A Clinical Score Sheet for Lyme Arthritis?
ABSTRACT & COMMENTARY
Synopsis: Use of a clinical score based on clinical presentation and history was helpful in distinguishing Lyme arthritis (LA) from other causes of arthritis, reducing unnecessary and expensive testing.
Source: Huppertz HI, et al. Diagnosis of pediatric Lyme arthritis using a clinical score. Eur J Pediatr 1998;157: 304-308.
In a national, prospective study in germany, 186 children in whom their attending physicians had considered the diagnosis of Lyme arthritis (LA) were examined. Ultimately, LA was confirmed in 87 patients. These were compared with the remaining 99 children in whom arthritis was attributable to other causes. In comparison to patients with other causes of arthritis, patients with LA had a higher frequency of episodic arthritis and initial knee joint arthritis, reported tick bites more frequently, were older, had a lower frequency of initial arthralgias, and had fewer large joints involved. A score was developed in the first of these patients and tested in a second group. The score was based on findings including: episodic arthritis arthralgia prior to onset of arthritis, age, initial arthritis of knee joint, history of tick bite, and number of involved large joints. Use of the score allowed patients with LA to be distinguished from those with other causes of arthritis within a score range of -7.0-12.0 points. A score of 2.5 or less excluded LA, whereas 6 or more points were highly indicative of LA. If only those children with a score result between 2.5 and 6 had been tested for antibodies to Borrelia burgdorferi, the number of ordered tests for LA would have been reduced by 63%. Huppertz and associates conclude that analysis of clinical presentation and use of a clinical score may help in distinguishing LA from other causes of arthritis and, thus, reduce unnecessary and expensive testing and uninterpretable test results.
COMMENT BY EUGENE D. SHAPIRO, MD, FAAP
Huppertz et al developed a convoluted scoring system based upon findings in an initial group of patients, in whom LA had been considered to be a possible diagnosis. The use of the score in a second validation group of similar patients with arthritis was useful in predicting which of the children had arthritis due to Lyme disease, rather than to another cause. Huppertz et al indicate that the score could be used to eliminate unnecessary diagnostic tests for LA in patients who were unlikely to have Lyme disease.
The number of patients studied was small and the study was conducted in Europe (where Lyme disease is caused by a strain of B. burgdorferi that is different from the strain that causes disease in the United States), so it is not clear that the results of the study can be generalized to American children. I question whether this kind of score could really have much of an effect on the ordering of tests that are unnecessary. Even if the clinical score correctly identified all of the children who did not have Lyme disease among the 186 children with arthritis in the study, only 99 children would not have the test ordered in this approximately six-year, multi-center study.
There is no doubt that Lyme disease is greatly over-diagnosed and that the unnecessary use of diagnostic tests for Lyme disease is a major factor in over-diagnosis. However, the major problem is that the tests are ordered in patients (such as those with only nonspecific symptoms such as arthralgia, headache, or fatigue) with a low probability of having Lyme disease. If a patient in an endemic area has objective evidence of subacute arthritis (as opposed to just arthralgia), particularly if the knee is affected, the "prior probability" (i.e., the probability based on the clinical history and exam prior to knowing the result of a diagnostic test for infection with B. burgdorferi) that the patient has Lyme disease will be high (perhaps as high as 50%). In this setting, the positive predictive value of a positive test is high, so obtaining a diagnostic test for Lyme disease clearly is indicated.
By contrast, in patients with only nonspecific complaints, the prior probability of Lyme disease is low (< 1%). Consequently, the positive predictive value of a positive test in this clinical setting is poor (< 10%), so the test should not be ordered.1 The improper use of diagnostic tests for Lyme disease in screening patients with a low probability of disease accounts for most of the tests that are inappropriately ordered (and for much of the erroneous publicity about the horrors of Lyme disease). It is this practice that should be avoided. Although Huppertz et al's scoring system fails to include a history of a characteristic rash (erythema chronica migrans), many of the features included in their scoring system should alert the clinician to a strong possibility of Lyme disease. (Dr. Shapiro is Professor of Medicine, Director of the Pediatric Lyme Disease Clinic, Yale University School of Medicine.)
Reference
1. Seltzer EG, Shapiro ED. Pediatr Inf Dis J 1996;15: 762-763.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content