Ehrlichia, Anaplasma & Others

TAXONOMY Back to top

Members of the genus Ehrlichia and Anaplasma are now recognized to be important human

pathogens. They are obligate intracellular bacteria currently placed in

the Proteobacteria phylum (Alphaproteobacteria), orderRickettsiales, and

family Anaplasmataceae. Although most closely related to the

genera Rickettsia and Orientia,organisms classically considered ehrlichiae are divided into

four major clades. Taxonomic classification (Fig. 1) is largely based upon sequence analysis

of rrs (16S rRNA genes) and groESL (heat shock operon) (47, 72), but also that of citrate

synthase (gltA), the β subunit of RNA polymerase (rpoB), FtsZ protein (ftsZ), and other rRNA

genes (71–73, 80, 89, 130). Serologic cross-reactions, similarities among major

immunodominant surface proteins, and the cellular tropisms of these bacteria further support

the phylogenetic approach (43, 74). All members of the former

tribes Ehrlichieae and Wolbachieae are now included in the family Anaplasmataceaeinstead of

the family Rickettsiaceae. Table 1 delineates current, proposed, and former names of

selectedEhrlichia and Anaplasma species that are known human and veterinary pathogens.

All tick-borneAnaplasmataceae are grouped within two closely related

genera, Ehrlichia and Anaplasma. A. phagocytophilum,within the genus Anaplasma, now

includes Ehrlichia phagocytophila, Ehrlichia equi, and the human granulocytic ehrlichiosis

(HGE) agent. Minor sequence differences in rrs exist between these organisms, which could

reflect biological and ecological differences (33, 91, 127). Genetic variants of A.

phagocytophilum have been reported in ticks and mammals in the northeastern United

States and Europe, including some that probably do not cause disease in humans (91, 127).

DESCRIPTION OF THE GENERA Back to top

Ehrlichia and Anaplasma spp. are gram-negative obligate intracellular bacteria that reside

and propagate within membrane-lined vacuoles found in the cytoplasm of bone marrowderived

cells, such as granulocytes, monocytes, erythrocytes, and platelets. E.

ruminantium and several other species also infect endothelial cells (51, 135). These

intracytoplasmic clusters of bacteria resemble elementary bodies of Chlamydia as small

dense core forms (0.2 to 0.4 μm); larger forms (0.8 to 1.5 μm) resemble reticulate bodies

(110). Both are capable of binary fission. At least in some species, specific ligands associated

with cell adhesion are differentially expressed on the infective dense form but not on the

metabolically active reticulate form. After a few days, the elementary bodies dividing in the

phagosome form an inclusion, also called a morula, that can be seen

microscopically. Neorickettsia sennetsu grows as bacterial cells that can maintain individual

vacuolar membranes when they undergo binary fission. Cell lysis leads to the release of

bacteria that can infect other competent cells (22).

Unlike Rickettsia, Ehrlichia and Anaplasma do not have a thickened outer membrane leaflet.

The outer membrane appears more ruffled in A. phagocytophilum than in N.

sennetsu or Ehrlichia chaffeensis (111). By transmission electron

microscopy, Ehrlichia and Anaplasma spp. have a very limited region that corresponds to the

peptidoglycan layer, and genes necessary for its synthesis are not present within the

genomes of Ehrlichia, Anaplasma, or Neorickettsia spp. (21, 53, 58). The full genome

sequences of 15Anaplasmataceae members are established and range in length from 860 kb

in N. sennetsu and 1,176 kb inEhrlichia chaffeensis to 1,471 kb in A. phagocytophilum.

Genome sequencing has identified a low G+C content common to endosymbionts (except

for Anaplasma marginale) and a large proportion of noncoding sequences. Many of the genes

required for glycolysis are absent (53). Carbon sources are proline and glutamine. In

contrast to Rickettsiaceae, all enzymes necessary for the Krebs cycle and for metabolism of

purine and pyrimidine are present. ATP synthesis is possible, since these organisms possess

the ATP synthase complex and other enzymes required for aerobic respiration. However,

unlike Rickettsia, ATP/ADP translocases are absent. Genes for type IV secretion systems

most likely contribute to virulence and are present as they are in

other Alphaproteobacteria (99). In contrast to the reductive evolution seen in other

intracellular bacteria, these organisms have many pseudogenes, which document gene

duplication events (21, 53). The active duplication of tandemly repeated sequences likely

results in new genes, antigenic variation, immune evasion, and thereby enhanced survival.

Anaplasma marginale infects ruminants, and Aegyptianella spp. infect birds, amphibians, and

reptiles. These organisms also reside in small membrane-bound inclusions (19, 115) and are

not known to cause human disease.

EPIDEMIOLOGY AND TRANSMISSION Back to top

Ehrlichia and Anaplasma spp. are zoonotic agents transmitted to animals and humans by

ticks. Neorickettsiasennetsu rarely causes human disease and is likely acquired by ingestion

of fish infested with Neorickettsia-infected flukes (98). Wolbachia spp. are symbionts of a

broad range of arthropods, helminths, and crustaceans (131). Thus, most species can infect

vertebrates and invertebrates (Table 1). Transovarial transmission of these organisms in

ticks does not occur, but transstadial, interstadial, and intrastadial transmission

do.Ehrlichia spp. are acquired when the immature tick (larva or nymph) feeds on an infected

animal and are transmitted when the next stage (nymph or adult) feeds on another

mammalian host. Ticks and mammals, the latter with high-grade and/or persistent

bacteremia, may serve as reservoirs (83). In contrast, humans are only inadvertently

infected and represent an end-stage host. Therefore, maintenance of tick-borne ehrlichiae in

nature depends upon the presence of appropriate tick vectors and mammalian hosts in the

local environment (6, 84, 114, 132).

