Confirmed and probable cases of human granulocytic ehrlichiosis, by month and year of onset, in New York during 1994 and 1995. A total of 64 cases were identified. The date of onset of illness was not reported for 1 case.
Percentage of cases of human granulocytic ehrlichiosis (HGE), with onset of illness during 1994 and 1995, and Lyme disease, with onset of illness during 1994, by patient age, in New York.
Wallace BJ, Brady G, Ackman DM, Wong SJ, Jacquette G, Lloyd EE, Birkhead GS. Human Granulocytic Ehrlichiosis in New York. Arch Intern Med. 1998;158(7):769-773. doi:10.1001/archinte.158.7.769
Human granulocytic ehrlichiosis (HGE), a potentially fatal tick-borne disease, was first described in the upper Midwest in 1994. Following reports of suspected cases of ehrlichiosis from New York physicians, descriptive and case-control studies were conducted to characterize the epidemiology and risk factors for HGE in New York residents.
Descriptive data were gathered from surveillance and laboratory reports and hospital records. A confirmed case was defined as either (1) a 4-fold change in total antibody titer to Ehrlichia equi by indirect immunofluorescence or (2) a polymerase chain reaction assay positive for Ehrlichia phagocytophila/E equi group DNA. A probable case was defined as an acute febrile illness and either (1) a single E equi titer greater than or equal to 80 or (2) morulae on a peripheral blood smear. The case-control study included patients with confirmed HGE 18 years of age or older with the onset of disease in 1995 and 2 to 3 neighborhood-matched controls.
During 1994 and 1995, the New York State Department of Health, Albany, received reports of 241 residents who were tested for HGE; 30 met the confirmed case definition and 34 met the probable case definition. The median age of patients was 46 years (age range, 9-90 years), 35 (55%) were male, and 25 (45%) were hospitalized. Fever, headache, malaise, and myalgia were the most frequently reported symptoms. Fifty-six (88%) of the 64 patients resided in areas in which Lyme disease is hyperendemic. In the case-control analysis, cases were more likely than controls to have sustained a tick bite during 1995 (matched odds ratio, 5.0; 95% confidence interval, 0.9-49.8). Cases and controls did not differ by occupational exposure to ticks, underlying chronic diseases, or measures taken to prevent tick bites.
This study, which, to our knowledge, is the first population-based study of HGE, demonstrates the recent recognition of HGE in the state of New York. Control measures should be integrated with those for Lyme disease and should focus on minimizing contact with ticks and obtaining early treatment for infection.
IN 1994, A NEW type of human ehrlichiosis was described in 12 patients from the upper Midwest.1 This illness differed from previous cases of human ehrlichiosis reported in the United States2- 6 in that the ehrlichiae infected neutrophils rather than mononuclear leukocytes. The terms human monocytic ehrlichiosis (HME) and human granulocytic ehrlichiosis (HGE) were used to distinguish the 2 diseases.7 Since 1994, cases of HGE have been reported with increasing frequency, primarily from the upper Midwest and the northeastern states.8- 12 Published reports have included referral-based case series and individual case reports, but thus far no population-based surveillance data have been available. We report the results of population-based surveillance for HGE in New York, including a descriptive study and risk factor analysis of the first 64 cases of HGE identified in the state.
During the summer of 1994, physicians in Westchester County, New York, reported several cases of possible human ehrlichiosis to the Westchester County Department of Health (WCDH). Although the diagnoses of the cases were inconclusive, the WCDH informed primary care physicians in the county of the possible presence of human ehrlichiosis in the region. In the early summer of 1995, additional suspected cases were reported to the WCDH, and several of these cases were confirmed as HGE by serologic tests and/or polymerase chain reaction (PCR) assay.11
The New York State Department of Health (NYSDOH) notified all local health units throughout the state of the presence of human ehrlichiosis in New York and requested that suspected cases be reported to the NYSDOH. The department's laboratory, the Wadsworth Center, offered free serologic testing for HME and HGE. Hospitals, large group practices, and physicians in Westchester County were contacted directly and asked to report suspected cases using a standard form.
A confirmed case of HGE was defined as an acute febrile illness in a resident of New York who had either (1) a 4-fold change in antibody titer by indirect immunofluorescence assay (IFA) against Ehrlichia equi or (2) a PCR assay positive for E phagocytophila/E equi group DNA. A probable case was defined as an acute febrile illness in a resident of New York who had either (1) a single E equi titer greater than or equal to 80 or (2) neutrophilic morulae identified on peripheral blood smear.
