• Clinical cure defined as no fever, no abdominal pain or bleeding. [ Time Frame: 10 days after hospital discharge ] [ Designated as safety issue: Yes ]
  

Septic abortion is still a major cause of maternal mortality in developing countries. According to the WHO, 1 woman dies for every 270 illegal abortion (Ahman E, 2004). Infected abortion has an important role in maternal morbidity and mortality (Stubblefield PG, 1994). the diagnosis of infected abortion must be considered when a patient presents a history of delayed menses, vaginal bleeding, abdominal pain and fever (Brasil, 2000)

Prompt diagnosis and treatment are paramount steps to prevent complications. At Hospital de Clínicas de Porto Alegre, the use of gentamycin plus clindamicin before curettage is preconized (Savaris R, 2006). Nevertheless, the time of treatment it is not well established, varying from 7-14 days (Brasil, 2000).

A recent study with post-partum endometritis has shown that it is not necessary to extend the treatment to 14 days, after clinical improvement (Turnquest MA, 1998; French LM, 2004)

A randomized clinical trial comparing placebo with the standard protocol of treatment would define weather both treatments are equivalent or not. Comparison: The prolonged use of antibiotics, after intravenous use of antibiotics and clinical improvement, will be compared to the use of placebo in cases of septic abortion.

Sample size and ethical issues The study protocol was approved by the ethics committee of Hospital de Clínicas de Porto Alegre.

To compare equivalence between the 2 treatments we calculated the sample size considering an alpha error of 0.05, a beta error of 0.1, and difference between the two groups of no more than 10%. We expected a 99% clinical cure with the standard protocol, and 95% for the alternative one. These figures yield a minimum of 42 patients in each group. Interim analysis will performed at 58 for possible early stopping, if clinical cure was < 95%, or for sample size re-estimation.

Randomization and treatment Subjects will be allocated in blocks of four at a time to create the allocation sequence. If the patient was eligible for the study, she will be allocated to one of the 2 treatments. The allocation will be concealed, coded and obtained from a central telephone number. Patients and those who assessed the outcomes were blind to group assignment. To avoid bias, both medications were manipulated by the hospital pharmacy and put in identically coded blisters and capsules.

Statistical analysis Student´s t-test, Mann-Whitney test, and Fisher´s exact test will be used for statistical analysis. The rates of cure were analyzed by "modified" intention to treat (Keech AC, 2003) and per protocol with 95% confidence intervals.