Geometric mean IELT and premature ejaculation: appropriate statistics to avoid overestimation of treatment efficacy.

Introduction: The intravaginal ejaculation latency time (IELT) behaves in a skewed manner and needs the appropriate statistics for correct interpretation of treatment results.
Aims: To explain the rightful use of geometrical mean IELT values and the fold increase of the geometric mean IELT because of the positively skewed IELT distribution.
Methods: Linking theoretical arguments to the outcome of several selective serotonin reuptake inhibitor and modern antidepressant study results.
Main Outcome Measures: Geometric mean IELT and fold increase of geometrical mean IELT.
Results: Log-transforming each separate IELT measurement of each individual man is the basis for the calculation of the geometric mean IELT. A drug-induced positively skewed IELT distribution necessitates the calculation of the geometric mean IELTs at baseline and during drug treatment. In a positively skewed IELT distribution, the use of the "arithmetic" mean IELT risks an overestimation of the drug-induced ejaculation delay as the mean IELT is always higher than the geometric mean IELT. Strong ejaculation-delaying drugs give rise to a strong positively skewed IELT distribution, whereas weak ejaculation-delaying drugs give rise to (much) less skewed IELT distributions. Ejaculation delay is expressed in fold increase of the geometric mean IELT.
Conclusions: Drug-induced ejaculatory performance discloses a positively skewed IELT distribution, requiring the use of the geometric mean IELT and the fold increase of the geometric mean IELT.