Coetzee K, Kruger T F, Lombard C J, Shaughnessy D, Oehninger S, Ozgür K, Pomeroy K O, Muller C H
Obstetrics and Gynaecology Department, Tygerberg Hospital, University of Stellenbosch, South Africa.
Fertil Steril. 1999 Jan;71(1):80-4. doi: 10.1016/s0015-0282(98)00421-x.
To evaluate the level of variance produced in a multicenter study with the use of a computer-assisted sperm morphology analyzer.
A multicenter, prospective, blinded study.
Assisted reproduction research laboratories.
PATIENT(S): Semen samples produced for assisted reproductive procedures.
INTERVENTION(S): Hamilton Thorne Research (Beverly, MA) integrated visual optical system semen analyzers were used at five different centers to evaluate the same set of 30 slides that were prepared and numerically coded at Tygerberg Hospital in Tygerberg, South Africa.
MAIN OUTCOME MEASURE(S): The percentage of normal sperm.
RESULT(S): Interlaboratory coefficients of variation (CVs) ranged between 16.31% and 23.09%. One of the participating laboratories produced an approximately 14% (-6.5-7.7) limits of agreement analysis, with a CV of 11.36%, for its duplicate readings. The use of a 10% normal sperm morphology cutoff point to determine discordance levels produced rates ranging between 10% and 23.3% for the interlaboratory and intralaboratory readings. This level of discordance equates with < or = 7 of the corresponding readings from two laboratories falling into a different normal sperm morphology group (< or = 10% or >10%).
CONCLUSION(S): The magnitudes of variation produced by the readings performed in our study reached the same level as for the manual evaluation of sperm morphology. A < 10% CV can be obtained if the correct quality control measures are implemented.
