DOI: 10.1097/CCM.0b013e3181aff568
,
Issn Print: 0090-3493
Publication Date: 2009/09/01
Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
The statistical evidence in this article is said to be derived from the log rank test. However, the log rank test is not suitable for the evaluation of long term effects of a short term therapy. The log rank test can only be used to compare two regimens when the therapies used are given over the whole period of the test, until an adverse event occurs, or till the death of a participant. In this study the therapies are only given over 24 hrs, whereas the conclusion includes up to 29 undocumented days thereafter; in such a case, the correct statistics should be derived from the chi-square test. In this study, this test would give a p value of 0.2; therefore, there is no statistical significance.
I would also like to make some methodological remarks. In the levosimendan group a fixed dose is given during 24 hrs, whereas in the enoximone group a variable dose is given, and there is no mention of the duration. Also, the therapy of both groups after cessation of both studied drugs is undocumented. At the end of the 24-hr administration period there is no difference in hemodynamic values; however, no further therapeutic and/or hemodynamic data are given, so there is no evidence that the deaths occurring within both studied groups after the 24-hr period can be attributed to the therapy in the first 24 hrs.
I would like to see a study with better methodology and correct statistical analysis.
The administration period should be equal, the dose of both drugs should be constant and compatible, the therapy after the studied period should be documented and taken into consideration, and last but not least, the log rank test should be replaced by the chi-square test.
The author received a patent for Q-PiDT.
