Physician Beliefs, Attitudes, and Approaches Toward Lyme Disease in an Endemic Area Stephen C. Eppes, M.D.1 Joel D. Klein, M.D.1 2 Gregory M. Caputo, M.D. Carlos D. Rose, M.D. 1

Summary: To assess the beliefs and practice habits regarding Lyme disease among practitioners, questionnaires were sent to physicians in a seven-county Lyme-endemic region. One hundred twenty-four evaluable responses were returned from 53 family physicians, 39 pediatricians, 27 internists, and five subspecialists who diagnosed three to four cases of Lyme disease per year, on average. The majority presented with erythema migrans (EM) or other early symptoms, although arthritis was the presenting sign in 16%. The enzyme-linked immunosorbent assay (ELISA) was the most frequently ordered diagnostic test, but 45% of respondents did not specify which test when ordering Lyme serology. The majority would use amoxicillin or doxycycline to treat EM in children or adults, respectively. Nearly all would use ceftriaxone for meningitis, and half would use it to treat Lyme arthritis or Bell’s palsy. Physicians differed markedly in the duration of therapy they would prescribe. Eighty-three percent would treat a patient for possible Lyme disease with antibiotics (many intravenously), even in the absence of EM or positive serology. Thirty-five percent of practitioners prescribed antibiotics for deer-tick bites. Our survey documents significant variation in approaches to Lyme disease among primary-care physicians and suggests the need for well-designed clinical trials, continuing basic research, and physician education.

Introduction

1 ymptomatic Borrelia

infection with

burgdorferi,

com-

monly known as Lyme disease, has become endemic in many areas of the United States. The 1

clinical manifestations of Lyme disease have been the subject of numerous papers in the medical literature and of frequent reports by the lay media. Physicians who practice in endemic areas are likely to be familiar with the classic find-

Division of Pediatric Infectious Diseases, Alfred I. duPont Institute, Wilmington, Delaware of Infectious Diseases, Department of Internal Medicine, Hershey Medical Center, Pennsylvania State University, Hershey, Pennsylvania

2 Division

Address correspondence to: Stephen C. P.O. Box 269, Wilmington, DE 19899

130

Eppes, M.D., Alfred I. duPont Institute,

in this illness.

However, many and symptoms signs are nonspecific, and Lyme disease is known to have varied presentations. Consequently, it can be a difficult diagnosis to make on clinical grounds. 1-3 Moreover, some serologic tests for B. burgdorferi lack optimal standardization and may lack sensitivity and/or specificity. 4-8 While guidelines for treatment have been published, 1,9,’0-&dquo; appropriate therapy may depend on patient-associated variables and the perceived

ings

severity of the disease. In receiving referrals to our Lyme disease clinic, it has been

our observation that practitioners differed widely in their approaches to the diagnosis and treatment of Lyme disease, perhaps as a result of the above uncertainties. In order to assess the beliefs, attitudes, and practice habits of physicians, we performed a survey in an area endemic for Lyme disease.

Materials and Methods

Most

physicians diagnosed beand eight cases of Lyme disease per year; 10% diagnosed none; and 6% diagnosed more than eight cases per year. The numbers of patients diagnosed with Lyme disease were similar for each subgroup of practitioners. Seventy-three percent of patients tween one

had EM or other symptoms of early disease (e.g., flu-like illness). Internists were more likely to see patients with acute neurologic signs ( 14% ) and chronic central nervous

system (CNS) A questionnaire was designed to evaluate physicians’ practices with regard to the number and nature of the patients they diagnosed with Lyme disease, the tests they were likely to order, the consultations they would obtain, and the treatment they would administer in several hypothetical situations. This questionnaire was mailed to 500 physicians in the three counties of Delaware; Salem County, New Jersey; Cecil County, Maryland ; and Chester and Delaware counties, Pennsylvania. This area is endemic for Lyme disease. 13 Our hospital database of area pediatri-

cians, internists,

family practitiongeneralists was used to generate the mailing list. One hundred thirty physicians returned the questionnaire during fall, 1991; six forms were incomplete and were not included in the analysis. Responses for the 124 completed questionnaires were analyzed according to medical specialty and as a group (Figure 1). ers, and

Results Of the 124 completed question-

naires, 53 were from family or general

practitioners, 39 pediatricians, 27 were nists, based

and five

were

were from from interfrom hospital-

subspecialists (Figure 2).

