<<

GRAND ROUNDS CLINICIAN’S CORNER AT THE JOHNS HOPKINS BAYVIEW MEDICAL CENTER

Human Monocytic

John H. Stone, MD, MPH A 56-year-old man with a history of Wegener granulomatosis presented with Kerry Dierberg 6 days of sinus congestion, , , , , and a mor- Ghazaleh Aram, MD billiform . An exacerbation of Wegener granulomatosis was the principal J. Stephen Dumler, MD concern because of the frequency of flares in that . The patient devel- oped acute renal failure, , transaminitis, and, finally, se- CASE PRESENTATION vere that led to congestive heart failure. Additional history- A 56-year-old white man with a history taking and the evolution of his clinical features led to empirical treatment with of Wegener granulomatosis (WG) pre- for monocytic ehrlichiosis (HME). The diagnosis of HME sented in May 2003 with 6 days of si- nus congestion, fever, malaise, myal- was confirmed by both a polymerase chain reaction assay for chaf- gias, and red eyes. The appearance of feensis and by the demonstration of morulae within peripheral blood mono- these symptoms in the context of his pre- nuclear cells. The patient improved promptly following institution of doxy- vious WG immediately led to hospital- cycline, and his cardiac function returned to normal over a period of 4 months. ization. On admission, his creati- JAMA. 2004;292:2263-2270 www.jama.com nine level was discovered to be 3.8 mg/dL (336 µmol/L) (patient’s baseline, 1.9 mg/dL [168 µmol/L]; normal range, Ͻ1.3 Despite daily treatment with cyclophos- mocystis prophylaxis; and lisinopril, 10 mg/dL [115 µmol/L]), heightening con- phamide and high-dose glucocorti- mg/d. He had discontinued glucocorti- cern about a potential WG flare. coids, the patient’s renal function had coids for 9 months and his ANCA se- The patient’s presentation with WG 16 continued to decline for 4 weeks after the rum samples had been negative for 8 months earlier had been characterized by start of therapy, requiring months. He was married and worked for 2 months of nasal and sinus conges- when his serum reached 8.3 a communications company. He re- tion, followed by episcleritis, tongue ul- mg/dL (734 µmol/L). The patient had ported a remote history of smoking and cers, and a nondestructive polyarthritis discontinued dialysis after 6 weeks and occasional alcohol consumption. He had involving small and large joints. He had achieved a new baseline serum creati- traveled to Maryland’s eastern shore in also developed digital ischemia with nine of 1.9 mg/dL. search of retirement properties 2 weeks splinter hemorrhages and elevation of he- Remission of his WG was induced by before admission. patic transaminase levels (aspartate and treatment with cyclophosphamide and The patient denied shortness of alanine aminotransferase levels [AST and prednisone for 5 months. He then breath, cough, , stiff- ALT] Ͼ1000 U/L; normal range, Ͻ7-40 switched to azathioprine, 100 mg/d, for ness, , or vom- U/L). of a solitary left upper lobe remission maintenance. In addition to iting, , and dysuria. On physi- lung mass had demonstrated necrotiz- conventional WG treatment, the pa- Author Affiliations: Johns Hopkins Center, Di- ing, granulomatous . A se- tient also participated in a randomized, vision of Rheumatology (Dr Stone), Johns Hopkins Uni- rum sample had tested positive for an- double-blind trial of etanercept (a versity School of Medicine (Ms Dierberg), Department of Medicine, Johns Hopkins Bayview Medical Center (Drs tineutrophil cytoplasmic soluble inhibitor of tumor fac- Stone and Aram), and Division of Medical Microbiol- (ANCA) by , with tor [TNF]) vs placebo for maintenance ogy, Department of Pathology, Johns Hopkins Univer- a cytoplasmic ANCA pattern and of WG remission.1,2 At the time of his sity School of Medicine (Dr Dumler), , Md. Corresponding Author: John H. Stone, MD, MPH, an antigen specificity for proteinase 3. current admission, the patient was re- Johns Hopkins Vasculitis Center, Johns Hopkins Bay- ceiving azathioprine, 100 mg/d; his ex- view Medical Center, 5501 Hopkins Bayview Cir, perimental (either etaner- 1B.23, Baltimore, MD 21224 ([email protected]). See also Patient Page. Grand Rounds at The Johns Hopkins Bayview Medi- cept, 25 mg twice per week, or placebo); cal Center Section Editors: David B. Hellmann, MD, CME available online at trimethoprim-sulfamethoxazole (1 Charles Weiner, MD, Stephen D. Sisson, MD, The Johns www.jama.com Hopkins Hospital, Baltimore, Md; David S. Cooper, MD, single-strength tablet per day) for Pneu- Contributing Editor, JAMA.