Recognized natural reservoirs for E. chaffeensis include deer (Odocoileus

virginianus and Blastocerus dichotomus), domestic dogs, and perhaps other animals that

host Amblyomma ticks (41, 82). Less important reservoirs include opossums, raccoons,

voles, coyotes, and goats (101). White-tailed deer (Odocoileus virginianus) also are a

reservoir of Ehrlichia ewingii and of another ehrlichial agent closely related to Anaplasma

platys, the white-tailed deer agent, which has not yet been associated with human disease

(8). The major reservoirs for A. phagocytophilum are incompletely investigated. However,

small mammals are frequent hosts of the immature stages of Ixodes scapularis or Ixodes

ricinus (132, 138). These include the white-footed mouse(Peromyscus leucopus) in the

eastern United States, chipmunks (Tamias striatus), voles (Clethrionomys gapperi),and wild

mice (Apodemus spp.). Although white-tailed deer can be persistently infected by A.

phagocytophilum,the A. phagocytophilum strains (AP variant 1) that naturally infect deer are

not infectious for small mammals, have not been identified in humans, and may represent

nonpathogenic variants (90, 96). Persistent or prolonged infection in animal reservoir hosts

is essential for maintenance of zoonoses. Mice infected with A. phagocytophilum can remain

infected for months, which contributes to transmission to different stages of developing I.

scapularis ticks (124). In Europe, red deer, sheep, cattle, and goats are persistently infected

and serve as reservoirs of A. phagocytophilum. The reservoir for N. sennetsu is not known;

however, epidemiological data suggest that consumption of raw fish is a risk factor for

sennetsu fever (98). Other species of Neorickettsia have complex transmission processes

involving trematodes. N. risticii, the agent of Potomac horse fever, is transmitted to horses

by accidental ingestion of insects carrying N. risticii-infected cercariae (85). Similarly, N.

helminthoeca infects dogs through ingestion of trematode-infested fish.

CLINICAL SIGNIFICANCE Back to top

Human Diseases

HME

The causative agent of human monocytic ehrlichiosis (HME) is E. chaffeensis, a

monocytotropic ehrlichia that was first identified as a human pathogen in a patient with a

severe febrile illness after tick bites in 1986 (86). More than 5,490 cases of HME were

reported to CDC through 2009; however, data from active surveillance efforts suggest that

HME occurs much more frequently than is reported (44, 50). The seroprevalence of E.

chaffeensis ranges from 1.3% to 12.5% in the regions of Arkansas and Tennessee in which it

is endemic. In contrast, active surveillance in Tennessee and Missouri identified 330 to 414

cases/100,000 population (44,50). Most cases are identified in the south-central and

southeastern United States, but increasingly infections are identified in the middle Atlantic

states. Prospective evaluation of heavily exposed cohorts shows that approximately 75% of

seroconversions are subclinical (57). Recent reports from Latin America, Africa, Europe, and

Asia indicate that E. chaffeensis, or closely related microorganisms, are also found there

(23, 59, 95, 106,117, 139). Using PCR to amplify ehrlichial rrs, E. chaffeensis nucleotide

sequences have been found in various tick species from different regions of China (29).

The median incubation period for HME is 9 days. The median age is 44 years, and 66% of

patients are males (42, 101). Patients often present with high fever (96%), headache

(72%), malaise (77%), myalgia (68%), and no localizing physical findings (50).

Gastrointestinal (nausea, vomiting, and diarrhea), respiratory (cough), and osteoarticular

(joint pain) symptoms are present in less than 50% of patients. Central nervous system

involvement (stiff neck, confusion, and meningitis), have been described (125). Petechial,

macular, and maculopapular rashes occur with varied distribution and onset (100, 101).

Rashes are more frequent in children (67% of cases) (121). Abnormal laboratory parameters

occur in at least 86% of patients and include thrombocytopenia (70 to 90%), leukopenia

(60%) with lymphopenia and/or neutropenia, and increased serum aspartate transaminases

(50%). Severe complications include meningoencephalitis and a toxic shock-like syndrome

with multiorgan failure, including adult respiratory distress syndrome. Fulminant infections

are more common in patients immunocompromised by HIV, high-dose corticosteroids, and

medications related to organ transplantation (56, 88, 102–104, 119, 126). The case fatality

rate is 2 to 3%. Male sex, advanced age, and an immunocompromised status are

independent risk factors for death (42).

HGA

The causative agent of human granulocytic anaplasmosis (HGA) is Anaplasma

phagocytophilum. HGA was first identified in 1990 in a patient from Wisconsin who reported

tick bites. Neither E. chaffeensis nor its tick vector was known to exist in that area (12, 33).

Although tick bite is likely the most frequent route of transmission, there are others

(30, 67, 148). Perinatal transmission has been reported (67), as has transmission by

accidental inoculation of infected blood and blood transfusion (30, 148).

