A tow vaginal "pool" amniotic fluid glucose measurement is a predictive but not a sensitive marker for infection in women with preterm premature rupture of membranes

Objective: We Sought to identify the use of vaginal amniotic fluid (vAF) glucose measurements in predicting infection of the amniotic fluid retrieved by transabdominal amniocentesis (aAF) in women with preterm premature rupture of the membranes (PPROM) Study design: Fluid was retrieved by aAF was retreived from 35 consecutive women with PPROM on whom an amniocentesis was clinically indicated to rule out intra-amniotic infection/inflammation and successfully completed. aAF was cultured for aerobic, anaerobic bacteria, Ureaplasma and Mycoplasma species. Clinical laboratory analysis for aAF included glucose concentration, Gram stain, lactate dehydrogenase, and white and red blood cell count. vAF was analyzed only for glucose concentration. Glucose concentration for the paired abdominal-vaginal AF samples (aAF-vAF) was determined by using well-established clinical and research laboratory methods. At the end of enrollment we stratified our patients into 2 groups: (1) positive microbial cultures ( + )AFC(n = 17, gestational age [GA]: 27.3 +/- 0.7 weeks) or (2) negative microbial cultures (-)AFC (n = 18, GA: 31.3 +/- 0.5 weeks). Cohen kappa measure of concordance and receiver operating characteristic (ROC) curve analysis were used to test the ability of the vaginal "pool" glucose measurements to discriminate between women with positive or negative AF Cultures. Results: Women with (+)AFC ruptured and delivered at an earlier GA compared with the (-)AFC group(p <.001). The latency period was similar (P = .35). There was a significant linear correlation between aAF and vAF glucose concentrations (r = 0.783, P < .001). Women with intra-amniotic infection (IAI) had significantly lower aAF [mean +/- SEM ( + )AFC: 11.4 +/- 3.2 vs (-)AFC 23.0 +/- 2.8 mg/dL, P =.01] and vAF glucose levels [(+)AFC: 10.1 +/- 2.8 vs (-)AFC: 19.8 +/- 2.9 mg/dL, P = .02] compared with the noninfected group. Cohen kappa measure of concordance indicated "substantial" agreement between aAF and vAF glucose measurements (kappa = 0.719, 95% CI = 0.491-0.947). The sensitivity of the vAF glucose level to detect IAI ranged from 82% to 47%, whereas specificity ranged from 100% to 56% depending on the threshold we used. A vaginal "pool" (vAF) glucose measurement less than 5 mg/dL had 47.1% sensitivity, 100% specificity, 100% positive predictive value, 66.7% negative predictive value, and 74.2% accuracy in identifying women with ( + )AFC. Conclusion: Vaginal glucose determination is a readily available, inexpensive, rapid AF marker that can be measured practically in any clinical laboratory. vAF glucose measurements less than 5 mg/dL have predictive value,. but low sensitivity for detection of IAI. (c) 2006 Mosby, Inc. All rights reserved.