Diagnostic specimens were evaluated for human ehrlichiosis at either the Wadsworth Center or the University of Maryland, Baltimore. Serum samples submitted to the Wadsworth Center were tested for antibodies to E equi and Ehrlichiachaffeensis using IFA methods.8 The E equi slides were prepared by harvesting peripheral blood neutrophils from an acutely infected horse (the E equi antigen slides were purchased from John Madigan, DVM, University of California, Davis). The E chaffeensis slides were prepared by using infected DH8M2 dog histiocytes. The serum samples were screened at a 1:80 dilution with a polyvalent fluorescent conjugate, and the specimens that were positive at a titer of 1:80 were retested in a 4-tube titration to a titer of 1:640. The specimens that were still positive at 1:640 were reported as "≥640." The specimens submitted to the Wadsworth Center were also tested for Borrelia burgdorferi antibodies by enzyme-linked fluorescent immunoassay and confirmatory immunoblotting; the results of the screening tests and immunoblots for Lyme disease are published elsewhere.13 The specimens submitted to the University of Maryland were tested by IFA and/or PCR as described previously.8 Peripheral blood smears were examined for neutrophilic morulae by some of the hospital laboratories collecting the specimens; these results were also reported to the NYSDOH.
A case-control study was conducted to determine the risk factors and exposures associated with HGE. Cases included patients 18 years of age or older who had confirmed HGE with onset of illness during 1995. Eligible cases were those reported to the NYSDOH before December 1, 1995. For each case, 2 to 3 neighborhood-matched controls were contacted (by using single-digit addition and subtraction from the last digit of the case telephone number) until eligible controls were identified. Controls were restricted to persons 18 years of age or older. Cases and controls were interviewed by telephone using a standard questionnaire and asked about tick exposures (eg, occupational, recreational, and animal exposures), measures used to prevent tick bites, and underlying medical conditions. In addition, cases were interviewed about presenting and residual symptoms.
Descriptive statistics were calculated using Epi Info software.14 All odds ratios and confidence intervals were calculated using methods designed for matched data. Crude odds ratios and the 95% confidence intervals were computed using StatXact software.15 Variables that were potential confounders (eg, age and sex) and variables that approached statistical significance in the univariate analysis were examined with conditional logistic regression analysis using LogXact software.16 Age was categorized into quartiles (ie, 18-37, 38-50, 51-63, and 64-90 years), with persons aged 18 to 37 years serving as the referent group. Because the dataset was small, exact methods were used in the analyses.
During 1994 and 1995, 30 patients had confirmed HGE and 34 had probable HGE. Fourteen patients had the onset of the disease during 1994, and 50 had the onset during 1995. The median age of patients was 46 years (age range, 9-90 years), and 35 (55%) were male. Cases were reported from 11 (18%) of 62 counties in New York; 45 patients (70%) resided in Westchester County, the county immediately north of New York City. Most patients (88%) resided in areas in which Lyme disease is hyperendemic. In 49 cases, the onset of illness was reported to occur in June through August (Figure 1). Of the patients with HGE for whom symptoms were reported, 100% (62/62) had fever, 93% (42/45) had malaise, 83% (40/48) had myalgias, 78% (40/51) had headache, and 69% (29/42) had arthralgias. Fever was defined by the reporting physician. Only 10 patients (16%) reported a rash. Descriptions of the rash were nondistinctive and included pustular or papular lesions, often at the site of a tick bite; a "flat, erythematous" rash on the abdomen; and an 8- to 10-cm "circular, raised, erythematous, nonpruritic" rash on the wrist.
Of the 56 patients for whom hospitalization status was known, 25 (45%) had been hospitalized for a median duration of 4.5 days (range, 1-47 days). Medical records of 20 (80%) of the 25 hospitalized patients were reviewed. Human granulocytic ehrlichiosis was considered as one of the possible diagnoses for 8 patients (40%) at the time of hospital admission. Other initial diagnoses included viral syndrome, heat stroke, sepsis, Lyme disease, cerebrovascular accident, pneumonia, and fever of unknown origin. The hospital courses of the patients included gastrointestinal bleeding (3 patients), toxic encephalopathy (2 patients), first-degree heart block (1 patient), and rhabdomyolysis (1 patient). Five (25%) of the 20 patients were admitted to intensive care units, and 1 required mechanical ventilation. Respiratory symptoms developed in 5 patients, including adult respiratory distress syndrome in 2 patients. One 9-year-old boy who had a probable case of HGE (ie, single titer of 1:80) died of multiorgan failure. Of the patients with HGE for whom clinical laboratory findings were available, 77% (43/56) had a white blood cell count lower than 5×109/L and 71% (37/52) had a platelet count lower than 150×109/L (Table 1). Most patients also had mildly increased serum liver enzyme levels.