HGA, like HME, is not a reportable illness in all states; thus, the true incidence and

prevalence of the infection are unknown. However, passive surveillance in northwestern

Wisconsin and Connecticut reveals yearly incidence rates of 24 to 58 cases per 100,000

population (44). As of 2009, more than 6,200 cases had been reported nationwide, with the

highest annual incidence rates in the Northeast and upper Midwest. Infected patients have

been identified in other states, as well as in several European and Asian countries

(9, 18, 38,106, 109, 113, 148). Most infections are likely subclinical, since between 0.6%

and 14.9% of the population in Connecticut and northwestern Wisconsin, respectively, are

seropositive (14, 69). The tick vectors for HME and HGA coexist in the middle Atlantic and

southern New England states, and in the southern Midwest. Thus, both diseases can occur in

these areas (32, 38, 42). Furthermore, human ehrlichioses, including HME, HGA, and E.

ewingii infections, are clinically indistinguishable.

HGA has a median incubation period of 5 to 11 days after the bite of an Ixodes sp. tick. The

median age is 43 to 60 years (1, 3, 16), and the male/female ratio is 2:1 (11). Patients most

often present with high fever (91%), myalgias (77%), headache (77%), and malaise (94%).

Gastrointestinal, respiratory, musculoskeletal, and central nervous system involvement

occurs in fewer patients. Rash is observed in 6% of patients, all attributable to erythema

migrans with concurrent Lyme disease (13). Leukopenia with lymphopenia,

thrombocytopenia, and increased serum aspartate transaminase activities are common early

in the disease and may normalize before antimicrobial treatment. Lymphocytosis with

atypical lymphocytes can occur after the first week of infection (11, 68). Severe

complications of HGA include a septic shock-like illness with multiorgan failure, adult

respiratory distress syndrome, and opportunistic infections (16, 63, 68, 81, 136). Meningitis

has not been documented. At least 6 deaths have been reported, 3 of which had

opportunistic infections including Candida esophagitis, Cryptococcus pneumonia, invasive

pulmonary aspergillosis, and herpes esophagitis (16, 63, 81, 136).

Ehrlichia ewingii Ehrlichiosis

E. ewingii was recognized to cause human infection when its DNA was found in the blood of

patients with an HME-like disease in Missouri (28). As with E. chaffeensis, the main vector is

the lone star tick (Amblyomma americanum); therefore, the distribution of the disease is

similar to that of HME. Evidence of E. ewingii inDermacentor variabilis ticks has also been

documented (140, 145). White-tailed deer are also reservoirs, and recent epidemiological

studies suggest that dogs may be as well (8, 145). Ewingii ehrlichiosis seems to affect

mostly immunosuppressed patients, including those with HIV, but is clinically milder than

coinfection with HIV and E. chaffeensis (28, 102). In dogs, E. ewingii is responsible for

canine granulocytotropic ehrlichiosis.

Sennetsu Ehrlichiosis (Neorickettsiosis)

Named after the Japanese term for glandular fever, Neorickettsia sennetsu was first isolated

from patients with suspected infectious mononucleosis in 1953 (94). It is rarely identified

now. Patients develop a self-limited febrile illness with chills, headache, malaise, sore throat,

anorexia, and generalized lymphadenopathy. Cases were identified in Japan and possibly

Malaysia, although at least one new case was recently recognized in Laos (98). Laboratory

findings include early leukopenia and atypical lymphocytes in the peripheral blood during

early convalescence. No fatalities or severe complications have been reported.

Other Human Ehrlichioses

In 1996, Ehrlichia canis was isolated from the blood of an asymptomatic man from

Venezuela. Since then, six additional symptomatic patients were identified as infected by

an E. canis strain that differs from canine strains by a single nucleotide polymorphism

in rrs (107, 108).

While Wolbachia spp. are not known to be directly pathogenic for humans or animals,

emerging evidence indicates a potential role for the intracellular bacterial components

(“symbionts”) of such helminths as Brugia malayi, Onchocerca volvulus, and Wuchereria

bancrofti as potentiators of the inflammatory reactions associated with parasitic infections

(65, 133).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

Currently, there are three methods for diagnosis of active infection with HME or HGA: (i) PCR

amplification of nucleic acids from Ehrlichia or Anaplasma species; (ii) detection of morulae

in the cytoplasm of infected leukocytes by nonspecific Romanowsky stains (e.g., Giemsa or

Wright) or by specific immunocytologic or immunohistologic stains, using E. chaffeensis or A.

phagocytophilum antibodies; and (iii) culture of Ehrlichia orAnaplasma from blood or

cerebrospinal fluid (CSF). In contrast, testing of single serum samples (acute or convalescent

phase) is rarely useful. However, concurrent testing of paired sera (obtained during acute

illness and in convalescence, 2 to 4 weeks later) does provide definitive retrospective

diagnosis.

EDTA-anticoagulated blood is a useful specimen for most tests (PCR, smears, and culture)

and should be obtained during the active phase of illness (1, 60, 100). Peripheral blood buffy

coat smears or cytocentrifuged preparations of CSF cells should also be obtained at the acute

phase and should be prepared within hours of obtaining the samples, since leukocytes

degenerate rapidly. Once prepared, air-dried blood smears and cytocentrifuged CSF

preparations are stable at room temperature for months or years.

Whole blood is the preferred sample for PCR since it contains infected leukocytes; serum

should not be used. Samples for PCR should be tested promptly or frozen at −80°C.

The preferred specimen for culture of Ehrlichia and Anaplasma is peripheral blood. Samples

should be obtained by sterile venipuncture or lumbar puncture and processed as soon as

possible. The culture conditions forEhrlichia and Anaplasma species are still being optimized.

