American Journal of Emergency Medicine 38 (2020) 2224

Contents lists available at ScienceDirect

American Journal of Emergency Medicine

journal homepage:

Inaccurate outcome evaluation and conflict in odds ratio in multivariable analysis

To the Editor,

We read with great interest the original article “Multicenter Retro- spective analysis of the risk factors for Delayed neurological sequelae after acute carbon monoxide poisoning” by Zhang et al. [1]. Carbon monoxide poisoning is still one of the serious hospital admissions in many countries. In this study, the authors wanted to analyze the fac- tors that determine the presence of delayed neurological sequelae due to carbon monoxide poisoning. However, we think that there are some serious mistakes and shortcomings methodologically.

Firstly, the word “correlation” is used incorrectly in the article. Basi- cally, correlation is one of the statistical methods, but it is used incor- rectly to express that there is no statistical significance between the two variables in daily use [2]. The main outcome of this study was deter- mined as predictors of delayed neurological sequelae, not correlation of potential variables of CO poisoning.

Secondly, the explanation of delayed neurological sequelae in the study is unclear. DNS is the ultimate outcome in CO poisoning, thus out- come evaluation should be elaborated clearly. Who evaluated the pa- tient records for this diagnosis, emergency physician or neurologist? Was the physician’s experience sufficient? What were the newly- onset neuroPsychiatric symptoms? Sufficient explanation should have been made for such a study.

The last and most controversial issue is that higher Glasgow Coma Scale (GCS) was associated with the occurrence of DNS. Authors re- ported that DNS group had older age, longer CO exposure, and lower GCS on-site and at emergency room compared with non DNS group in the univariate analysis. However, after multivariable regression analy- sis, it was found that older age (OR, 1.11; 95% CI, 1.06-1.17), longer

duration of CO exposure (OR, 1.54; 95% CI, 1.06-2.22), higher GCS on- site (OR, 2.06; 95% CI, 1.39-3.05), and higher GCS at emergency room (OR, 1.33; 95% CI, 1.00-1.76) were independent predictors for DNS. Older age and long duration of CO exposure came out as expected after multivariate analysis. On the other hand, the odds ratio of GCS above 1 is not compatible with univariate analysis. According to the uni- variate results in the article, the odds ratio of the coma score should have been below 1 [3]. As readers, we could not fully understand how this result was achieved.

Declaration of Competing Interest

No financial disclosure was declared by the authors.

References

1. Zhang Y, Lu Q, Jia J, Xiang D, Xi Y. Multicenter retrospective analysis of the risk factors for delayed neurological sequelae after acute carbon monoxide poisoning. Am J Emerg Med. 2020. https://doi.org/10.1016/j.ajem.2020.06.090.
2. Akoglu H. User’s guide to Correlation coefficients. Turk J Emerg Med. 2018;18(3): 91-3. https://doi.org/10.1016/j.tjem.2018.08.001.
3. Szumilas M. Explaining odds ratios. J Can Acad Child Adolesc Psychiatry. 2010;19(3): 227-9.

   Mehmet Ali Aslaner MD

   Department of Emergency, Gazi University School of Medicine, Ankara,

   Turkey

   Corresponding author.

   E-mail addresses: [email protected], [email protected]

   Nurettin Ozgur Dogan MD

   Department of Emergency, Kocaeli University School of Medicine, Kocaeli,

   Turkey

   28 July 2020

   https://doi.org/10.1016/j.ajem.2020.07.087

   0735-6757/(C) 2020