Analysis of risk factors of prolonged intensive care unit stay of critically ill obstetric patients: a 5-year retrospective review in 3 hospitals in Beijing

Abstract

To identify the risk factors of prolonged intensive care unit (ICU) stay of critically ill obstetric patients. A retrospective analysis of cases of critically ill obstetric patients admitted to the ICUs of Peking University Third Hospital, Capital Medical University Affiliated Beijing Chaoyang Hospital, and PLA 306 Hospital from January 1st 2006 to December 31st 2010 was made. Data included demographics, causes of critical illness or complications that prompted ICU admission, the acute physiology and chronic health evaluation II (APACHEII) scores, the time intervals between onset of acute symptoms and ICU admission, laboratory test results, treatment measures, length of ICU stay and the final maternal mortality. Data were used to identify univariate and multivariate predictors for prolonged ICU stay. During the 5-year period there were 207 obstetric patients [mean age (31.74±2.32) years old, mean gestational age (34.86±4.72) weeks] were transferred to the ICU for critical care (42 ICU admissions per 10 000 deliveries), and among them 4 women died (mortality rate 1.93%). The pathogenesis of the cases could be divided into direct obstetric pathologies (n=138) and indirect or coincidental pathologies (n=69). The most common obstetric causes of admission were massive postpartum haemorrhage (n=42, 20.29%) and pregnancy associated hypertension (n=36, 17.39%), followed by acute fatty liver of pregnancy (AFLP, n=27, 13.04%), obstetric disseminated intravascular coagulation (DIC, n=23, 11.11%). The most common non obstetric causes of admission were acute heart failure (n=26, 12.56%) and acute respiratory failure (n=22, 10.63%), followed by severe acute pancreatitis (SAP, n=11, 5.31%). The incidence of prolonged ICU stay (ICU stay >3 days) was 52.66% (n=109). Univariate predictors of prolonged ICU stay included inadvertent antenatal care, high serum creatinine, abnormal international normalized ratio of prothrombin time (INR), abnormal oxygenation index (PaO(2)/FiO(2)), AFLP, obstetric DIC, SAP, use of mechanical ventilation, indications for inotropic support, and blood purification, and >24 hour interval between onset of acute symptoms and ICU admission. Multivariate modeling identified that inadvertent and irregular antenatal care [odds ratio (OR) 1.68, 95% confidence interval (95%CI) 1.14-2.69, P=0.011], PaO(2)/FiO(2) (OR 4.73, 95%CI 1.46-11.37, P=0.013), AFLP (OR3.21, 95%CI 1.13-4.76, P=0.026), DIC (OR 2.73, 95%CI 1.28-4.02, P=0.018), SAP (OR 4.78, 95%CI 1.83-7.42, P=0.021 ), indications for inotropic support (OR 1.96, 95%CI 1.24-3.15, P=0.001), blood purification (OR 11.02, 95%CI 3.04-58.02, P=0.015) and >24 hour interval between onset of acute symptoms and ICU admission (OR 2.04, 95%CI 1.21-4.25, P<0.001) were the independent predictors for prolonged ICU stay. The incidence of prolonged ICU stay is high for critically ill obstetric patients. Patients with identified multivariate predictors carry a high risk of prolonged ICU stay, they may benefit from enhanced regular antenatal care, prevention of critical obstetric and medical complications, shortening the interval between onset of acute symptoms and ICU admission and strengthening the support of organ function.