©2004 American Medical Association. All rights reserved. (Reprinted) JAMA, November 10, 2004—Vol 292, No. 18 2263 HUMAN MONOCYTIC EHRLICHIOSIS cal examination, he had a maximal in TABLE 1, along with results obtained day 3, the ANCA assay by immuno- temperature of 40°C, a pulse of 104/ during subsequent hospitalization. fluorescence was reported as negative. min, a blood pressure of 91/54 mm Hg, On hospital day 4, the patient be- and an oxygen saturation of 96% on Early Hospital Course came progressively dyspneic. A chest ra- room air. He had ocular con- The patient’s inactive sediment diograph showed mild pulmonary con- sistent with episcleritis. The patient had partly assuaged concerns about a WG gestion but no focal infiltrates. His no oral thrush, tongue ulceration, or na- flare. The patient’s azathioprine and ex- intravenous fluids were stopped, but the sal . There was no evidence of perimental medication were sus- dyspnea continued to worsen. His oxy- . His cardiac exami- pended because of concern about a gen saturation was 93% on room air, and nation revealed a normal S1 and S2 with possible opportunistic . Tri- an arterial blood gas measurement dem- no murmurs, rubs, or gallops, and his methoprim-sulfamethoxazole and lisin- onstrated a pH of 7.4, a PO2 of 66.0 lungs were clear to auscultation. The ab- opril were also discontinued because of mm Hg, and a PCO2 of 29.0 mm Hg. An domen was nontender and showed no his declining renal function. He was electrocardiogram indicated normal volt- signs of . The musculo- treated with intravenous fluids for acute- age and no signs of ischemia or pericar- skeletal examination was remarkable for on-chronic renal failure, presumed sec- dial inflammation. The patient’s car- absence of . The skin had a faint ondary to prerenal azotemia. Despite on- diac enzymes, however, were strikingly but diffuse, confluent macular rash over going fever, empirical were elevated: serum troponin I level was the arms, trunk, and face that spared the not begun because there was no clear in- Ͼ200 ng/mL (normal range, 0.0-0.5 ng/ palms and soles and blanched on pres- fectious source. Blood and urine cul- mL) and creatine kinase was 1842 U/L sure. There were no splinter hemor- tures remained negative. (normal range, 24-195 U/L), with a cre- rhages. The neurological examination On hospital day 2, the patient devel- atine kinase–MB isoenzyme of 99 ng/mL was nonfocal. The patient’s laboratory oped voluminous, watery that (normal range, 0-5 ng/mL). An echocar- results on admission were remarkable was nonbloody. difficile diogram showed left ventricular sys- for a sodium level of 129 mEq/L (nor- assays of stool were negative. His tolic dysfunction and an ejection frac- mal range, 135-148 mEq/L), creatinine persisted, his serum creatinine tion of 10%. The patient was transferred level of 3.8 mg/dL, AST level of 36 U/L, level rose to 4.5 mg/dL (398 µmol/L), to the cardiac care unit. ALT level of 22 U/L, albumin level of 3.8 and he developed brownish nasal crusts g/dL (normal range, 3.5-5.3 g/dL), and without epistaxis. By hospital day 3 the Diagnostic Considerations an erythrocyte sedimentation rate of 12 patient’s serum creatinine level had sta- and the Case Breaker mm/h (normal range, Ͻ20 mm/h). A bilized but he had developed signifi- Because of the frequency with which urinalysis with microscopy showed 2+ cant elevations of his hepatic transami- WG recurs, the principal concern at the protein, no red blood cells, 5 to 10 white nases and a decreased count time of the patient’s hospital admis- blood cells per high-power field, and no (Table 1). The transaminitis height- sion was a flare of that disease. Follow- casts. A was normal. ened concern about a potential WG ing the initial control of WG with con- For comparison, the results of blood flare because the patient’s AST and ALT ventional therapies in approximately work obtained 2 weeks prior to admis- concentrations had been elevated at the 90% of patients, between 60% and 80% sion (for monitoring of WG) are shown time of his WG diagnosis. On hospital eventually experience disease flares.2-4

Table 1. Pertinent Laboratory Test Results 2 Weeks Before Hospitalization and After Admission Hospital Day*

Laboratory Test −14 1 4 5 6 8 9 16 Body temperature, °C 36.2 39.9 37.0 37.1 37.0 36.0 36.0 36.4 Sodium, mEq/L 141 129 135 129 131 137 135 138 Creatinine, mg/dL 2.0 3.8 3.3 3.4 3.6 2.4 2.0 2.1 AST, U/L 26 36 286 536 NA 934 143 37 ALT, U/L 19 22 74 114 NA 948 525 106 count, ϫ103/µL 7.7 3.5 2.9 3.0 3.1 4.4 5.6 7.8 Platelet count, ϫ103/µL 410 169 77 81 85 99 222 401 Total CK, U/L† NA NA 1842 2062 1318 NA NA NA CK-MB, ng/mL† NA NA 99 77 30 8 NA NA Troponin I, ng/mL† NA NA Ͼ200 173.69 58.4 NA NA NA Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CK, creatine kinase; NA, not analyzed because laboratory test not performed. SI conversion: To convert creatinine to µmol/L, multiply values by 88.4. *Day 1 was the day of hospital admission. Patient received doxycycline on hospital days 5 through 16. †Normal ranges for total CK, CK-MB, and troponin I are 24 to 195 U/L, 0 to 5 ng/mL, and 0 to 0.5 ng/mL, respectively.

