CORRESPONDENCE

8. Hollander DI, Nagey DA, Pupkin MJ. Magnesium sulfate and ritodrine hydrochloride: a randomized comparison. Am J Obstet Gynecol. 1987;156:631-637.

9. Elliott JP. Subtherapeutic doses of magnesium sulfate do not inhibit preterm labor. Am J Obstet Gynecol. 1992;167:568-569.

10. Travis BE, McCullough JM. Pharmacotherapy of preterm labor. Pharmacotherapy. 1993;13:28-36.

In reply.

We thank Dr. Li for his letter regarding the optimal dosage of intravenous magnesium sulfate for acute asthma. Dr. Li suggests that dosing details were not duly noted in either ofthe 2 meta-analyses ~,2 and thatthis issue represents an important source of hetero- geneity. Although we agree that the dosing issue is important, we would like to remind Dr. Li of the potential limitations of system- atic reviews. First and foremost, there is often a limited abilityto pool data on the basis of dosing, given the underreporting of information in this domain and the failure on the part of authors to provide this informa- tion at a later date. Indeed, we found this to be the case for our meta-analysis. Second, the only alternative approach to this com- mon problem is to perform an individual patient data systematic review that would provide data for analysis with weight- adjusted dosing. This approach is very costly, time consuming, and difficult to perform. Consequently, dosing is not a commonly per- formed sensitivity analysis within Cochrane systematic reviews.

Dr. Li further suggests that a dose- response relationship should be "self-evi- dent" because it is such a common phe- nomenon. In fact, several analyses in the Cochrane Airways Review Group do not fol- low this self-evident truth. For example, a comparison of different doses of cortico- steroids in the treatment of acute asthma

failed to identify any dose-response curve. 3 Moreover, the subgrouping we identified (severe versus moderate-mild asthma) explained most of the observed heterogene- ity. Thus it is highly unlikelythat dosing issues are closely linked to the heterogeneity observed in our meta-analysis.

In addition, although we believe systematic reviews based on randomized, controlled trials are the highest level of evidence, we note that Dr. Li's evidence comes from case reports, a study design that is of some value but is gener- ally regarded as the lowest level of evidence. We believe the results from the evidence pro- duced in the systematic reviews are more valid than the anecdotal reports cited by Dr. Li.

Finally, the issue of magnesium levels in serum and cells and their association with bronchodilator response is very poorly described and understood. For every article that Dr. Li cites, we could cite similar articles providing contradictory information. Although we agree that the doses used in the treatment of asthma appear to be lower than those used in tocolysis, this only serves to reinforce our point regarding the safety of intravenous magnesium sulfate in the clinical setting. Furthermore, we note that concern has been expressed that higher doses of magnesium sulfate may cause bradycardia, respiratory depression, and signs of toxicity in some patients. 4

As we stated in our conclusion, there are many questions that remain regarding the treatment of acute asthma with magnesium sulfate. One of these questions certainly relates to the optimal dose for intravenous therapy, but questions of more clinical rele- vance need to be answered first, such as whether magnesium sulfate may be as effec- tive delivered by inhalation as intravenously 5 and whether aggressive standard therapy in

the placebo arm negates the effect of intra- venous magnesium sulfate. 6

What is clear at this point is that an intra- venous dose of 2 g in adults and up to 2 g in children provides little benefit to those with mild-moderate asthma exacerbations but appears to be effective in the management of severe exacerbations. Debates about opti- mal dosing are merely conjecture, and we agree with Dr. Li that they require further study. However, when systematic reviews demonstrate consistent benefit at lower doses than some would recommend, one could argue that this only enhances the con- clusion that magnesium sulfate is an effec- tive therapy.

Brian H. Rowe, MD, MSc University of Alberta Edmonton, Alberta, Canada Carlos A. Camargo, Jr., MD, DrPH Harvard University Boston, MA

47/8/114468 doi: 10.1067/mem.2001.114468

1. Rowe BH, Bretzlaff JA, Bourdon C, et at. Intravenous magnesium sulfate treatment for acute asthma in the emer- gency department: a systematic review q~ the literature. Ann Emerg Meal. 2000;36:181-190.

2. Alter Hj, Koepsell TD, Hilly WM. Intravenous magne- sium as an adjunct in acute bronchospasm: a meta-analysis. Ann Ernerg Meal. 2000;36:191-197.

3. Manser R, Reid D, Abramson M. Corticosteroids.for acute severe asthma in hospitalised patients (Cochrane Review). In: Cochrane Library, issue 4. Oxford, United Kingdom: Update Software; 1999.

4. Fletcher 5J, Parr MJ. Life-threatening magnesium toxic- ity. Intensive Care Med. 2000;26:257.

5. Mangat H5, D'Souza GA, Jacob MS. Nebulized magne- sium sulphate versus nebulized salbutamol in acute bronchial asthma: a clinical tdal. Eur Respir J. 1998;12:34t-344.

6. Rowe BH, Bretzlaff jA, Bourdon C, et al. Systematic review q~ magnesium sulfate in the treatment c!f acute asthma (Cochrane Review). In: Cochrane Library, issue 4. Oxford, United Kingdom: Update Software; 1998.

CORRECTION

In the December 2000 issue, the article by Tanen et al ("Hypertonic Sodium Bicarbonate Is Effective in the Acute Management of Verapamil Toxicity in a Swine Model"; pages 547-553), the unit of measure of ionized calcium in Table 1 and Figure 2 is incorrect. The concentration of ionized calcium should be in mg/dk, not mmol/k.

MAY 2001 37"5 ANNALS OF EMERGENCY M E D I C I N E 5 53