (all in the placebo group) in our study would all have had uncured OME. The fact that more children in the placebo group had no assessment of the primary endpoint than children in the drug therapy group is coincidence, and is unrelated to therapy preference because treatment allocation was random and our trial was done in a double-blind fashion with no early looks. We further fail to see how the mention of the two meta-analyses and the randomised trial would have altered our results or conclusions. Our study proved that antibiotic treatment has a short-term effect in a general practice population of children with persistent bilateral OME. However, we emphasise that we do not advise antibiotic treatment for all children with persistent bilateral OME. We clearly state that: “the decision whether or not to prescribe antibiotics in children with persistent OME should depend on the individual child and the possible sequelae in that child”. If Cantekin is trying to tell us to be careful in prescribing antibiotics in patients with persistent OME we fully agree. However, our study pointed out that antibiotic therapy should be considered in certain children with persistent OME. With respect to Bisset, our results can only be generalised to children with bilateral OME because only these children were included in our trial. The sequelae of OME include recurrent episodes with acute otitis media,
1
2
3 4
5
Van Balen FAM, de Melker RA, Touw-Otten FWMM. Double-blind randomised trial of co-amoxiclav versus placebo for persistent otitis media with effusion in general practice. Lancet 1996; 348: 713–16. Cantekin EI, McGuire TW, Griffith TL. Antimicrobial therapy for otitis media with effusion (‘secretory’ otitis media). JAMA 1991; 266: 3309–17. Bailar JC. The practice of meta-analysis. J Clin Epidemiol 1995; 48: 149–57. Cantekin EI. Antibiotics to prevent acute otitis media and to treat otitis media with effusion. JAMA 1994; 272: 203–04. Mandel EN, Rockette HE, Paradise JL, Bluestone CD, Nozza RJ. Comparative efficacy of erythromycin-sulfisoxazole, cefaclor, amoxicillin or placebo for otitis media with effusion in children. Pediatr Infect Dis J 1991; 10: 899–906.
SIR—Rigorous randomised control trial evidence in otitis media is very much to be encouraged. However, the results of van Balen and colleagues’ trial1 should be interpreted with caution. First, these investigators use tympanograms as an outcome measure. Although tympanograms are sensitive and specific2 they may lead to over-referral of mild cases. Resolution of symptoms (if any) would clearly be of greater clinical interest. Unilateral OME is likely to be less detrimental to speech and language development than bilateral OME. Is treatment of unilateral OME justified? Second, the follow-up time is short at 2 weeks. To justify widespread use of antibiotics, follow-up at 3 months or even longer would be of interest, especially in view of the fluctuating nature of OME. Third, why did general practitioners continue to treat with antibiotics after the trial period? Did the trial tend to encourage irrational prescribing? For example, decongestant drops were included in the trial although they are known to be ineffective. In population terms, widespread use of antibiotics is likely to be expensive, and runs the risk of increasing antibiotic resistance and treatment of children who do not have a clinical problem. More research is needed on
134
structural damage to the eardrum, and behavioural problems.1–3 These sequelae also occur in unilateral OME and therefore we believe that these children need careful follow-up and could also benefit from antibiotic treatment. Moreover, OME is usually not restricted to one ear, and often occurs alternately in both ears. The fluctuating nature of OME is an important fact, that tends to affect the magnitude of the observed effect in trials. Nevertheless, we did record a relevant difference between the two arms of the study. It would certainly be of interest to extend the follow-up period beyond 3 months. We could not do so mainly because the Dutch guidelines on OME advise referral to a surgeon after 3 months, and the dropout rate in our study would therefore have increased considerably. At the follow-up visit general practitioners were given three possible management decisions: (1) to refer the patient to a surgeon (according to the Dutch guidelines); (2) to extend the follow-up period for 3 more months without treatment; or (3) with antibiotic treatment. These options were included because Thomsen et al4 found a longer lasting effect when antibiotics were given for 30 days. The reason for giving decongestant nosedrops were described in the method section. In our discussion we mentioned that our findings to not imply that antibiotics should be given to all children with persistent OME. We agree that more research is needed to assess which children are likely to benefit most from antibiotic treatment. *F A M van Balen, R A de Melker Department of General Practice/Family Medicine, Universiteitsweg 100, 3584 CG Utrecht, Netherlands
1
2
3 4
Teele DW, Klein JO, Rosner BA. Otitis media with effusion during the first three years of life and development of speech and language. Pediatrics 1984; 74: 282–87. Haggard MP, Birkin JA, Browning GG, Gatehouse S, Lewis S. Behavior problems in otitis media. Pediatr Infect Dis 1994; 13: S43–50. de Melker RA. Treating persistent glue ear in children. BMJ 1993; 306: 5–6. Thomsen J, Sederberg-Olsen J, Balle V, Vejlsgaard R, Stangerup SE, Bondessen G. Antibiotic treatment of children with secretory otitis media. Arch Otolaryngol Head Neck Surg 1989; 115: 447–51.
Sudden unexplained death SIR—Sudden unexplained death in epilepsy, upon which Nashef and Brown comment (Nov 16, p 1324),1 may well have a link to the excessive sympathomimetic activity seen in acute stroke and subarachnoid haemorrhage.2 We have found that the increased plasma catecholamine concentrations in
Vol 349 • January 11, 1997