Culture is currently performed only in a few public health and research laboratories. Samples

for culture should be maintained at approximately 4°C during shipping but not frozen. A.

phagocytophilum is easier to culture than E. chaffeensis, including from EDTA-anticoagulated

blood stored for up to 18 days at 4°C (75). Ehrlichia- and Anaplasma-infected cells can be

stored frozen within infected host cells at −80°C for months. Storage of infected cells is best

accomplished when more than 50 to 90% of the host cells are infected and is achieved by

suspension of at least 106 cells per ml in tissue culture medium that contains 10% dimethyl

sulfoxide and at least 30% fetal bovine serum.

LABORATORY CONFIRMATION OF EHRLICHIA

CHAFFEENSIS Back to top

Direct Examination

Microscopy by Romanowsky Staining of Peripheral Blood

Patients with suspected HME should have Romanowsky-stained (Giemsa or Wright stain)

peripheral blood or buffy coat leukocytes examined for the presence of ehrlichial morulae.

However, the sensitivity is low (≤29%) compared to culture (125). E. chaffeensis is detected

predominantly in monocytes and is more frequently detected in severe infection. When

present, morulae are small (1 to 3 μm in diameter) round-to-oval clusters of bacteria that

appear as basophilic to amphophilic stippling within cytoplasmic vacuoles with Romanowsky

stains (Fig. 2C). Detection of morulae is accomplished more frequently in

immunocompromised patients. Since the percentage of cells infected ranges from 0.2% to

10%, up to 500 cells should be examined.

Antigen Detection by Immunohistology

Immunohistochemistry may identify E. chaffeensis in bone marrow, liver, and spleen.

However, the sensitivity of detecting active infection with examination of bone marrow is

only 40% (49). A monoclonal antibody can specifically detect E. chaffeensis in human tissues

(146), but most studies use polyclonal antibodies that react with other Ehrlichia species.

Commercial assays are not currently available.

Nucleic Acid Detection Techniques

The most widely used method is PCR amplification of DNA from E. chaffeensis in clinical

samples using the HE1/HE3 primer set (7, 55, 125). This primer pair amplifies a 389-bp

fragment of rrs (16S rRNA gene). The product may be detected by simple nucleic acid

staining (e.g., ethidium bromide) after agarose gel electrophoresis, by Southern

hybridization of the amplified products using an internal probe, or by real-time PCR with

either 5′-nuclease or molecular beacon probes that increase analytical sensitivity (37). A

clinical evaluation of E. chaffeensis PCR using the HE1/HE3 system showed a sensitivity of 79

to 100% compared with serology; however, nucleic acids from E. chaffeensis were frequently

detected in patients who never developed antibodies (7, 125). Similar sensitivity was shown

with a nested PCR that employs broad-range “Ehrlichia genus” primers in an initial step

followed by PCR with the HE1/HE3 primer pair (77). A nested-PCR assay with broadrange

rrs primers (8F and 1448R) followed by a second (nested) reaction with primers 15F

and 208R on whole blood yielded results similar to those obtained with culture (125). Other

targets for PCR that have not been fully evaluated for clinical sensitivity or specificity include

the groESL operon, the outer-surface variable-length PCR target protein gene present in E.

chaffeensis (102, 125, 129), the 120-kDa antigen gene that encodes an immunodominant

antigen with tandemly repeated subunits that vary among E. chaffeensis isolates, the

quinolinate synthase A gene, nadA, the disulfide bond formation protein gene (dsb),and the

p28 multigene family (43, 102, 134, 147). In a prospective study, the overall sensitivity and

specificity of PCR were 56% and 100%, respectively, using the 16S rRNA subunit, nadA, and

120-kDa protein genes (100). However, in this study several samples had high titers of

antiehrlichial antibodies by immunofluorescence assay (IFA), suggesting that the pathogen

may have already been cleared (few circulating ehrlichiae). When sensitivity was calculated

using seroconversion as the gold standard, it increased to 84%. Posttest probabilities for a

positive and negative PCR result were 96% and 11.1%, respectively. Posttest probabilities

depend critically on prevalence (100).

Real-time multicolor PCR and real-time multiplex reverse transcriptase PCR assays have

been developed recently with extremely high analytical sensitivity and specificity comparable

to that of nested PCR (43, 123). Advantages include improved specificity (lower risk of

contamination), increased speed, lower cost, and the detection of multiple ehrlichial

pathogens simultaneously.

Isolation Procedures

Ehrlichia chaffeensis has been isolated from peripheral blood of a limited number of patients

with HME, and anE. canis-like organism was isolated only once from an asymptomatic human

(40, 48, 104, 108, 125). The most frequently used cell for primary isolation is the canine

histiocytic cell line DH82; however, E. chaffeensis has been successfully cultivated in other

cells including the human macrophage-like THP-1 cells, the fibroblast-like HEL-22 cells, Vero

cells, and HL-60 cells (human promyelocytic cell line differentiated to monocytic pathway),

among others (64). Isolation may be successful even when infected leukocytes are not

observed on peripheral blood examination (125). Isolation usually involves direct inoculation

of leukocyte fractions or whole blood into flasks with confluent layers of adherent cells or into

flasks that contain approximately 2 × 105 to 1 × 106 nonadherent cells per ml of tissue

culture medium. Macrophage-like cells that are highly phagocytic may be adversely affected

by the presence of erythrocytes; thus, it is recommended that either (i) leukocytes be

fractionated from erythrocytes by density gradient centrifugation (e.g., Ficoll-Paque); or (ii)

leukocytes be harvested after erythrocyte lysis (hypotonic lysis, NH4Cl lysis, etc.); or (iii) cell

confluency be reestablished after cultivation with erythrocyte-containing samples by addition

of uninfected host cells. Since E. chaffeensismay be present in few peripheral blood

leukocytes, it is advisable to inoculate cultures with as many peripheral blood leukocytes as

possible (which may be difficult with leukopenia). Use of 2 to 3 ml of EDTA-anticoagulated

blood diluted in 2 volumes of sterile Hanks’ balanced salt solution followed by Histopaque

(Sigma, St. Louis, MO) gradient separation of leukocytes has been effective (125).

