Research
Effect of systemic steroids on post-tonsillectomy bleeding and reinterventions: systematic review and meta-analysis of randomised controlled trials
Cite this as:
BMJ
2012;345:e5389
- Our online table of contents is updated at least twice each day. Read all articles published in the last 7 days.
- You can use bmj.com to help you with your continuing medical education. Find out about CME/CPD credits for BMJ articles
- Keep up to date with cardiology: Access the latest cardiovascular medicine resources from across BMJ Group.
-
OPEN ACCESS: All research articles are freely available online, with no word limit. Find out more about the BMJ's open access policy. Submit your paper. - Find out how study types differ in our How to read a paper section.
Recent rapid responses
Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on bmj.com. Although a selection of rapid responses will be included as edited readers' letters in the weekly print issue of the BMJ, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window.
Displaying 1-4 out of 4 published
16 October 2012
In a systematic review and meta-analysis of randomised controlled trials, Plante and colleagues1 conclude that systemic steroids in perioperative period do not appear to increase post-tonsillectomy bleeding events, but are associated with an increased incidence of surgical reinterventions for bleeding episodes. Actually, the results of a meta-analysis are only as good as the quality of the collected data, and meta-analysis addressing the same question can reach different conclusions, depending on the quality of the trials that are included. Other than the limitations reported in the discussion, we note that additional weaknesses of this meta-analytic study may have made interpretation of their conclusions questionable.
First, both grade of surgeon and operation technique have a significant influence on the frequency of post-tonsillectomy bleeding events.2 Among 29 randomised controlled trials included in this meta-analysis, some of them had clearly defined and compared those two factors. For example Czarnetzki et al’ study,3 it involved 22 different surgeons, their surgical method of tonsillectomy varied, and most of severe bleeding episodes in dexamethasone group occurred with three of the surgeons. If removing those three surgeons with most bleeding cases (total of 12 cases), bleeding events are not significantly different between dexamethasone and placebo groups. Moreover, Czarnetzki et al.3 did not standardise many determinants of post-tonsillectomy bleeding events, such as diagnosis, surgical technique, and administration of ibuprofen. Thus, it is unfair that all post-tonsillectomy bleeding events are contributed to a single use of dexamethasone.
Second, this meta-analysis had pooled results from different studies referred to systemic steroids administration as a single dose, or repeated doses over a specific period of time (16 h to 8 days) for different reasons on different patient populations. This would further have confounded their findings. The potential adverse effects of long-term use of systemic steroids include difficulty in controlling intraoperative blood sugar levels, delayed wound healing, increased incidence of wound infection and avascular necrosis, suggesting cautious use in surgical patients.4 However, a single intravenous dose of dexamethasone is an effective and safe treatment for reducing morbidity from paediatric tonsillectomy, such as postoperative nausea and vomiting, and pain.5 In the available literature, there is no report of side-effects from use of a single intravenous dose of steroids during tonsillectomy. Recently, in a large meta-analysis, Shargorodsky et al.6 concluded that there was no overall association between single dose dexamethasone administration and postoperative bleeding in children undergoing tonsillectomy or adenotonsillectomy. Moreover, a randomised controlled trial of 314 children undergoing tonsillectomy showed that use of dexamethasone 0.5 mg/kg (up to a maximum of 20 mg) at the start of the operation did not result in more level Ⅱ or Ⅲ bleeding events than placebo.7 Because of relative safety, low cost, and reduced postoperative morbidity, it is generally recommended routine use of a single intravenous dose dexamethasone during paediatric tonsillectomy.5,8 Thus, their conclusion that systemic steroids should be used with caution in patients undergoing tonsillectomy is needed to be reconsidered.
Acknowledgement: None of the authors received financial support and there are no potential conflicts of interest for this work.
References
1 Plante J, Turgeon AF, Zarychanski R, Lauzier F, Vigneault L, Moore L, Boutin A, Fergusson DA. Effect of systemic steroids on post-tonsillectomy bleeding and reinterventions: systematic review and meta-analysis of randomised controlled trials. BMJ 2012; 345:e5389.
2 Sarny S, Ossimitz G, Habermann W, Stammberger H. Hemorrhage following tonsil surgery: a multicenter prospective study. Laryngoscope 2011; 121:2553-60.
3 Czarnetzki C, Elia N, Lysakowski C, Dumont L, Landis BN, Giger R, Dulguerov P, Desmeules J, Tramèr MR. Dexamethasone and risk of nausea and vomiting and postoperative bleeding after tonsillectomy in children: a randomized trial. JAMA 2008; 300:2621-30.
