To determine whether extended-spectrum antibiotic prophylaxis that targets Ureaplasma urealyticum reduces postcesarean endometritis.
Following cesarean, the frequency of endometritis, wound infections, and a combination of the 2 was significantly lower among treated women than women given placebo.
The 597 women enrolled in this randomized, double-blind trial all were given cefotetan prophylaxis after cord clamping at cesarean delivery. Subjects then were randomized to receive doxycycline plus azithromycin (n = 301) or placebo (n = 296). Both groups were monitored for endometritis, defined as a fever of 100.4° F or higher with 1 or more supporting clinical signs (maternal tachycardia, foul-smelling or purulent lochia, tender uterus, and maternal leukocytosis) or as a physician diagnosis of endometritis and no nonpelvic source of fever. Among study participants, 56% were black, with an age of 25.5±6.2 years, and 43% were nulliparous. Groups were similar for race, parity, maternal age, and most risk factors for postcesarean endometritis.
Postcesarean endometritis occurred in 16.9% of treated women versus 24.7% of controls (P = .02), and wound infections affected 0.8% of treated women versus 3.6% of controls (P = .03).
Although the 2 groups were dissimilar for maternal leukocytosis (24.9% of treated women versus 12.5% of controls, P = .042) and classic uterine incision (7.6% versus 12.5%, P = .048), adjusting for these factors did not alter the risk ratio for postcesarean endometritis in the active versus placebotreated groups (relative risk 0.65, 95% confidence interval 0.43-0.98).
Length of stay was longer in the placebo group (104±56 versus 95±32 hours, P = .016) and among women with endometritis (146±52 versus 127±46 hours, P = .047).
This study tried to demonstrate that U. urealyticum is a significant pathogen and the etiologic agent for postpartum endometritis. However, the fact that U. urealyticum is found in the genital tract of approximately 70% to 90% of women does not support the thesis that it plays a major role in the microbial pathogenesis of postpartum endometritis.
Just because a microorganism is present in the lower genital tract does not mean that, in a state of infection, it is the etiologic agent. For example, many women harbor Enterococcus feacalis or Staphylococcus epidermidis in the genital tract; these often are isolated along with other bacteria from the site of infection. Yet the infection often is treated with antibiotics that offer no activity against these bacteria.
Lack of bacteriology limits relevance. Upon first analysis, this study appears to be sound, since it is both randomized and blinded. However, a major flaw weakens the conclusions significantly: lack of bacteriology.
The researchers neglected to obtain specimens for culture of bacteria from the uterus of each infected patient. Instead, they relied on statistical analysis, comparing the endpoint of infection versus no infection to extrapolate as to the cause. They failed to realize that the antibiotics used for prophylaxis—specifically doxycycline and azithromycin—also provide activity against Gram-positive and Gram-negative bacteria that make up the endogenous bacteriology of the vagina.
If endometrial specimens had been obtained from each infected patient, they would have provided a database on the frequency of involvement of the bacteria causing endometritis.
The authors do not state why they chose to use both doxycycline and azithromycin. In regard to the activity of these antibiotics against U. urealyticum there is probably not much difference as to efficacy.
The use of combinations of antibiotics for surgical prophylaxis is not advisable because it can lead to the selection of resistant strains that will remain in the patient’s lower genital tract. Therefore, I would not use a cephalosporin plus doxycycline and/or azithromycin as a regimen for surgical prophylaxis. Before such a combination can be recommended, further study is necessary that includes microbiology to establish which bacteria are responsible for postpartum endometritis. The microbiological studies must be quantitative—not qualitative—before conclusions can be drawn and clinical recommendations made.