The blood mononuclear cells are resuspended in a 2-ml volume of tissue culture medium

supplemented with 5% fetal bovine serum and allowed to interact with adherent host cells in

a 25-cm2 flask for 3 h, usually enhanced by incubation with rocking at 37°C in 5% CO2. The

inoculum is removed if significant erythrocyte contamination is present, and the monolayer is

replenished with 5 ml of fresh tissue culture medium. SinceEhrlichia species are bacteria,

antibiotics in the medium must be avoided. The generation time of E. chaffeensisis

approximately 19 h (22), and thus, cultures must be maintained to allow a slow logarithmic

or stable growth phase to avoid the host cells outgrowing the ehrlichiae.

Identification

The presence of infected cells is determined by sampling the medium (DH82 cells and THP-1

cells) or by lightly scraping part of the monolayer. Aliquots of the culture are cytocentrifuged

and then stained with Romanowsky or immunofluorescent stains, and cells are examined for

the presence of intracytoplasmic morulae or Ehrlichia chaffeensis antigen (Fig. 2A). Culture

may require one month or more but has been achieved in as short a time as a few days

(40, 48, 125). Confirmation of the infectious agent is currently best achieved by PCR

amplification using species-specific primers (5).

Serologic Tests

The gold standard for the diagnosis of HME is demonstration of a fourfold rise in

immunoglobulin G (IgG) titer or seroconversion by examination of paired (acute- and

convalescent-phase) sera. Thus, diagnosis is at best retrospective. The most frequently used

serologic method is the indirect IFA. Other methods have included enzyme immunoassays

(enzyme-linked immunosorbent assay [ELISA]) and protein (Western) immunoblotting.

Ehrlichial antigens may be difficult to prepare and are available mostly through public health

and research laboratories, although commercial production and distribution are now

available. Commercial sources of IFA serodiagnostic kits include Focus Technologies

(Cypress, CA), Scimedx Corp. (Denville, NJ), and PanBio Diagnostics (no longer available in

the United States).

Currently, there is little standardization for any method of ehrlichial serology, and cutoff

titers are dependent upon validation in individual laboratories that perform these assays. The

algorithm for serologic testing by IFA includes an initial screen at a dilution of 1:64 or 1:80

for IgG antibodies to E. chaffeensis. Reactive samples are then titrated to the end point.

Ehrlichia chaffeensis IFA

E. chaffeensis IgG is detected by IFA using E. chaffeensis Arkansas strain-infected DH82

canine macrophage-like cells. Reactive sera are serially diluted starting at a dilution of 1:64.

The presence of antibodies is detected after incubation with fluorescein isothiocyanateconjugated

anti-human IgG. The test is positive if classic intracytoplasmic morulae are seen.

It is important to identify the appropriate proportion of infected cells as determined by a

positive control serum and the appropriate morphology for each antigen preparation to

preclude false-positive interpretations. Prescreening for autoantibodies or routine removal of

rheumatoid factors will lessen the risk of misinterpretation due to antibodies reactive with

cellular components including nuclear or cytoplasmic antigens that could have the

morphologic appearance of morulae. A fourfold increase in IgG antibody titer or

seroconversion confirms the diagnosis of acute HME. A single specific IgG titer of ≥64, like

identification of morulae in monocytes or macrophages for E. chaffeensis by microscopy, is

suggestive. Antibody titers may be detected in a small proportion of subjects without HME

owing to the presence of antigens that are highly conserved among bacterial species

(39, 141). Acute-phase sera should be obtained at the time of presentation with acute

illness, and convalescent-phase sera are best obtained 3 to 6 weeks later (39).

The sensitivity and specificity of IFA for the diagnosis of infection with E. chaffeensis are not

known but are assumed to be high because of documented correlation between new or rising

antibody titers against E. chaffeensis and characteristic clinical findings (57). In the early

phase of infection, IFA testing is not sensitive compared to PCR (36). A systematic evaluation

of the usefulness of IgM testing has not been conducted, but a preliminary evaluation based

on nine culture-confirmed cases suggests that it might be slightly more sensitive than IgG

for the diagnosis of HME during the acute phase (36). Previously, the serologically crossreactive

E. canis was used as a surrogate antigen; however, this serodiagnostic assay has a

lower sensitivity than that obtained using E. chaffeensis, and its use should be discouraged

(7). The role of immunoblots in diagnosis is not well established; however, many patients

with E. chaffeensis infection can be differentiated from patients with HGA by the

demonstration of antibodies reactive with one or more of the 22-, 28-, 29-, 46-, 54-, or 120-

kDa antigens of E. chaffeensis (24, 34). Alternative methods based on recombinant proteins

show promise but are not commercially available (147). Antibodies to E. chaffeensis have

also been detected in patients diagnosed with Rocky Mountain spotted fever, Q fever,

brucellosis, Lyme disease, and Epstein-Barr virus infections, suggesting that false-positive

reactions do occur (39, 141). Antigenic diversity among E. chaffeensisisolates is well

described (35) but may not affect the detection of polyclonal antibody responses generated

with human infection. Several reports characterized patients with suspected HME who lacked

antibody responses, even long after onset of symptoms (55, 118, 125). However, in the few

cases in which E. chaffeensis infection was proven by isolation of the agent, patients who

survived developed clear convalescent serologic reactions by IFA (35, 40, 48, 102, 125).

