Perioperative antibiotics to prevent infection after first-trimester abortion: Cochrane systematic review
Assessed as up to date: 2011/06/10
There are two main strategies for the prevention of post-abortal upper genital tract infection: antibiotics given around the time of surgery for all women; and 'screen-and-treat', in which all women presenting for abortion are screened for genital infections and those with positive results are treated.Objectives
1. the effectiveness of antibiotic prophylaxis in preventing post-abortal upper genital tract infection;
2. the most effective antibiotic regimen;
3. the most effective strategy.
We searched the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, EMBASE, POPLINE and LILACS. The search was last updated in May 2011.Selection criteria
Randomised controlled trials (RCTs) in any language including women undergoing induced first trimester surgical or medical abortion, comparing: 1) any antibiotic regimen to placebo, nothing, or another antibiotic; 2) screen-and-treat versus antibiotics. The primary outcome was the proportion of women diagnosed with post-abortal upper genital tract infection.Data collection and analysis
Two reviewers independently selected references and extracted data. We calculated risk ratios (RR) with 95% confidence intervals (CI). We used meta-analysis where appropriate and examined between trial heterogeneity using the I2 statistic. In the presence of between trial heterogeneity we also estimated the 95% prediction interval (PI).Main results
A total of 703 unique items was identified. We included 19 RCTs. There was evidence of small study biases (Egger test, P = 0.002). In 15 placebo-controlled RCTs there was an effect of antibiotic prophylaxis (pooled RR 0.59, 95% CI 0.46 to 0.75, 95% PI 0.30 to 1.14, I2 = 39%). There were insufficient data (three trials) to determine whether one regimen was superior to another. In one trial, the incidence of post-abortal upper genital tract infection was higher in women allocated to the screen-and-treat strategy (RR 1.53, 95% CI 0.99 to 2.36).Authors' conclusions
Antibiotic prophylaxis at the time of first trimester surgical abortion is effective in preventing post-abortal upper genital tract infection. Evidence of between trial heterogeneity suggests that the effect might not apply to all settings, population groups or interventions.
This review did not determine the most effective antibiotic prophylaxis regimen. Antibiotic choice should take into account the local epidemiology of genital tract infections, including sexually transmitted infections.
Further RCTs comparing different antibiotics or combinations of antibiotics with each other would be useful. Such trials could be done in low and middle income countries and where the prevalence of genital tract infections in women presenting for abortion is high.
Low Nicola, Mueller Monika, Van Vliet Huib AAM, Kapp Nathalie
Antibiotic prophylaxis for first trimester induced abortion
Infection of the upper genital tract, including the uterus and fallopian tubes, can cause complications after induced abortion. Antibiotics given around the time of the abortion (prophylaxis) could prevent this complication. We found 19 randomised controlled trials that looked at the effect of antibiotic prophylaxis on post-abortal upper genital tract infection amongst women requesting induced abortion in the first trimester of pregnancy. We looked at the effect of any antibiotic prophylaxis regimen on the outcome. Overall, the risk of post-abortal upper genital tract infection in women receiving antibiotics was 59% that of women who received placebo. There were, however, differences between the trial results over and above what would be expected by chance alone. It should be noted that, if the infection is caused by a sexually transmitted organism, antibiotic prophylaxis will not protect the woman from becoming re-infected if her sexual partner has not been treated. None of the trials was done in lower or middle income countries, which is where the risk of post-abortal complications is highest. Further trials are needed to determine whether combinations of antibiotics can prevent more infections than single antibiotics, or whether antibiotic prophylaxis should be restricted to women with positive results of screening tests before the abortion.
Implications for practice
A general strategy of perioperative antibiotics at the time of first-trimester surgical abortion is effective in preventing post-abortal upper genital tract infection, with an average reduction of 41% (95% CI 25 to 54%, random-effects model). The level of between trial heterogeneity suggests that this effect might not, however, apply to all settings, population groups or interventions. To take this into account, we also estimated a 95% PI, which is wider than the 95% CI (RR 0.59, 95% PI 0.30 to 1.14).
There are sub-groups amongst whom antibiotic prophylaxis had a beneficial effect, with no evidence of between trial heterogeneity: women receiving nitroimidazole antibiotics and single dose regimens; and settings in which the rate of post-abortal upper genital tract infection was 12% or more. In this review, there was a beneficial effect both in women with and without a history of PID.
