Human Granulocytic Ehrlichiosis: A Case Series from a Medical Center in New York State

Methods

Patients

Patients were evaluated at the Westchester County Medical Center in 1994 and 1995; one patient (patient 18) has been discussed in another report [6]. At the first visit, samples of blood were obtained for complete blood cell count, differential count, chemistry profile, buffy-coat smear for examination of morulae, polymerase chain reaction (PCR) assay for HGE, and serologic examination to test for presence of antibodies to Ehrlichia equi and E. chaffeensis. Blood samples were also collected during the first to sixth weeks of convalescence.

Laboratory Testing

Buffy-coat smears were stained with Wright stain, and at least 1000 granulocytes were microscopically examined for Ehrlichia morulae. The presence of antibodies was tested for by indirect immunofluorescent antibody assay using horse granulocytes infected with E. equi at an initial serum dilution of 1:80 as described elsewhere [1]. We tested for antibodies to E. chaffeensis by using immunofluorescent antibody assay with infected DH82 cells [1, 7]. Peripheral blood was tested by PCR using EDTA-anticoagulated blood and the primer pair GE9f/GE10r as described elsewhere [8]. Samples of EDTA-anticoagulated blood from two patients (patients 14 and 15) were inoculated into an HL-60 cell line for culturing as described elsewhere [9].

Statistical Analysis

The two-tailed Student t-test, Pearson correlation (SPS software, Chicago, Illinois), and two-tailed Fisher exact test (EPISTAT software, Richardson, Texas) were used for data analysis.

Results

Eighteen patients received a diagnosis of HGE (3 in 1994 and 15 in 1995) in the Westchester County Medical Center. Most of these patients (83%) presented between June and August, Sixteen patients resided in Westchester County, and 2 resided in Putnam County; these counties are contiguous and located slightly north of New York City.

Thirteen patients (72%) reported that they had been bitten by a tick: Four ticks were positively identified as Ixodes scapularis (three nymphs and one adult tick). The other nine patients who reported a tick bite described a tick whose appearance was compatible with that of I. scapularis.

Symptoms appeared an average of 5.5 days after patients received the tick bite (Table 1). Two patients, neither of whom recalled being bitten by a tick, described a rash that they characterized as erythematous, pruritic, localized, and smaller than 5 cm in diameter. The described rash was not visible when these patients first visited the medical center. Patients were evaluated an average of 4.7 days after onset of symptoms (median, 3 days [range, 1 to 15 days]). Five patients (28%) required hospitalization for an average of 5 days. One patient (patient 15) was briefly admitted to an intensive care unit. None of the 18 patients died.

Laboratory test results frequently showed leukopenia or thrombocytopenia and abnormal liver enzyme levels (Table 2). Lymphopenia (defined as <1.5 × 10³ lymphocytes/mL) was seen in 8 of 13 (62%) patients, and neutropenia (defined as <2 × 10³ neutrophils/µL) was seen in 3 of 13 (23%). Lymphopenia was seen in 8 of 9 (89%) patients who presented with an illness that lasted 4 days or less and in none of 4 patients with an illness that lasted 5 days or longer (P = 0.006). Neutropenia was seen in 3 of 4 (75%) patients who had HGE for longer than 4 days but in only 1 of 9 (11%) patients who had the disease for 4 days or less (P = 0.05). Pearson correlation analysis showed that the initial neutrophil count was inversely related to the duration of symptoms before the first visit (r = −0.58[95% CI, −0.851 to −0.019];P = 0.038) and that the lymphocyte count was directly related to the duration of symptoms before the first visit (r = 0.543 [CI, −0.033 to 0.836]; P = 0.055). Patients who had HGE for longer than 4 days usually had atypical lymphocytosis of at least 10% of the differential count. No patients had elevated serum creatinine or blood urea nitrogen levels, and only one patient had a low hemoglobin level. Leukocyte and platelet counts returned to normal during observation in all patients who received treatment; these values generally returned to normal earlier than did liver enzyme levels.

Intracytoplasmic morulae were seen in 3 of 12 patients (25%). The frequency of infected granulocytes ranged from 0.3% to 6% (Table 2), and morulae were mostly seen in neutrophils (patient 15 had morulae in eosinophils). The presence of morulae correlated with age: Three of 4 patients older than 50 years of age and 0 of 8 patients younger than 50 years of age had morulae that could be detected. Results of polymerase chain reaction assay for HGE were positive in 9 of 12 patients (75%). All but 1 patient developed antibodies to E. equi. Three patients showed antibodies to E. chaffeensis in convalescent-phase specimens that had a higher titer (≥ 2560) to E. equi. The HGE agent grew in the HL-60 cell line from the blood of 2 patients (patients 14 and 15) who presented with HGE in the autumn of 1995.

All 18 patients received doxycycline therapy, beginning an average of 13.2 days (median, 6 days [range, 1 to 78 days]) after the onset of symptoms (Table 1). Seven patients were initially treated with other antimicrobial agents. The therapy of 2 patients was changed from amoxicillin to doxycycline within 2 days of treatment. The other 4 patients, who were initially treated with amoxicillin, had recurrent fever or persistence of symptoms after 1 to 2 weeks of therapy; for these patients, therapy was changed to doxycycline. Symptoms resolved in all patients 24 to 48 hours after the start of treatment with doxycycline, which was administered for an average of 14.2 days (range, 1 to 21 days). No symptoms have been seen after follow-up as long as 1 year.