Twenty-nine (45%) of the persons with confirmed and probable cases of HGE reported having had a tick bite during the month preceding the onset of illness. The median duration between the tick bite and the onset of symptoms was 6 days (range, 0-21 days). Seven patients identified the tick: 6 as a deer tick and 1 as a "wood tick." Of the 6 patients reporting a deer tick bite, 3 stated that the ticks were nymphs, 2 stated that the ticks were adults, and 1 was uncertain.
Detailed clinical information was obtained from the 19 cases interviewed for the case-control study. Among the 19 cases, the median number of days after the onset of symptoms and until medical care was sought was 2 days (range, 0-7 days); the median number of days until any antibiotic treatment was initiated was 2 days (range, 1-7 days); and the median number of days until doxycycline treatment was initiated was 3 days (range, 1-21 days).
During 1994 and 1995, a reported total of 241 New York residents were tested for HGE. Of the 30 confirmed HGE cases, 18 were confirmed by a 4-fold rise in titer, 8 by positive PCR assay results, and 4 by both. Three patients with confirmed HGE cases also had characteristic morulae in granulocytes on a peripheral blood smear. Not all patients were tested serologically and by PCR assay; however, 7 persons with confirmed cases who had elevated antibody titers had negative results on PCR assay, and 4 persons with confirmed cases who had positive PCR assay results had negative results on serologic testing.
Nineteen (30%) of 64 patients whose illness met the confirmed or probable HGE case definitions also had elevated E chaffeensis antibody titers. In 13 (68%) of these patients, the E chaffeensis titers were lower than the E equi titers. Five patients had the same level of E chaffeensis and E equi antibody titers, and 1 patient had an E chaffeensis titer higher than the E equi titer. Additional testing on a subset of patients by immunoblot analysis demonstrated that serum samples that were reactive to both E equi and E chaffeensis represented infection with E equi.13
Twenty-two patients were eligible for the case-control study, but 2 patients declined and 1 could not be contacted. Of the 122 eligible controls contacted by telephone, 53 (43%) agreed to complete the survey. Cases and controls were similar in age and sex distribution (mean age of cases, 45 years [range, 22-90 years]; mean age of controls, 53 years [range, 18-90 years]; P=.72). Twelve cases (63%) and 28 controls (53%) were male (P=.44).
In the unadjusted analysis, a history of 1 or more tick bites during 1995 and engaging in hiking or camping during the summer were significantly associated with HGE (Table 2). Cases also were more likely than controls to have a history of Lyme disease, to do yard work usually more than twice a week, to have a dog in the household, and to have seen ticks on their pets, although these differences were not statistically significant. Occupational exposures, underlying medical illnesses, deer sightings, and measures taken to prevent tick bites (eg, wearing long pants, wearing light-colored clothing, using insect repellent, and checking for the presence of ticks after outdoor activities) did not differ between cases and controls. In the conditional logistic regression analysis, sustaining a tick bite during 1995 and engaging in hiking or camping resulted in elevated odds ratios; however, the association was not statistically significant (Table 2). Yard work, dog ownership, and ticks on pets had odds ratios close to 1 and were dropped from the model.
Of the 18 cases and controls who reported having had 1 or more tick bites during 1995, 12 (67%) stated they were doing activities around their homes when the bite occurred. Recreational activities away from home accounted for 3 (17%) of the tick bites, and occupational exposure as a landscaper accounted for 1.
Cases and controls were questioned about how often they used specific measures to prevent tick bites when they were outdoors during the summer. As stated earlier, cases and controls did not differ in this regard and the following data are for cases and controls combined. Of all respondents, 45 (61%) reported wearing long pants most or all of the time; 39 (54%) reported wearing light-colored clothing; 36 (49%) reported checking themselves for ticks on returning indoors; 18 (25%) reported wearing long-sleeved shirts; 17 (20%), reported tucking their pant legs into their socks; and 13 (18%) reported using tick and insect repellent.
The emergence of HGE highlights the importance of advances in clinical laboratory science and the public health role in disease surveillance. During 1994 and 1995, a combination of active and passive surveillance detected 64 cases of HGE among New York residents. Most of these cases occurred among persons residing in the lower Hudson Valley and Long Island areas; both of these are areas in which Lyme disease is hyperendemic. The cases of HGE were characterized by influenzalike symptoms of fever, headache, malaise, and myalgias and were often accompanied by thrombocytopenia and leukopenia.