complaints or fatigue

(25%)

attributed to Lyme disease than were other practitioners. Arthritis

was

the

presenting sign

in

16% of patients, similar for all groups of practitioners. Cardiac involvement was rarely reported as the presenting feature of Lyme disease. Eighty-five percent of practitioners believed that the presence of EM was diagnostic of Lyme disease ; many also considered Bell’s palsy, recurring arthritis, and heart block to be diagnostic. Six percent felt that no physical findings could be considered pathognomonic of Lyme disease. To facilitate diagnosis, the majority of physicians (55%) obtained serum ELISA tests. Pediatricians were likely to also order serum Western blot assays for antibodies to B. burgdorferi. Urine antigen tests and antibody tests on cerebrospinal fluid were not commonly ordered. Fortyfive percent of physicians did not specify which testwhen they ordered

Lyme serology. Most physicians managed presumptive cases of Lyme disease without consultation. Twenty-one percent referred patients to infectious disease specialists; 10% to rheumatologists; and 7% to neurologists, orthopedic surgeons, or other specialists. Referral patterns did not differ among the groups of

practitioners. There was distinct variability in the choice of antibiotic and dura-

tion of therapy for several hypothetical patients with various manifestations of Lyme disease. Some internists did not respond to questions involving children, and some

pediatricians did not answer questions about adults. Eightyone percent of respondents EM in adults with oral doxycycline and, in children, with amoxicillin (with or without probenecid) or oral penicillin.

would

treat

in pregnancy often treated with penicillin or amoxicillin by the physicians who responded to this question. Most would treat a patient with multiple EM the same as for solitary lesions. Ninety-six percent would treat Lyme meningitis with intravenous (IV) ceftriaxone. half of Roughly respondents would treat Lyme-associated Bell’s palsy with IV ceftriaxone, as opposed to oral antibiotics; 28% of pediatricians would treat this manifestation in a child with IV ceftriaxone, while 74% of internists would use IV ceftriaxone for a similarly affected adult. For treating Lyme ar-

Erythema migrans

was most

thritis, respondents

were

equally

divided between oral doxycycline and IV ceftriaxone for adults and between amoxicillin and IV ceftriaxone for children. Few practitioners would treat Lyme disease in any stage for less than 2 weeks. Durations of 2, 3, and 4 weeks were chosen with almost equal frequency, except with arthritis, for which the longer courses were favored.

Eighty-three percent responded they would treat a patient for suspected Lyme disease even in the absence of EM or positive serology. Of these physicians, a quarter said they would also consider treating such a patient intravenously, some for as long as a month. Almost half of family and general practitioners would prethat

-

scribe

an

antibiotic for

a

known

131

Figure 1. Lyme disease questionnaire with responses of physicians (expressed

132

as a

percentage of those responding to questions).

frequency of these symptoms in adult patients. Conversely, adults with other neurologic diseases may receive the diagnosis of Lyme disease based on nonspecific epidemiologic evidence or falsely positive

generally sound approach to treatment of hypothetical cases of Lyme disease. Most physicians treated early Lyme disease manifestations in accordance with accepted rec-

irrelevant serum tests. 14 The vast majority of respondents correctly identified EM as a diagnostic feature of Lyme disease. It is noteworthy, however, that a substantial number of each group of physicians felt that Bell’s palsy, recurring arthritis, or heart block were also diagnostic. While each of these conditions in the appropri-

jority used IV ceftriaxone in patients with meningitis, as is considered appropriate. While published recommendations consider an oral regimen sufficient for treating isolated Bell’s palsy, many re-

or

epidemiologic setting may sugthe diagnosis, none is pathognomonic. A smaller number apparently believed that no clinical finding could be considered diagnostic of Lyme disease. ate

Figure 2. Breakdouvn of practitioners responding to Lyme disease survey. deer-tick attachment; 21 % of pediatricians and 30% of internists would provide an antibiotic in that situation.