Discussion

We describe 18 adult patients with HGE from the northeastern United States. Symptoms in these patients developed an average of 5.5 days after a presumed or proven bite from an Ixodes tick and frequently included fever (body temperature >38.3 °C), arthralgia and myalgia, and headache. Compared with the patients from the midwestern United States studied by Bakken and colleagues [1, 2], our patients were significantly younger (mean age ±SD, 47 ± 16 years compared with 57 ± 20 years; P = 0.04) and presented with a milder illness based on clinical and laboratory evaluations. Compared with the patients in our study, more midwestern patients were hospitalized (54% of midwestern patients compared with 28% in our study) and died (5% of midwestern patients compared with 0% in our study). Differences in case referral or in the threshold for diagnostic testing may explain some of these disparities.

Although HGE has a nonspecific clinical presentation, abnormal results on certain routine laboratory tests considerably narrow the differential diagnosis. A diagnosis of Lyme disease may be considered in a patient with fever after a bite from I. scapularis, but cytopenia is extremely rare in that disease [10]. In contrast, approximately 80% of our patients with HGE had leukopenia or thrombocytopenia. The explanation for these abnormal hematologic values and for abnormal liver enzyme levels is not yet known. Leukopenia can be caused by lymphopenia or neutropenia, with lymphopenia in the early stages of infection, followed by lymphocytosis with atypical lymphocytes.

Among the methods for specifically diagnosing HGE, microscopic detection of morulae is the quickest; for our patients, however, this method lacked sensitivity. Morulae were seen in 3 of 12 patients (25%) in our series compared with 80% of the midwestern patients. Visualization of intragranulocytic inclusions is probably an indication of massive infection and has been described as a risk factor, along with older age, for more serious illness [2]. In our series, PCR was the most sensitive diagnostic test during the acute phase. Nine of 12 (75%) of our patients had positive PCR results compared with 43% in the midwestern series [2]. As suggested by Bakken and colleagues, the lower sensitivity may have been caused by poor preservation of specimens and conditions that were suboptimal for the PCR assay.

Another diagnostic approach that offers definitive evidence of infection is isolation of the organism in culture [9]. In our study, culture was done for two patients in whom morulae were seen in peripheral blood granulocytes; Ehrlichia was isolated in both cases. However, this technique is probably not easily done in most clinical laboratories, and its sensitivity has not yet been determined.

Detection of antibodies to E. equi by immunofluorescent antibody assay is useful for confirming the diagnosis of HGE, but often only in retrospect. During the acute phase of HGE, the results of immunofluorescent antibody assay were positive in only 29% of our patients and 23% of the patients in the midwestern cohort. Current serologic assays to detect antibodies to Ehrlichia, however, lack standardization. Although all but one patient (94%) in our series developed antibodies to E. equi, the sensitivity and specificity of serologic testing have not been defined. Because Bakken and colleagues [2] reported that E. equi antibodies persist for longer than 1 year in nearly 50% of persons with HGE, some background seroreactivity probably exists in highly endemic areas; this possibility should therefore be considered when patients are evaluated. Another potential source of diagnostic confusion is cross-reactivity between E. equi and E. chaffeensis (the agent of human monocytic ehrlichiosis). Three (17%) of our patients had antibodies to E. chaffeensis (2 in low titer) as did 3 (8%) of 40 midwestern patients [2] and another patient with HGE in Massachusetts who was described in another study [4]. The practitioner must be aware that human monocytic ehrlichiosis and HGE are caused by distinct pathogens and that distinct diagnostic tests must therefore be requested for diagnosis to be confirmed or excluded.

Patients with HGE respond clinically within 48 hours of treatment with tetracyclines; empirical treatment may thus be warranted when strong clinical suspicion exists. The optimal duration of therapy has not yet been established, but short courses may be as effective as longer ones. Because of intolerance to the drug, one of our patients received doxycycline for only 1 day, but clinical and laboratory evidence of infection nonetheless resolved. In the studies by Bakken and colleagues [1, 2], the symptoms of six (15%) patients completely resolved without specific treatment. Many cases of HGE may be subclinical and self-limiting, a phenomenon that has been well documented with human monocytic ehrlichiosis [11, 12]. Tetracyclines are also excellent therapy for early Lyme disease, which is particularly important when Lyme disease alone, or as a coinfection with HGE, cannot be excluded [13]. Our case series either represents a milder form of illness caused by the same pathogen that was described in the midwestern United States or reflects the clinical disease caused by a closely related organism that has different virulence properties. Only when diagnostic testing becomes more widely available and more reliable and when large patient samples are systematically studied will the true incidence and clinical spectrum of this potentially serious illness be understood.

Drs. Horowitz, Wormser, and Nowakowski: Division of Infectious Diseases, Westchester County Medical Center, Macy Pavilion 209 SE, Valhalla, NY 10595.

Ms. McKenna: Main Nursing Office, Westchester County Medical Center, Valhalla, NY 10595.

Dr. Munoz: New York Medical College, Department of Pediatrics, Munger Pavilion Room 113, Valhalla, NY 10595.

Dr. Dumler: Division of Medical Microbiology, Department of Pathology, Meyer B1-193, The Johns Hopkins Medical Institutions, 600 North Wolfe Street, Baltimore, MD 21287.