The findings from a multi-country clinical trial published today in the New England Journal of Medicine show that the use of preventive antibiotics before miscarriage surgery did not result in a significantly lower risk of pelvic infection than the use of a placebo.
Roughly 10% to 20% of the 208 million pregnancies that occur each year globally end in miscarriage, and surgery is often required to remove the contents of the uterus. But whether prophylactic antibiotics help prevent post-surgical infection remains unclear.
And in low- and middle-income countries (LMICs), where rates of miscarriage and post-surgical infection are higher but antibiotics are frequently overused, more definitive evidence of the value of antibiotic prophylaxis is needed.
While the results suggest the value is limited, the international team of researchers who conducted the study and an outside expert say, however, that the broadening of the definition of pelvic infection in the middle of the trial muddies the picture.
The results of the Antibiotics in Miscarriage Surgery (AIMS) trial showed that, under pragmatic, broad criteria for pelvic infection that included clinical judgment, women who were given prophylactic antibiotics had a lower rate of infection 14 days after miscarriage surgery than those who received a placebo, but the risk of infection was not significantly lower (a 23% drop). Under a stricter definition of infection based on international diagnostic criteria, however, the risk of infection in women who had received antibiotics before surgery was 40% lower, suggesting "a possible benefit when pelvic infection was defined by strict criteria," the authors of the study write.
Expanded definition of infection
The randomized, double-blinded, placebo-controlled study involved women and adolescents recruited from hospitals in Malawi, Tanzania, Uganda, and Pakistan from June 2014 through April 2017. All the patients in the study had experienced miscarriages at less than 22 weeks before gestation and were scheduled to undergo surgery. Patients were assigned 1:1 to receive either 400 milligrams (mg) of doxycycline or 400 mg of metronidazole orally or five matched placebos 2 hours before surgery.
At the start of the trial, the initial primary outcome—pelvic infection 14 days after surgery—was defined according to criteria designated by the US Centers for Disease Control and Prevention and the World Health Organization. Under this definition, diagnosis of a pelvic infection requires two or more of four clinical features.
But during the trial, there was concern that this strict definition could lead to some missed infections. So the criteria were broadened to require only one of four clinical features of infection, along with the clinician's judgment that the patient had a pelvic infection and needed antibiotics for treatment. The initial primary outcome then became the secondary outcome. The changes were made before the data were un-blinded.
A total of 3,412 patients were enrolled in the study, with 1,705 assigned to receive antibiotics and 1,707 to receive placebo. Primary outcome data were obtained for 3,360 patients. The rate of pelvic infection was 4.1% in the antibiotic arm (68 of 1,676 pregnancies) and 5.3% (90 of 1,684 pregnancies) in the placebo arm (risk ratio [RR], 0.77; 95% confidence interval [CI], 0.56 to 1.04, P = 0.09).
For the secondary outcome, the rate of pelvic infection—diagnosed according to the stricter definition—was 1.5% (26 of 1,700 pregnancies) for the antibiotics group and 2.6% (44 of 1,704 pregnancies) for the placebo group (RR, 0.60; 95% CI, 0.37 to 0.96). The authors of the study note that they did not adjust these rates for multiple comparisons.
There were no significant differences in the rates of adverse events—including diarrhea, vomiting, and blood transfusion—between the two groups.
More sensitivity, less specificity
The investigators, led by David Lissaur, MBChB, PhD, of the University of Birmingham in England, acknowledge that while broadening the criteria for diagnosis increased the sensitivity for identifying pelvic infection, it may have also decreased the specificity. Under the broader criteria, there were an additional 42 infections diagnosed in the antibiotic-prophylaxis arm and an additional 46 in the placebo arm. "The addition of clinician judgment to the pragmatic definition is likely to have diluted the observed treatment effect," they write.
But were those additional diagnosed infections truly pelvic infections?
Writing in an accompanying commentary, Ugandan physician and infectious disease expert David Serwadda, MBChB, MPH, says that including clinician judgment is problematic because there was no clear definition of what constituted the perceived need for antibiotics among the clinicians. And in low- and middle-income countries with poor diagnostic facilities, he notes, healthcare providers may err on the side of caution and prescribe antibiotics for a miscarriage patient who's feverish and had surgery performed in less-than-optimal settings.
Serwadda believes that this less specific definition of infection may have inadvertently reduced the odds of observing a true difference in pelvic infection rates between the two groups.
As a result, he writes, "I would interpret the results as indicating that antibiotic prophylaxis prevented pelvic infections as defined by international diagnostic criteria. Given the risks associated with pelvic infections in LMICs, these data provide reasonable support for prescribing prophylactic antibiotics in these settings. Antibiotic resistance, however, will need to be monitored."
Mar 13 N Engl J Med study
Mar 13 N Engl J Med commentary