I wish to thank von Mandach for her comments regarding our paper1.
Women undergoing elective caesarean section are at low risk as they avoid the risk factors of labour and its duration, rupture of membranes and multiple vaginal examinations that their counterparts undergoing emergency caesarean section are exposed to. This difference is recognised by the reviewers of the Cochrane Library who have analysed antibiotic prophylaxis in elective caesarean section separately2.
Our reported infection rates are comparable to other reported studies in Europe3, and those serving indigent populations4. The study by von Mandach et al., defined wound infection as that having a purulent discharge and positive bacteriology. Our figures for purulent wound infection were 3.3% for placebo and 1.3% for cefoxitin, which are comparable to the von Mandach study. We also recorded wounds that revealed cellulitis and serous/serosanguinous discharge as evidence of wound infection. Regarding bacteriology, 4.6% and 3.4% of women in the placebo and cefoxitin groups, respectively, had microbiologically confirmed wound infection.
The aim of antibiotic prophylaxis is to reduce the number of bacteria colonising tissues at the operation site and not necessarily to eliminate all the bacteria present. Therefore, the drug chosen need only be active against most, but not all, of the pathogens that are likely to contaminate the wound. Expensive broad-spectrum antibiotics should be reserved for the treatment of serious wound infections and not used for routine prophylaxis, as they are not more effective than the cheaper limited spectrum antibiotics5. At King Edward VIII Hospital, Durban, cefoxitin is reserved for prophylaxis and is not used in patients with established infection.
von Mandach et al., compared three doses of cefoxitin to a single dose of broad-spectrum ceftriaxone and found no difference in the rate of fever, endometritis and wound infection. They have not compared single with multiple dose cefoxitin and are therefore not able to conclude that higher and more frequent doses will result in lower infection rates. Single dose regimens are recommended as they are effective as multiple dose regiments6–8.
The relative risk for endometritis in our study of 0.5 (0.09–1.92) was not found to be significant and together with a low baseline infection rate of 1.7%, one would need to treat 119 women to prevent one case of endometritis.
We regret the error regarding the anaesthesia data and confirm that over 90% of the women in both groups had spinal anaesthesia, as stated in the text.