4 Ho CM, Wu HL, Ho ST, Wang JJ. Dexamethasone prevents postoperative nausea and vomiting: benefit versus risk. Acta Anaesthesiol Taiwan 2011; 49:100-4.
5 Steward DL, Grisel J, Meinzen-Derr J. Steroids for improving recovery following tonsillectomy in children. Cochrane Database Syst Rev 2011; 8:CD003997.
6 Shargorodsky J, Hartnick CJ, Lee GS. Dexamethasone and postoperative bleeding after tonsillectomy and adenotonsillectomy in children: A meta-analysis of prospective studies. Laryngoscope 2012; 122:1158-64.
7 Gallagher TQ, Hill C, Ojha S, Ference E, Keamy DG, Williams M, Hansen M, Maurer R, Collins C, Setlur J, Capra GG, Brigger MT, Hartnick CJ. Perioperative dexamethasone administration and risk of bleeding following tonsillectomy in children: a randomized controlled trial. JAMA 2012; 308:1221-6.
8 Steward DL, Welge JA, Myer CM. Do steroids reduce morbidity of tonsillectomy? Meta-analysis of randomized trials. Laryngoscope 2001; 111:1712-8.
Competing interests: None declared
Plastic Surgery Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 33 Ba-Da-Chu Road, Shi-Jing-Shan District, Beijing100144, People's Republic of China
Click to like:
14
+Vote up!
22 September 2012
We(1) agree with Drs Diakos and Gallos that meta-analyses of rare events need careful statistical approaches, choices, and interpretation. We chose Peto odds ratios (OR) as this method is more conservative than alternatives and preferred when baseline risk is low, effect size is small and randomization groups are well balanced.(2-4) Indeed, Peto ORs tend to underestimate associations in the presence of strong effect sizes, thus offering a more conservative analytic approach in the case of rare events.(3) The randomization balance in our included studies ranged from 1:1 to 1:3. While large imbalances (ie. >8:1) can be problematic for Peto ORs, this large imbalance leads to an underestimation of the treatment effect rather than overestimation.(5) As such, our pooled estimate would tend to underestimate the association between steroids and postoperative bleeding, readmission and reintervention with the use of Peto ORs. In addition, our decision to use a continuity correction for studies with no events in both groups allows the inclusion of all eligible randomized controlled trials in the analyses. The inclusion of these studies dilutes our effect size estimate in comparison to not using a continuity correction,(6) and thus, the bias is again toward an underestimation of treatment effect size.
While the study by Czarnteski et al. may be criticized for stopping early, there is no sound justification for excluding their data from our meta-analysis. In fact, exclusion of a particular study may generate the very bias we want to avoid in performing meta-analyses. Although we agree with Diakos and Gallos that steroids reduce postoperative pain, nausea and vomiting, their systematic review considered only studies performed in adults (age > 16 years old). Indeed, we observed in our a priori planned sensitivity analyses based on age that children seem to be more prone to reintervention than adults (OR 3.43 [1.29 to 9.13]). This could explain an absence of effect for reintervention due to bleeding in their meta-analysis.
Most studies included in our meta-analysis used systemic steroids for the prevention of postoperative nausea and vomiting (PONV). Considering that a reintervention for bleeding following tonsillectomy outweights PONV in terms of clinical significance and effective alternatives (e.g serotinergic 5-HT3 receptors antagonist agents) are available, the routine use of steroids in tonsillectomy procedures should be reconsidered.
References
1. Plante J, Turgeon AF, Zarychanski R, Lauzier F, Vigneault L, Moore L, et al. Effect of systemic steroids on post-tonsillectomy bleeding and reinterventions: systematic review and meta-analysis of randomised controlled trials. BMJ 2012;345:e5389.
2. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]: The Cochrane Collaboration. Available from http://www.cochrane-handbook.org, 2011.
3. Bradburn MJ, Deeks JJ, Berlin JA, Russell Localio A. Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events. Stat Med 2007;26(1):53-77.
4. Lane PW. Meta-analysis of incidence of rare events. Statistical methods in medical research 2012.
5. Greenland S, Salvan A. Bias in the one-step method for pooling study results. Stat Med 1990;9(3):247-52.
6. Friedrich JO, Adhikari NK, Beyene J. Inclusion of zero total event trials in meta-analyses maintains analytic consistency and incorporates all available data. BMC Med Res Methodol 2007;7:5.