Hypothetical reasons for false-negative results include infection by antigenically diverse

strains (unproven) and abrogation of antibody response by early therapy (125).

LABORATORY CONFIRMATION OF ANAPLASMA

PHAGOCYTOPHILUM Back to top

Direct Examination

Microscopy by Romanowsky Staining of Peripheral Blood

Examination of Romanowsky-stained (Giemsa or Wright stain) peripheral blood or buffy coat

leukocytes for the presence of morulae is highly valuable in the diagnosis of HGA. Usually,

800 to 1,000 granulocytes are examined under magnification of ×500 to ×1,000 for the

presence of morulae (1, 3, 16). Since most patients presenting with positive smears have

<1% of infected granulocytes and usually have leukopenia, buffy coat preparations yield

more than peripheral smears. Infection rates as high as 40% of granulocytes have been

described (12). As for HME, the presence of detectable infected granulocytes in peripheral

blood correlates modestly with severity of infection (3, 11, 12). The sensitivity of the buffy

coat smear examination in the acute phase of HGA is approximately 60% (11). When

present, ehrlichial morulae are small (1 to 3 μm in diameter) round-to-oval clusters of

bacteria that stain basophilic to amphophilic with Romanowsky stains (Fig. 2D). These

clusters are present in the cytoplasm of neutrophils or eosinophils and have a stippled

appearance owing to individual bacteria within the vacuole.

Immunohistology for Antigen Detection

Immunohistologic methods may also be used to identify A. phagocytophilum within human

tissues, including bone marrow, liver, and spleen (81).

Nucleic Acid Detection Techniques

Multiple PCR assays for detection of A. phagocytophilum nucleic acids have been published

(54, 62, 92, 123,128). Most utilize regions of rrs (16S rRNA gene) that are relatively specific

as targets for amplification. The most frequently applied and evaluated method employs the

primer set ge9f and ge10r, which amplifies a 919-bp fragment, most often used as a singlestage

reaction, with or without a hybridization probe to enhance sensitivity (33, 54). A

popular alternative is the use of nested PCR with an outer set of primers that anneal to and

amplify eubacterial 16S rRNA genes, followed by a nested internal PCR with A.

phagocytophilum-specific primers (92, 125). Amplification of the groESL region using a

nested PCR has also been useful to detect ehrlichial DNA in blood during the acute phase.

Primers HS1 and HS6 are used in the primary reaction followed by primers HS43 and HS45.

The size of the amplified product distinguishes E. chaffeensis from A. phagocytophilum, 528

bp versus 480 bp, respectively (128). The analytical sensitivity and specificity of several

published primer sets were evaluated by using DNA extracted from serial dilutions of A.

phagocytophilum-infected HL-60 cells (92). Specificity was evaluated using DNA extracted

from cultures of E. chaffeensis, Rickettsia rickettsii, and Bartonella henselae. The primer sets

with the greatest sensitivity and specificity were those used in a nested reaction to

amplify rrs, i.e., ge3a-ge10 and ge9-ge2, and those amplifying the msp2gene, i.e., msp2-3fmsp2-

3r (92). Both PCR assays detected as few as 0.25 infected HL-60 cell per μl of blood.

Additional recent methods employ real-time PCR and the 5′ nuclease (TaqMan) approach and

target the >100-copy msp2 gene family, which provides increased sensitivity to as low as 1

infected cell per μl of blood (122). A multiplex assay to detect Ehrlichia and Anaplasma spp.

by real-time reverse transcriptase PCR was developed and evaluated in peripheral blood of

dogs suspected of ehrlichiosis (123). The assay has a sensitivity of 100rrs transcripts, which

corresponds to about 1 infected cell in a test sample, can detect single or multiple infections,

and has the potential for automation.

Isolation Procedures

A. phagocytophilum has been successfully cultivated more often from human patients than

has E. chaffeensis,E. canis, or N. sennetsu, probably owing to the quantity of organisms

present in the peripheral blood of infected patients (1). HGA is described in Europe and Asia,

and several successes at isolating the organism have been reported. At least one human

isolate of A. phagocytophilum from outside the United States (Slovenia) has been verified

(61). Isolation is best achieved in the human promyelocytic cell line HL-60 (60) and has been

accomplished even when morulae are not observed in peripheral blood smears. The optimal

conditions for recovery of these bacteria have not been conclusively determined. Because

erythrocytes do not adversely affect HL-60 cells, direct inoculation of EDTA-anticoagulated

blood is effective. Fractionation of blood into buffy coat or granulocyte fractions by density

gradient centrifugation is also effective (116). Approximately 100 to 500 μl of EDTAanticoagulated

blood, containing 102 to 104 infected granulocytes, is inoculated into 100-fold

more uninfected HL-60 cells that are in the exponential growth phase. Cultures are

subsequently maintained at a concentration between 2 × 105 and 1 × 106 cells per ml of

tissue culture medium.

