Recommendation ID

What is the optimal antibiotic prophylaxis regimen for women who are having a surgical abortion?

Any explanatory notes
(if applicable)

Why the committee made the recommendations

The evidence on antibiotic prophylaxis for women who are having medical abortion showed lower rates of severe infection with antibiotic prophylaxis compared with no antibiotic prophylaxis. However, the committee had concerns with the quality of the evidence, and the absolute risk of severe infection was very low. In addition:

  • routinely prescribing antibiotics after medical abortion may increase the risk of antibiotic resistance

  • adherence is likely to be low

  • if using routine antibiotics the potential reduction in the risk of post-abortion infection is uncertain.

With these points in mind, the committee did not recommend routine antibiotic prophylaxis for women who are having a medical abortion. The committee believed that prophylaxis may be appropriate for women who are at high risk of infection, or who would find it difficult to access treatment at a later date if they screened positive for a sexually transmitted infection. However, there was no evidence to support recommending prophylaxis for a specific group.

Antibiotic prophylaxis is part of current clinical practice for women having a surgical abortion. The committee wanted to encourage this, so they made a recommendation in support.

The evidence reviewed did not identify which specific antibiotic regimen is most effective. Because of this, the committee agreed that further research would be beneficial and made a research recommendation. However, the committee were aware of a Cochrane review which showed the effectiveness of nitroimidazoles (such as metronidazole), tetracyclines (such as doxycycline) and beta lactams (such as amoxicillin).

A 7-day course of doxycycline is currently used in practice. There was no evidence comparing doxycycline with other antibiotics; however, there was some limited evidence on duration of doxycycline. The evidence was unclear on whether or not there were clinically important differences between 3‑day and 7‑day courses of doxycycline in the rates of pelvic inflammatory disease after abortion, patient adherence, vomiting, or diarrhoea. The committee recommended a 3‑day course because this may be as effective and adherence is likely to be better with a shorter course.

Metronidazole in combination with another broad-spectrum antibiotic is not routinely recommended because:

  • compared with doxycycline alone for surgical abortion, it was unclear if it made a clinically important difference to the rate of pelvic inflammatory disease after abortion in women who had elevated vaginal pH and amines in vaginal discharge, or a positive gram stain for bacterial vaginosis

  • although there was no evidence on the gastrointestinal side effects when compared with doxycycline alone, the committee agreed that in clinical practice metronidazole may be poorly tolerated with significant side effects.

However, the committee agreed that metronidazole is effective for anaerobic infections, so there may be situations where it is clinically indicated.

The evidence for the recommendations on doxycycline and metronidazole was based on antibiotic prophylaxis for surgical abortion. However, the committee agreed that the recommendations would also be appropriate if there is a need to use antibiotic prophylaxis for medical abortion as the causes of infection would be similar for both procedures, although the level of risk may differ.

The committee could not make recommendations about screening for sexually transmitted infections, because they did not review the evidence this. A cross-reference has been included to the NICE guideline on preventing sexually transmitted infections and under-18 conceptions.

Source guidance details

Comes from guidance
Abortion care
Date issued
September 2019

Other details

Is this a recommendation for the use of a technology only in the context of research? No  
Is it a recommendation that suggests collection of data or the establishment of a register?   No  
Last Reviewed 30/09/2019