The purpose of this study was to compare guideline recommendations and day-to-day practice of serological testing for Lyme borreliosis (LB) in a laboratory located in Amsterdam, the Netherlands, serving both regional hospitals and primary care physicians. By telephone interview, we obtained clinical information regarding 488 requests for LB serology. Screening for LB was performed with a C6-peptide EIA and confirmed by recombinant immunoblot. A total of 82 % of the requests were not supported by guideline's recommendations and either originated from patients with atypical symptoms and a low a priori chance for LB or from patients for which testing on LB was not recommended for other reasons. C6-EIA screening was positive in 5 % of patients with atypical symptoms, comparable to the seroprevalence in the Dutch population. Interestingly, 10 % of the requests were from patients with atypical skin lesions, of which 20 % was positive, suggesting that serological testing is of additional value in a selection of such patients. Strikingly, only 9 % of the requests were supported by recommendations by guidelines. The percentage of positive confirmatory IgM and/or IgG immunoblots did not differ substantially between the groups and ranged from 56 to 75 %. Guidelines for testing for LB are not adequately followed in the Netherlands. Better education and adherence to the guidelines by physicians could prevent unnecessary diagnostics and antibiotic treatment of supposed LB patients.