Identification

Cultures are examined every 2 to 3 days by Romanowsky staining of cytocentrifuged

preparations of 20 to 50 μl of culture suspensions. Anaplasma morulae appear as small

aggregates of basophilic bacteria in the cytoplasm of the HL-60 cells (Fig. 2B). Since HL-60

cells can contain a variety of cytoplasmic granules, immunocytochemistry or

immunofluorescence is very helpful for the inexperienced laboratorian. Unfortunately,

immunohistological reagents are currently not commercially available. Cultures usually

require between 5 and 10 days before morulae are clearly identified; but infected cells may

be detected as early as 3 days postinoculation. Time to detection of organisms in culture

correlates with the number of bacteria present in blood at the time of culture (75). Definitive

identification is achieved by PCR amplification using species-specific primers (33, 60) or by

sequence analysis of PCR-amplified rrs (33). The exact length of incubation before cultures

are considered negative is not determined, but they should be kept for at least 14 days,

maintaining the cell density adjusted to about 2 × 105/ml.

Serologic Tests

Anaplasma phagocytophilum IFA

Although testing can be performed using A. phagocytophilum antigens prepared from

infected circulating leukocytes of horses, the preferred method for testing human sera is the

use of a human isolate propagated in the HL-60 promyelocyte cell line

(4, 38, 60, 70, 112, 137). It is now well demonstrated that antigenic diversity exists among

isolates of A. phagocytophilum, but such diversity has not been shown to affect detection in

clinical specimens (10, 137). Interpretation of immunofluorescent patterns is similar to that

for E. chaffeensis and requires an experienced microscopist. Commercial sources of IFA

serodiagnostic kits include Focus Technologies (Cypress, CA), Scimedx Corp. (Denville, NJ),

and PanBio Diagnostics (no longer available in the United States).

Sera should be screened at a single dilution (1:64 or 1:80), and the presence of antibodies is

determined after incubation with fluorescein isothiocyanate-conjugated anti-human IgG. If

specimens test reactive, they are serially diluted to determine the end point titer. A serologic

confirmation diagnosis is achieved when a fourfold rise in titer is demonstrated in

convalescence with a minimum IgG titer of 80 or when a single antibody titer of ≥80 is

demonstrated in a patient with typical clinical features of HGA (3, 15, 16). Approximately 25

to 45% of infected patients have antibodies at the time of presentation (3, 4, 15, 16);

however, up to 11 to 14% of the population possess antibodies in some regions of high

endemicity, rendering a single serologic test less useful (2, 14). The typical response during

acute infection is a rapid rise (within 2 weeks of onset) in antibody levels reaching high titers

(≥640) within the first month (4, 15). In treated patients whose diagnosis was confirmed by

culture, antibody titers declined gradually over the several months, and about one-half of

these patients had antibodies detectable by IFA 1 year after infection. However, many

patients have antibodies detectable for months to years after the initial infection (15).

The sensitivity and specificity of the HGA serologic tests are both believed to be high because

of good correlation between typical clinical cases and serologic reactions to A.

phagocytophilum group antigens (3, 4,15). Seroconversion was documented in 21 of 23

patients (91.3%) with culture-confirmed HGA from whom a convalescent-phase sample was

available (4). In an inter- and intralaboratory evaluation, paired serology had a median

sensitivity of 95% for the detection of acute HGA in a group of 28 patients diagnosed by

culture, PCR, or the presence of morulae in blood smears (137). IgM testing appears to be a

useful tool for identification of recent infection, but neither the sensitivity nor the specificity

is as high as testing paired sera for IgG (137). Although ELISA and immunoblots have been

described (45, 70, 112), they are not routinely used methods for the serodiagnosis of HGA.

Patients with HGA have serologic reactions to E. chaffeensis in up to 15% of cases but often

show higher titers with antigens of the homologous infecting agent (3, 137). Thus, when

ehrlichiosis is clinically suspected, screening for antibodies against E. chaffeensis and A.

phagocytophilumis recommended (31). Immunoblots may be used to differentiate among A.

phagocytophilum and E. chaffeensisinfections by demonstration of a major A.

phagocytophilum antigen of approximately 44 kDa in sera of HGA patients (45, 70, 141).

False-positive reactions can be observed in patients infected with other rickettsiae, Q fever,

and Epstein-Barr virus. Many patients with HGA develop antibodies that react with Borrelia

burgdorferi by ELISA and demonstrate diagnostic IgG or IgM immunoblots (144). Most of

these likely represent false positives for B. burgdorferi, although some patients have been

confirmed by culture to have concurrent infection with A. phagocytophilum and B.

burgdorferi (97, 141). Another explanation is previous exposure to another tick-borne agent

(sequential tick bites). Indeed, antibodies to multiple agents are common in individuals living

in areas of high endemicity (14, 87). Autoantibodies to platelets and other leukocyte

components also can cause false-positive IFA tests (142).

Technologies that use commercially prepared recombinant A. phagocytophilum msp2

proteins or peptides in devices that enable rapid detection have been employed for serologic

diagnosis in veterinary laboratories, but these antigens have not been evaluated for

diagnosis of human infections (17).