The majority of trials included in the review did not evaluate a strategy of universal antibiotic prophylaxis as it would be applied in practice, i.e. giving prophylaxis to all women without doing tests to screen for existing gonorrhoea and chlamydia. This is because many trials had planned or actual exclusions (or treatment) of women who had infections diagnosed preoperatively. The prophylactic effect of antibiotics was actually weakest in the group of trials that did not use universal prophylaxis, perhaps because the opportunity to prevent post-abortal infections was reduced by the exclusion of those with infections. The antibiotic prophylactic regimen selected in practice should take into account the local epidemiology of lower genital tract infection.
This review did not determine the most effective antibiotic regimen because there were too few trials making such comparisons. In stratified analyses of placebo controlled trials nitroimidazoles prevented post-abortal upper genital tract infections with no evidence of between trial heterogeneity. Anaerobes or organisms associated with bacterial vaginosis might, therefore, be important aetiologically. In addition, two trials showed an effect of antibiotics active against chlamydia in women who were infected with C. trachomatis at baseline. Only one included trial used a combination of antibiotics; gave metronidazole and doxycycline. This antibiotic combination has been recommended in guidelines as it covers bacterial vaginosis and C. trachomatis. In a trial that was not included in the review because only half the women were in the first trimester, compared a combination of a seven day course of metronidazole and doxycycline with doxycycline alone in women with bacterial vaginosis. They found that the addition of metronidazole did not reduce the incidence of post-abortal infectious complications, defined using a symptom score. Single dose regimens also appeared to be associated with a consistent reduction in the risk of post-abortal upper genital tract infection. Of note, four of these trials assessed the outcome at two weeks or sooner and three of the trials used nitroimidazoles, which also showed a consistent effect.
The findings of this review are consistent with existing guidelines on antibiotic prophylaxis. In the USA, the American College of Obstetrics and Gynecology did not recommend any particular regimen, whilst the Society of Family Planning states that both nitroimidazoles and tetracyclines are effective. Guidance about the duration of the prophylactic regimen differs. The US Society of Family Planning recommends that antibiotics should not be given for more than three days. In the UK, Royal College of Obstetrics and Gynaecology guidelines recommend single dose metronidazole with single dose azithromycin or a seven day course of doxycycline. The Scottish Intercollegiate Guidelines Network has published general guidelines about antibiotic prophylaxis for surgical procedures and notes that in 'several studies... longer dose duration has no increased benefit' but no specific evidence about abortion was identified.
The implications of lower genital tract infections that are sexually transmitted or sexually transmissible for women and their sex partner(s) should be taken into consideration when developing strategies for the prevention of post-abortal upper genital tract infection. If pre-abortion screening tests for infection are not done, practitioners should give women information about the specific infections not covered by the prophylactic regimen, so that they can seek diagnosis, treatment and partner services. If pre-abortion infection screening tests are done, practitioners should provide full treatment and follow-up care for women diagnosed with a sexually transmitted infection. The single trial by did not determine whether or not there is a difference in the effectiveness of screen-and-treat and universal antibiotic prophylaxis strategies. There were fewer episodes of post-abortal upper genital tract infection in women receiving universal antibiotic prophylaxis, but 95% CI were wide. Furthermore, the authors of the trial tried to ensure treatment for partners to prevent re-infection but very few were known to have attended a clinic for treatment. The implications of this for re-infection are not known; the low partner notification success rate could reflect an inability to reach partners in partnerships that had ended, or a failure to reach ongoing sex partners.
The results of the review cannot be generalised to women having medical abortions because we did not find any relevant trials.
The results of the review cannot be generalised to women in the second trimester of pregnancy because the protocol specified only first-trimester abortion. Future updates should include second-trimester abortion.
Since all included trials were conducted in high income countries where testing is available, the results cannot necessarily be generalised to low and middle income countries, where the prevalence of sexually transmitted and endogenous infections in women requesting abortion might well differ and where screening tests might not be available.
Implications for research
Further RCTs comparing prophylactic regimens of different antibiotics with each other or combinations of antibiotics with a single antibiotic would be useful. Such trials could be done in low and middle income countries and settings in which the prevalence of lower genital tract infections in women presenting for abortion is high.
Observational cohort studies of women who have had abortions could give valuable information about the risk of re-infection and of upper genital tract damage as longer term consequences of abortion. Follow- up of RCTs could include a time period that is long enough to investigate the incidence of re-infection and the outcomes of partner notification, where appropriate, in women who have received antibiotic prophylaxis.
Further research to improve the accuracy and reproducibility of diagnostic criteria for upper genital tract infection would help to improve objective diagnosis.Get full text at The Cochrane Library
Evidence Central is an integrated web and mobile solution that helps clinicians quickly answer etiology, diagnosis, treatment, and prognosis questions using the latest evidence-based research. Complete Product Information.