Context Erythema migrans, while not pathognomonic, is the most common manifestation of early Lyme disease. Accurate diagnosis of this rash is essential to initiating appropriate antibiotic therapy.
Objective To determine the sensitivity of history and physical examination characteristics for the diagnosis of erythema migrans.
Data Sources Structured MEDLINE searches of articles written only in English, 1966 through March 2007.
Study Selection Studies were included if they enrolled at least 15 consecutive patients with the diagnosis of erythema migrans and reported original data regarding the history and physical examination characteristics of the patients.
Data Extraction One author abstracted data from the studies.
Results We separately analyzed the studies from Europe and analyzed both Lyme-endemic and nonendemic areas of the United States to search for potential differences in the clinical presentation. Thirty-two studies from Europe, 20 studies from the United States, and 1 from Europe and the United States met inclusion criteria for a total of 8493 patients. Sensitivity was calculated for each of the variables. No studies included patients without erythema migrans, so specificity data and likelihood ratios could not be determined. Many patients do not recall a tick bite. Associated systemic symptoms, such as fever and headache, are frequently reported. Nausea and vomiting are rare. A solitary lesion is the most frequent presentation in both US (81%; 95% confidence interval [CI], 72%-87%) and European patients (88%; 95% CI, 81%-93%). Central clearing is less common in the endemic United States (19%; 95% CI, 11%-32%) vs Europe (79%; 95% CI, 69%-86%) and the nonendemic United States (80%; 95% CI, 63%-90%).
Conclusions Our analysis of the current available literature suggests that there is no single element in the history or physical examination that is highly sensitive by itself for the diagnosis of erythema migrans. Clinicians should be aware of the wide variability in the clinical presentation of erythema migrans and the need to factor in multiple components of the clinical examination and epidemiological context.