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

Routine antimicrobial susceptibility testing of Ehrlichia or Anaplasma species isolates is

unnecessary. These bacteria are maintained enzootically by transmission among ticks and

feral mammalian reservoir hosts (5, 82,105, 132, 138). The level of exposure of such

vertebrate and invertebrate hosts to antimicrobial selection factors is very low, and thus,

antimicrobial pressure that results in resistance is very unlikely. Most patients with either

HME or HGA defervesce within 48 hours of therapy with doxycycline, the drug of choice

(13, 57). Tetracyclines are uniformly bactericidal for Ehrlichia and Anaplasma species,

whereas the MICs of chloramphenicol cannot be safely achieved in humans with HME or HGA

(20, 25, 66, 77, 93). In contrast, many antibiotics prescribed for undifferentiated fever, such

as penicillins, cephalosporins, aminoglycosides, and macrolides, do not inhibit the growth of

ehrlichiae in vitro. The rifamycins (rifampin and rifabutin) can achieve effective inhibition or

killing of Ehrlichia and Anaplasma species in vitro, and the fluoroquinolones (ofloxacin and

levofloxacin) have very low MICs for human isolates of A. phagocytophilum (66, 93).

However, at least one report documents recrudescence of infection with A.

phagocytophilum after levofloxacin was discontinued; the patient ultimately responded

appropriately to doxycycline treatment (143). Rifampin has been successfully used to treat

HGA during pregnancy and could be a useful alternative for patients who cannot receive

tetracyclines (27, 79). In vitro susceptibility testing by real-time PCR found that E.

chaffeensis was susceptible to doxycycline and rifampin and was partially susceptible to the

fluoroquinolones. Resistance to macrolides, co-trimoxazole, and beta-lactam compounds was

confirmed (20).

Whereas persistent infections with Ehrlichia and Anaplasma species may occur in naturally

and experimentally infected animals even after treatment with tetracycline, persistence of

ehrlichiae in humans is rarely documented and is not believed to have any clinical

importance (46, 52, 118). Therapy is usually highly effective at eliminating ehrlichiae from

the blood of infected humans.

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

Identification of infections with Ehrlichia and Anaplasma species requires clinical suspicion

followed by laboratory confirmation. Since rapid specific diagnosis is infrequently possible,

empiric therapy should be initiated when the diagnosis is suspected, since delays may lead to

increased morbidity and perhaps mortality. Collection of diagnostic samples should ideally

occur before therapy is initiated, and patients should be encouraged to return for clinical and

serologic follow-up 2 to 4 weeks later.

The presence of intracytoplasmic inclusions within a leukocyte in peripheral blood is helpful,

but they can be difficult to distinguish from overlying platelets, Dohle bodies, toxic

granulation, nuclear fragments, Auer rods, other bacteria, yeast, inorganic materials, or

normal granules. If the typical morphology of an Ehrlichia orAnaplasma spp. morula is

observed, an assessment as to the hematopoietic lineage and the percentage of cells that

contain morulae should be made and reported.

A positive PCR result should be reported as such, indicating the presence of E.

chaffeensis or A. phagocytophilum DNA, and it should be made clear that a positive PCR is

not equivalent to the culture of ehrlichiae from blood. Laboratories that use a broad-range

PCR to identify Ehrlichia or Anaplasma spp. DNA in blood may also detect E. ewingii infection

that may mimic either HME or HGA (28).

IFA serologic results should be reported as the titer of antibodies determined to be reactive

with E. chaffeensis or A. phagocytophilum, including the positive cutoff values determined in

the laboratory. The gold standard for the diagnosis of E. chaffeensis or A.

phagocytophilum infection by serology is a fourfold increase in IgG titer or documentation of

seroconversion. An interpretation should indicate whether the titers are considered

“significant” or “positive” based on a fourfold increase or fourfold decrease or only as a single

high serum IgG titer. The use of IgM titers is not advocated to establish a diagnosis, since it

is not as sensitive as IgG alone. It should be remembered that infections with E. ewingii yield

serologic patterns considered diagnostic for E. chaffeensis. Although not routinely available,

immunoblot analyses could provide information about antibodies that react with specific

antigens considered unique or diagnostic of infection with a single species

of Ehrlichia or Anaplasma.

Human ehrlichioses became nationally notifiable in 1999, but not all states report these

diseases. For the purpose of surveillance, the Council of State and Territorial Epidemiologists

and the CDC developed a case definition that was amended in 2008 to include HME, HGA, E.

ewingii infection, and ehrlichiosis/anaplasmosis —undetermined

(http://www.cdc.gov/ncphi/disss/nndss/casedef/ehrlichiosis_2008.htm). According to this

definition, a case presenting with fever, headache, myalgia, anemia, leukopenia,

thrombocytopenia, or any hepatic transaminase elevation as described above could be

classified as confirmed based on specific laboratory findings. A confirmed HME or HGA case is

supported by (i) a fourfold change in IgG antibody titer toE. chaffeensis or A.

phagocytophilum antigen, respectively, by IFA in paired serum samples; or (ii) positive PCR

and confirmation of E. chaffeensis or A. phagocytophilum DNA, respectively; or (iii)

immunostaining of E. chaffeensis or A. phagocytophilum antigen, respectively, in a biopsy or

autopsy sample; or (iv) culture of E. chaffeensis or A. phagocytophilum, respectively, from a

clinical sample. Supportive laboratory evidence is provided by an IFA IgG titer of ≥64 to E.

chaffeensis or A. phagocytophilum antigen or by identification of morulae in monocytes

(HME) or neutrophils (HGA). The diagnosis of E. ewingii infection can only be established by

nucleic acid amplification methods since specific antigens and serologic methods are not

available. The statement “ehrlichiosis or anaplasmosis ‘undetermined’ ” is used when

serological tests cannot distinguish between E. chaffeensis and A. phagocytophilum as

agents of infection.