Over the study period, from July 1998 to December 2004,
119 HIV infected women delivered by cesarean section at the above two teaching
institutions. Data was collected on 264
HIV-negative women who delivered by cesarean section during this time and served
as the controls. Maternal characteristics of the two groups are presented in . Overall, the majority of the patients
were African American (92.1% of HIV-infected group and 84.3% of HIV-negative
group, P < .05). Both groups
demonstrated a high body mass index (32.7 ± 7.8 kg/m2 in
HIV-infected group versus 33.6 ± 8.2 kg/m2 in HIV-negative group,
P > .05). The CD4 lymphocyte
count and viral load near time of delivery of the HIV-infected group were
evaluated; 78.8% of the HIV-infected women had a CD4 lymphocyte
count of ≥200 and 19.2% had undetectable viral load ().
characteristics of the two groups are compared in . There was no significant difference between
the HIV-infected group and the HIV-negative group in terms of gravidity,
parity, number of previous cesarean sections, estimated gestational age at time
of delivery, and rate of chorioamnionitis.
HIV-infected women were significantly more likely to have a lower
preoperative hematocrit than the HIV-negative women (31.9 ± 3.8% versus 33.6 ± 3.8%,
resp., P < .01). HIV-negative
women were significantly more likely than HIV-infected patients to deliver by
emergent cesarean section (78.0% versus 51.3%, resp., P < .01), to labor
prior to delivery (69.4% versus 48.3%, resp., P < .01), to have ruptured membranes
prior to delivery (63.5% versus 34.8%, resp., P < .01), and to have
application of internal monitoring devices during labor (31.5% versus 6.0%,
resp., P < .01).
The surgical characteristics of cesarean delivery in both
HIV-infected and HIV-negative women are presented in . Pfannenstiel skin incision were more often
performed in HIV-infected women than their controls (94.5% versus 82.7%,
resp., P < .01). There was no
statistical difference in the other characteristics reviewed which included the
type of anesthesia, the type of uterine incision, the estimated blood loss
during surgery, or the operative time.
Postoperative infectious morbidities are presented in . There was no significant difference in the
infectious morbidities between the HIV infected women and the control group. The
most common infectious morbidity after cesarean delivery for our study
population is postpartum endometritis.
Description of surgical procedures.
infectious complications among HIV-infected women and HIV-negative women.
was no statistical significant difference between the mean ± SD CD4 lymphocyte count between the HIV-infected women with infectious morbidity and
those without infectious morbidity (413.2 ± 257.9 cells/mm3 and 465.4 ± 283.7 cells/mm3, resp., P > .05), and between the mean
± SD viral load of those with infectious morbidity and those without infectious
morbidity (26,967.1 ± 79,491.6 copies/mL and 43,7242.9 ± 123,389.8 copies/mL,
resp., P > .05).
Postoperative infectious morbidity in both groups was analyzed
according to whether the cesarean section was an elective or emergent
delivery. Both groups of women were
statistically more likely to experience postpartum endometritis when being
delivered by emergent cesarean section than by elective cesarean section, (21.3%
versus 3.4%, resp., P < .05 in HIV-infected women and 14.6% versus 3.5%,
resp., P < .05 in HIV-negative women).
Postpartum endometritis composed the majority of the
post-operative infectious morbidity, and HIV-negative women in our study had
significantly more obstetrics risk factors for postpartum endometritis than the
HIV-infected women. Those risk factors were: delivery by emergent cesarean
section, rupture of membranes and labor prior to delivery, and insertion of
internal monitoring. Because of that, we further performed logistic regression
analysis, controlling for these risk factors to determine whether HIV infection
play a significant role in postpartum endometritis. To identify risk factors
influencing the risk of postpartum endometritis, we constructed a stepwise
proportional odds model. HIV infectious
status, emergent cesarean delivery, ruptured membranes prior to delivery,
application of internal monitoring devices during labor, chorioamnionitis, and
preoperative hematocrit were included in the model-building. HIV infectious status, rupture of membranes
prior to cesarean delivery, application of internal monitoring devices during
labor, and preoperative hematocrit were not found to be significant predictors
of endometritis (P > .05). However,
the odds of having endometritis were almost 4 times higher in emergent cesarean
delivery (odds ratio (OR) 4.1, 95% Confidence interval (CI) 1.41–1.91, P = .009) and 3 times higher in those
with chorioamnionitis during labor (OR 3.02, 95% CI 1.13–8.03, P = .027). It should be noted that because colinearity
was present between labor prior to delivery by emergent cesarean section and
emergent cesarean delivery, we only included emergent cesarean delivery in the
multivariate stepwise logistic regression model ().
variables influencing postpartum endometritis in a stepwise odds model.*