Competing interests: None declared
Department of Anesthesiology, Division of Critical Care Medicine, CHU de Québec, Université Laval, 1401, 18eme rue, Québec City, Québec
Click to like:
4 September 2012
Systemic steroids have been found to reduce post-operative nausea and vomiting along with pain in children and adults following tonsillectomy [1, 2]. This intervention is now widely used. This study finds weak evidence that this intervention may be harmful (p=0.04). Specifically, it finds that the bleeding episodes are not increased, but their severity assessed by re-operation. The number of the events is low as this is a rare occurrence (19/634 vs 8/544). Interestingly, 8 out of the 19 events in the tonsillectomy group are from the Czarnetzki trial that was stopped early for safety reasons [3]. The Czarnetzki trial had a high incidence of bleeding (12.3%), which is higher than the mean from this meta-analysis (4.4%). It was criticised for stopping early and that their conclusions were not supported by their data. A sensitivity analysis in this meta-analysis excluding the Czarnetzki trial will find no significant difference in the severity of bleeding. In summary, the possibility that the findings of the Czarnetzki trial could be chance findings could not be excluded. These data are also included in this meta-analysis. The only difference in events in the two groups is exactly the number of events included in the Czarnetzki trial. This may be another chance finding?
Caution has also been advised about the meta-analytical methods that can be used to combine summary-level data from trials with low event rates [4]. The Peto method used in this study is one of the methods. The Mantel-Haenszel (MH) method has an odds ratio of 2 (0.91 to 4), but a p value that is not significant at the conventional level (p=0.09). A random-effects MH model to take into account the clinical heterogeneity ignored by the fixed-effects model used in this study has an odds ratio of 1.83 (95% CI 0.77 to 4.33) and also not significant p value (p=0.17). It can also be argued that the Peto method is not preferable because it behaves poorly when randomisation ratio are more than 1 as is the case in several trials from this meta-analysis [4].
To conclude, in a meta-analysis for systemic steroids following tonsillectomy for adults we found this intervention to be protective for post-operative bleeding [1]. We presumed this was a chance finding as there was no biological plausibility for such a finding. Our events were few, the association was weak and there was no consistency in the published literature. The case here I suspect is no different.
References
1. Diakos E, Gallos ID, El-Shunnar S, Clarke M, Kazi R, Mehanna H. High dose peri-operative systemic steroids reduce pain, vomiting and overall complications following tonsillectomy in adults: A systematic review and meta-analysis of randomised controlled trials. Clin Otolaryngol. 2011;36:531-42.
2. Steward DL, Grisel J, Meinzen-Derr J. Steroids for improving recovery following
tonsillectomy in children. Cochrane Database Syst Rev 2011:CD003997.
3. Czarnetzki C, Elia N, Lysakowski C, Dumont L, Landis BN, Giger R, et al. Dexamethasone and risk of nausea and vomiting and postoperative bleeding after tonsillectomy in children: a randomized trial. JAMA 2008;300:2621-30.
4. Lane PW. Meta-analysis of incidence of rare events. Stat Methods Med Res 2012 DOI: 10.1177/0962280211432218.
Competing interests: None declared
University of Birmingham, Queen Elizabeth Hospital, Birmingham B15 2TH
Click to like:
43
+Vote up!
Post Tonsillectomy bleeding has the potential for catastrophic consequences. As such, it is right that much effort goes into devising strategies to minimise this occurrence.
What is rarely considered in preventing unintended excessive bleeding is the consideration that a patient might have an underlying bleeding disorder. This is surprising, as the incidence of Von Willebrands disease is 1 in 50.
In the assessment of someone’s risk of postoperative bleeding, it doesn’t take much trouble to take a bleeding history. That way, patients at high risk can be identified pre operatively and managed appropriately intraoperativley. For example, with the concurrent use of desmopressin, clotting factors and cyclokapron.
It would be a feather in the cap to ENT surgeons were they to pilot this – After all, up to 25% of women having a hysterectomy for heavy menstrual bleeding with a normal sized uterus have Von Willebrands disease. Yet gynaecologists rarely, if ever, take this into consideration when preparing to undertake major surgical procedures upon women.
Competing interests: None declared
Royal Oldham Hospital, Rochdale Road, Oldham, Lancs, OL1 2JH
Click to like:
43
+Vote up!
Latest comments and most commented
Replacement of lead by depleted uranium: no reasons for radiophobia.
Published 23 May 2013
Please could we have some fact.
Published 23 May 2013
Re: CT radiation risks coming into clearer focus
Published 23 May 2013
Re: Should we sequence everyone’s genome?
Published 23 May 2013