Discussion The results of this study confirm our suspicion that practitioners’ approaches to Lyme disease

substantially. Physicians responding to the questionnaire

vary

cared for an average of three to four patients with a diagnosis of Lyme disease in the 1 year prior to the survey. This result may not have been representative of all physicians in this area, inasmuch as response to the questionnaire may have selected for physicians who were more likely to have an interest in the disease. As might be expected in primary-care settings, most patients were seen with early disease symptoms and were diagnosed and treated by their primarycare physicians. It is of interest that internists were more likely to see both acute and chronic neurologic complaints, as well as chronic fatigue ascribed to Lyme disease. This may reflect a relatively greater

gest

This is curious because, while EM may take several forms, the classic expanding lesion with central clearing is considered by most authorities to be pathognomonic. 3,15-18 Respondents varied in their choice of serologic tests, but almost half did not specify a particular assay when ordering Lyme serology. More pediatricians ordered the Western blot assay, a useful confirmatory test, perhaps owing to its ready availability at our children’s hospital. Few physicians indicated that they used antibody tests on cerebrospinal fluid (CSF) as an adjunct to the diagnosis of neuroborreliosis, despite the substantial number of patients with neurologic symptoms. It is possible that the variability in utilization of certain tests reflects a lack of familiarity with the technologies or with the relative sensitivities and specificities of the tests. On the other hand, physicians may be aware of shortcomings of certain serologic methodologies available for Lyme disease diagnosis, as well as interand intralaboratory variability. 17 We were impressed with the

ommendations.1,10-12 The

vast ma-

spondents, especially internists, chose ceftriaxone in that situation. Recent evidence from both Europe and the United States indicates that many such patients will often have concomitant CNS involvement, 19,20 which would suggest that a parenteral regimen might be warranted. Respondents were divided regarding the selection of antibiotic treatment for Lyme arthritis. This may reflect the individual styles of practitioners, but may also indicate a lack of confidence in oral therapy for arthritis. For arthritis that persists on oral therapy, IV antibiotics have been recommended, but well-designed trials comparing oral and IV therapy are lacking. There was considerable variability in the duration of therapy chosen by practitioners for all conditions described in the survey ; this may be due to the range of durations given in recommended treatment

regimens.

One of the most striking findings of the survey was the fact that the vast majority of physicians would treat a patient for Lyme disease, even in the absence of a firm diagnosis (i.e., EM or later clinical finding with positive serology), and that 25% would consider using IV antibiotics in that situation. This may reflect the respondents’ recognition that EM is not detected or remembered in half of cases of proven Lyme disease and that current serologic tests are sometimes not reliable. Delayed production

133

of antibody in early Lyme disease and interruption of the humoral response by antibiotic administration are two often-cited reasons for seronegative Lyme disease.21,22 Of greater concern is the possibility that inherent imprecision in the

diagnosis of Lyme disease may lead to

unnecessary

of some who do not

treatment

seronegative patients have Lyme disease.23 Patient, parent, and community pressure to diagnose Lyme disease in an endemic area may drive some physicians toward this diagnosis and its treatment, even in the absence of objective clinical and serologic evidence of infection with B. burgdor-

aware, and also suggests the need for physicians to remain current about this important public-health problem. Moreover, uncertainties about the diagnosis and management of Lyme disease, at the level of the practitioner, underscore the need for continued basic research and clinical trials.

proven

2.

3.

necticut.31

quite varied, and the results of the study supported the importance of educating the community about Lyme disease. Ours is the only such survey of physicians of which we are

134

4.

Lyme disease. 26-28

The wisdom of prophylactic treatment of tick bites with an antibiotic has been controversial. A recent study found a low rate of infection (1.2%) following confirmed deer-tick attachments and suggests that this practice is not routinely necessary.29 In our survey, less than half of physicians prescribed an antibiotic in this situation ; some who did cited a favorable cost-benefit ratio as the rationale. However, a model analyzing cost effectiveness of this approach concluded that empiric treatment of deer-tick bites was warranted only when the probability of infection in a given endemic area exceeds 3.5%.30 A recent survey concerning Lyme disease assessed the knowledge, attitudes, and behaviors of 200 heads of households in Con-

Responses

. 1991;40:217-221. 1989-1990. MMWR Berger BW. Cutaneous manifestations of Lyme borreliosis. Rheum Dis Clin North Am. 1989;15:627-634. 17. Eichenfield AH, Athreya BH. Lyme disease : of ticks and titers. Pediatr. 1989; J

114:328-333. 18.

REFERENGES 1.

Steere AC.

Lyme

N Engl J Med.

