Risk of retained placenta: multivariate approach.

BACKGROUND

To identify important covariates related to retained placenta using logistic regression analysis.

METHODS

The study was carried out in the King Khalid University Hospital, Saudi Arabia, and involved 114 women who had retained placenta, and 116 women with normal deliveries. Chi-square test and logistic regression analysis were used for analysis of data. Adequacy of predictor variables was examined using indices of sensitivity and specificity and plots of probability histograms for prediction of retained placenta among patients and controls.

RESULTS

Logistic regression analysis highlighted multiparity, induced labor, small placenta, and large amount of blood loss to be significantly associated with retained placenta. In addition, high pregnancy number, previous injury to uterus, and pre-term labor related significantly to retained placenta. Predictor variables had sensitivity of 65.5% and 87.9% specificity and achieved 77% overall correct classification.

CONCLUSIONS

These findings could be used to develop a predictive procedure for identifying high-risk cases of retained placenta.