A case series of 41 patients with HGE from the upper Midwest has been reported by Bakken et al.8 The clinical presentations of the cases in the upper Midwest were similar to those in New York; however, the patients in New York were younger (median, 46 years [New York] vs 59 years [upper Midwest]) and included more women (49% vs 22%). Few of the diagnosed cases from the upper Midwest and New York have involved persons younger than 20 years: 2 (5%) of 41 patients in the upper Midwest and 4 (6%) of 64 in New York. This increased incidence among persons in older age groups is similar to New York's experience with another tick-borne disease, babesiosis, which has occurred mostly in older persons,17 but differs from that of confirmed New York cases of Lyme disease, 22% of which have occurred in persons younger than 20 years (Figure 2). Lyme disease and HGE presumably share the same vector, and the discrepancy between the ages of patients who have these 2 respective diseases raises the issue of whether children are less susceptible to HGE and/or have less severe, undiagnosed HGE infection than do adults.
Cross-reactivity between the HGE agent and E chaffeensis was an unexpected finding in serum specimens from patients with HGE residing in New York. Almost one third of the tested specimens from patients with HGE demonstrated reactivity to both agents in the IFA diagnostic test. This reactivity differed from the findings in the upper Midwest, where cross-reactivity in human serum to the ehrlichial agents was not detected. An immunoblot study of serum specimens from a subset of patients indicated that the infecting organism was E equi– like and not E chaffeensis.13 Geographic variation of bacterial strains, which is well documented for B burgdorferi, may occur for the Ehrlichia species and therefore account for the cross-reactivity observed in New York specimens.
An unadjusted analysis of the case-control study indicated that sustaining a tick bite during 1995 and engaging in hiking or camping were significant risk factors for HGE. The conditional logistic regression model did not produce significant results for the exposure variables at the .05 level; however, the study was limited by the small number of cases. Occupational exposures, underlying medical illnesses, preventive measures, and other recreational activities did not differ significantly between cases and controls. Similar studies of specific risk factors for Lyme disease have had mixed results. Some investigations have shown an increased risk associated with outdoor leisure activities,18 while other investigations have not.19,20 Risk factors associated with seroconversion for B burgdorferi included number of years at residence, rural residence, history of medical problems, and occupational exposures to ticks.21,22 In an outbreak of HME among residents of a golf community, an increased risk for infection was associated with tick bites, golfing, and never using insect repellent.23
Although yard work was not a statistically significant risk factor for HGE, many patients reported that they had sustained the tick bite while outside in their yards. This finding is consistent with those of several earlier studies of Lyme disease in New York, which indicated that most tick bites were acquired peridomestically.24,25 We did not distinguish between different types of yard activities (eg, gardening, clearing brush, and mowing the lawn), which may have shown specific activities to be a risk factor for HGE. The present study did not identify a protective effect of preventive measures against tick bites. Two explanations could account for this finding. First, the survey did not differentiate between preventive measures used while outside in the yard and during activities away from home. Many respondents reported taking precautions more often when away from home than when outside in their yards. Although cases and controls did not differ in the frequency with which they used preventive measures, the setting in which such measures were used might have differed. Second, although cases and controls were matched by neighborhood, the endemicity of HGE infection in ticks at the residences of cases and controls might have differed.
The investigation had several limitations. If persons who had sustained a tick bite were more likely to be tested for HGE than those who had not, a biased association between tick bites and disease could have occurred. Patients with HGE also may have been more likely to recall a tick bite than controls. This investigation represented population-based surveillance for HGE; however, some diagnostic and selection biases were likely present. The cluster of cases in Westchester County might have reflected the intense interest in HGE by local physicians, the media, and residents in the county. The human ehrlichioses were added to New York's reportable disease list in January 1996, and efforts are underway to increase surveillance activities. As disease awareness and surveillance increase, the geographic extent of HGE in New York will be better understood.
Human granulocytic ehrlichiosis has emerged as yet another tick-borne disease in New York. Issues remaining about HGE include its clinical presentation in children, geographical distribution of disease, and the development of standardized case definitions and diagnostic tests. While work in these areas continues, educational efforts and control measures should be integrated with those for Lyme disease and should focus on minimizing contact with ticks and early diagnosis and treatment of infection.
Accepted for publication July 31, 1997.
Ehrlichia chafeensis– infected DH82 canine histiocytes were provided by Jacqueline Dawson, PhD, Centers for Disease Control and Prevention, Atlanta, Ga.
We thank Rory Duncan, Lance Kingsley, and Reginald Taylor of the Wadsworth Center for their technical assistance and Hwa-Gan Chang for her statistical assistance.
Reprints: Barbara Wallace, MD, Bureau of Communicable Disease Control, New York State Department of Health, Room 649, Corning Tower, Empire State Plaza, Albany, NY 12237.