Abstract
Aim of the study: Is to evaluate factors most correlated to homeostatic model analysis (HOMA-IR) and to construct a formula which can predict HOMA from the cheaper biochemical, biophysical and hormonal risk factors.
Study design; case control
Methods: A total of 34 patients with polycystic ovarian syndrome were recruited from infertility clinic in Al-Yarmuok teaching hospital. From the epidemiological characteristics the following were recorded: BMI, waist to hip ratio, menstrual cycle, hirsutism which was scored by modified Ferriman Gallwey system and the mean ovarian volume from measuring both ovaries while from the hormonal and biochemical variables, luteinizing hormone and follicle stimulating hormone, total testosterone and lipid profile were measured for each patient. For each group of variables stepwise regression followed by best subset regression were implemented to isolate the factors most correlated with HOMA. Coefficient of Mallows (Cp) was used to sort out the single most predictive variable.
Results: Coefficient of Mallows (Cp) was the smallest for BMI among the demographic and clinical characteristics hence the most correlated to HOMA-IR while from the hormonal and biochemical characteristics, high density lipoprotein was the factor most correlated to HOMA-IR as assessed by Coefficient of Mallows (Cp). Accordingly a formula which predicts HOMA-IR directly from BMI and HDL was designed through which HOMA can be estimated easily through the constructed table from the formula. In addition, its percentile can be accessed from an associated table to measure insulin resistance severity.
Conclusion: HOMA was shown in this study to be best correlated with BMI and HDL and from those factors predictive formula was constructed which can predict HOMA-IR and its percentile directly. However extreme caution is called when such formula is used in clinical practice as its validity is neither established nor such tables were constructed before, and not until their validity is established by further studies we advise extreme caution in interpreting results.
Study design; case control
Methods: A total of 34 patients with polycystic ovarian syndrome were recruited from infertility clinic in Al-Yarmuok teaching hospital. From the epidemiological characteristics the following were recorded: BMI, waist to hip ratio, menstrual cycle, hirsutism which was scored by modified Ferriman Gallwey system and the mean ovarian volume from measuring both ovaries while from the hormonal and biochemical variables, luteinizing hormone and follicle stimulating hormone, total testosterone and lipid profile were measured for each patient. For each group of variables stepwise regression followed by best subset regression were implemented to isolate the factors most correlated with HOMA. Coefficient of Mallows (Cp) was used to sort out the single most predictive variable.
Results: Coefficient of Mallows (Cp) was the smallest for BMI among the demographic and clinical characteristics hence the most correlated to HOMA-IR while from the hormonal and biochemical characteristics, high density lipoprotein was the factor most correlated to HOMA-IR as assessed by Coefficient of Mallows (Cp). Accordingly a formula which predicts HOMA-IR directly from BMI and HDL was designed through which HOMA can be estimated easily through the constructed table from the formula. In addition, its percentile can be accessed from an associated table to measure insulin resistance severity.
Conclusion: HOMA was shown in this study to be best correlated with BMI and HDL and from those factors predictive formula was constructed which can predict HOMA-IR and its percentile directly. However extreme caution is called when such formula is used in clinical practice as its validity is neither established nor such tables were constructed before, and not until their validity is established by further studies we advise extreme caution in interpreting results.
Keywords
body mass index
insulin resistance
polycyst
Keywords
-