Lyme disease in children. J Pediatr. 1992;121:157-162. Corpuz M, Hilton E, Lardis MP, et al. Problems in the use of serologic tests for the diagnosis of Lyme disease. Arch In-

1364-1367. 21.

for

Schwartz BS, Goldstein MD, Ribeiro et al. Antibody testing in Lyme disease: A comparison of results in four laboratories. 1989; 262: 3431JAMA. 3434. Bakken LL, Case KL, Callister SM, et al. Performance of 45 laboratories partici-

Lyme

disease

serology. JAMA .

266:359. 26.

28.

Ceftriaxone-associated biliary complications of treatment of suspected disseminated Lyme disease. MMWR . 1993; 42:39-42. Nadelman RB, Zalmen A, Wormser GP. Life threatening complications of empiric ceftriaxone therapy for "seronegative Lyme disease." South Med J. 1991; 84:1263-1264. Feder HM, Rosenthal KE. Letter. JAMA.

29.

Shapiro ED, Gerber MA, Holabird NB, et

1992;268:891-895. 8.

9.

10.

Dattwyler RJ, Luft BJ. Immunodiagnosis of Lyme borreliosis. Rheum Dis Clin North Am. 1989;15:727-734. Luft BJ, Dattwyler RJ. Treatment of Lyme borreliosis. Rheum Dis Clin North Am. 1989;15:747-755. Treatment of Lyme disease. The Medical Letter.

11.

12.

1992;34:95-97.

Committee

Infectious Diseases. Treatment of Lyme borreliosis. Pediat. 1991;88:176-179. rics Recommendations for

were

treatment.

27.

264;693.

on

Rahn DW, Malawista SE.

al. A controlled trial of antimicrobial pro-

phylaxis for Lyme disease after deer-tick bites. N Engl J Med. 1992;327:1769-1773.

Lyme disease: diagnosis and

Ann Intern Med. 1991;114:

30.

472-481. 13.

Lyme

disease

—

United States, 1991-2.

L Jr. Lyme Disease and the Nervous System. New York, NY Thieme Medical

Reik

Magid D, Schwartz B, Craft J, et al. Prevention of Lyme disease after tick bites.

. 1993;42:345-348. MMWR 14.

1988;319:1441-1446.

Magnarelli LA. Laboratory diagnosis of Lyme disease. Rheum Dis Clin North Am. 1989;15:735-745. 23. Steere AC, Taylor E, McHugh GL, et al. The overdiagnosis of Lyme disease. JAMA. 1993;269:1812-1816. 24. Sigal LH. Summary of the first 100 patients seen at a Lyme disease referral center. Am J Med. 1990;88:577-581. 25. Caputo GM. Lyme anxiety. JAMA. 1991; 22.

pating in a proficiency testing program for

Dattwyler RJ, Volkman DJ, Luft BJ, et al. Seronegative Lyme disease: Dissociation of specific T- and B-lymphocyte responses to Borrelia burgdorferi. N Engl J Med.

tests

JM,

7.

20.

1991;151:1837-1840.

Luger SW, Kraus E. Serologic

Lyme disease: Interlaboratory variability. Arch Intern Med. 1990;150:761-763. 6.

Consensus conference on Lyme disease. Can . Med Assoc J 1991;144:1627-1632. Christen HJ, Bartlau N, Hanefeld F, et al. Peripheral facial palsy in childhood Lyme borreliosis to be suspected unless proven otherwise. Acta Paediatr Scand. 1990;79:1219-1224. Luft BJ, Steinman CR, Neimark HC, et al. Invasion of the central nervous system by Borrelia burgdorferi in acute disseminated infection. JAMA. 1992;267:

-

disease.

1989;321:586-596. Ostrov BE, Arhreya BH. Lyme disease: Difficulties in diagnosis and management Pediatr Clin North Am 1991; 38:535-553. Gerber MA, Shapiro ED. Diagnosis of

tern Med.

5.

Lyme disease surveillance: United States

16.

19.

feri.24,25 Unfortunately, significant adverse effects have occurred in a number of patients treated with intravenous antibiotics for un-

Publishers; 1991:107-108. 15.

31.

N Engl J Med. 1992;327:534-541. Lyme disease knowledge, attitudes, and behaviors-Connecticut, 1992. MMWR. 1992;41:505-507.