These recommendations are based on guidelines by the Infectious Diseases Society of America (IDSA) on the treatment of Lyme disease in women who are pregnant or breastfeeding [Wormser et al, 2006], which have been endorsed by the UK Health Protection Agency (HPA) [HPA, 2009e], and on evidence of the efficacy and safety of antibiotics for erythema migrans in adults.

• Basis for recommending amoxicillin or cefuroxime axetil and avoiding doxycyline during pregnancy or breastfeeding
• IDSA guidelines recommend that adults and children with early localized or early disseminated Lyme disease associated with erythema migrans should be treated with doxycyline, amoxicillin, or cefuroxime axetil [Wormser et al, 2006].
• Due to the risk of its deposition in developing bone and teeth, doxycyline should not be given to women who are pregnant or breastfeeding [Wormser et al, 2006; ABPI Medicines Compendium, 2008; NTIS, 2008b].
• In adults, limited evidence from a systematic review, and one subsequently published open-label randomized controlled trial (RCT), suggests that penicillins and cefuroxime axetil (for 10–21 days) are equally effective in treating erythema migrans and preventing late complications of Lyme disease, and are superior to erythromycin or azithromycin. Following treatment, the risk of developing a major late complication (such as myocarditis, meningoencephalitis, or recurrent arthritis) is 0–8%, and the risk of developing a minor late complication (such as cranial neuropathy, transient arthritis, fatigue, or arthralgia) is 6–24%.
• The available data on the use of both penicillins and cephalosporins in pregnancy do not suggest they are associated with an increased risk of congenital abnormalities above the background rate for the population [NTIS, 2008a].
• Amoxicillin and cefuroxime can be used during breastfeeding. Only low levels of amoxicillin or cefuroxime are found in breast milk. Occasionally rash, diarrhoea or oral candidiasis have been reported in the infant [Toxnet, 2009a; Toxnet, 2009c].
• Basis for recommending cefuroxime axetil only if amoxicillin cannot be given
• CKS only recommends cefuroxime axetil if amoxicillin is contraindicated, because it may be associated with an increased risk of Clostridium difficile-associated illness [HPA, 2009a], and it is currently more expensive than the other recommended antibiotics [Wormser et al, 2006; ABPI Medicines Compendium, 2008; NTIS, 2008b; BNF 57, 2009].
• The recommendation that between 0.5% and 6.5% of people who are hypersensitive to penicillins are also hypersensitive to cephalosporins is based on information from the British National Formulary [BNF 58, 2009].
• Basis for choice of antibiotic if bacterial cellulitis could be present
• IDSA guidelines recommend that if erythema migrans cannot be reliably distinguished from bacterial cellulitis, treatment should be with co-amoxiclav or cefuroxime [Wormser et al, 2006].
• The recommendation to avoid co-amoxiclav in people with hepatic impairment is based on advice from the Commission on Human Medicines that cholestatic jaundice may (rarely) occur during or shortly after the use of co-amoxiclav [CSM, 1997].
• As with amoxicillin, co-amoxiclav is considered to be suitable for use during both pregnancy and breastfeeding [NTIS, 2008a; Toxnet, 2009b].
• The option of adding flucloxacillin to amoxicillin is in line with national guidance on the treatment of cellulitis in primary care [HPA, 2009a]. This was recommended by one CKS expert reviewer because of the risk of Clostridium difficile associated with the use of cefuroxime and co-amoxiclav.
• Basis for contacting an obstetrician for women who are pregnant
• There is a theoretical possibility of uterine contractions precipitated by Jarisch–Herxheimer reactions in pregnant women with Lyme disease, extrapolated from reports of this occurring in women being treated for syphilis [Schaefer et al, 2007]. However, none of the CKS expert reviewers (including one reviewer who contacted international colleagues) have ever seen a woman with a Jarisch–Herxheimer reaction precipitating uterine contractions in clinical practice.
• Ongoing monitoring may be required because there are case reports of cardiac defects in the infant following untreated maternal Lyme disease during pregnancy [Elliot et al, 2001]. However, causality has not been established; several case series found no increase in adverse pregnancy outcome in women with Lyme disease compared with controls. Overall, maternal Lyme disease is associated with a very low risk of adverse pregnancy outcome, particularly if it is treated with suitable antibiotics [Elliot et al, 2001].