2264 JAMA, November 10, 2004—Vol 292, No. 18 (Reprinted) ©2004 American Medical Association. All rights reserved. HUMAN MONOCYTIC EHRLICHIOSIS

Renal disease, the most feared compli- ever, the patient’s wife recalled that dur- DISCUSSION cation of WG, may progress swiftly; the ing the couple’s trip to the eastern shore rise in the patient’s creatinine level from of Maryland, she had found and re- Human Monocytic Ehrlichiosis baseline prompted appropriate alarm. moved a from her skin. This sug- Bacteriology and Pathophysiology Many features of his acute presenta- gested a tick-borne disease as a possible A group of human previ- tion, particularly renal dysfunction, epi- etiology for the patient’s illness. Sero- ously known merely as “ehrlichiosis” re- scleritis, brown nasal crusts, and trans- logical assays for burgdorferi sults from the bite of infected . The aminitis, were consistent with a WG () and rickettsii name of this disorder derives from that flare. (Significant hepatic inflamma- (Rocky Mountain ) and of the causative agent, E chaffeensis. The tion is not typical of WG,5 but the pa- polymerase chain reaction (PCR) as- Ehrlichia was named to honor the tient had manifested this complica- says for human monocytic ehrlichiosis famous German bacteriologist Paul tion at the time of his WG diagnosis.) (HME; caused by ) Ehrlich, who did not actually contrib- Urine sediment without or and human granulocytic ute to its discovery or characterization. casts and the presence of (HGA; phagocytophilum) were The infectious , members of only baseline both argued obtained. When the patient began to de- the recently reorganized family Anaplas- against recurrent WG. velop increasing respiratory distress on mataceae,7 are obligate intracellular bac- The absence of many features of his hospital day 4, these results were still teria. Human in this family in- previously active WG supported a pending. Doxycycline, 100 mg/d, was in- clude organisms in the genera Ehrlichia non-WG cause of his current clinical pre- stituted empirically because of the strong and Anaplasma: E chaffeensis, Ehrlichia sentation. The patient had neither epi- suspicion that ehrlichiosis was the cause ewingii, and A phagocytophilum. These staxis nor bloody nasal crusts, both of of the patient’s illness. The assay result Rickettsia-like infect either the which suggest active nasal inflamma- for HME was reported later as positive / line (E chaffeen- tion, a hallmark of WG. The nasal crusts (and the HGA PCR result as negative). sis) or (E ewingii and A observed in the hospital were likely re- Review of the patient’s peripheral blood phagocytophilum).8 Comparisons of the lated to the effects of dry hospital air su- smear from the day of hospital admis- major features of the organisms that perimposed on damaged nasal mucosa sion showed several contain- cause ehrlichioses (and anaplasmosis) are from his previously active disease. In ad- ing characteristic cytoplasmic bacterial shown in TABLE 2. The life cycle of E dition, the patient had no , digi- inclusions (morulae) (FIGURE 1A). chaffeensis involves both mammalian and tal ischemia, or pulmonary lesions, and tick reservoirs (FIGURE 2). his erythrocyte sedimentation rate was Response and Resolution normal. Finally, several features of his The patient’s fever abated immediately Epidemiology and Ecology current illness, particularly the fever, with the institution of doxycycline; he The CDC recorded at least 1508 cases , rash, and diarrhea—which did not had no further fever after receiving his of HME in the through develop until hospital day 2—marked a first dose. His abnormal laboratory mea- 2003. That the disease is not report- significant departure from his original surements, including cardiac enzymes, able in all states probably leads to sub- WG presentation. Wegener - serum creatinine, and platelet count, also stantial underreporting.11,12 In highly tosis may be associated with improved promptly. The AST and ALT regions such as southeastern through several mechanisms (eg, alveo- continued to rise after initiation of ,13 recent studies have esti- lar hemorrhage, anemia of end-stage re- therapy but resolved prior to discharge. mated the annual incidence rates of nal disease, or the effects of medica- Delayed improvement in liver function HME to be as high as 158 cases per tions), but thrombocytopenia is distinctly tests is known to occur after the start of 100000 population. Cross-sectional se- uncharacteristic of WG. On the con- effective antiehrlichial therapy in HME. roprevalence studies have detected an- trary, because are acute-phase This is believed to be secondary to lin- tibody titers to E chaffeensis of at least reactants, thrombocytosis is the rule for gering inflammation within the hepatic 1:80 in 13% of children and at least patients presenting with generalized WG. lobules and subsequent death. 1:160 in 3% of children living in the In addition to a WG flare, the remain- The patient started a medical regimen for midwestern “tick belt.” This seroprev- der of the cen- heart failure (lisinopril and atenolol) and alence rate, which remains high even tered around possible infections. The ma- was discharged from the hospital 11 days with more stringent cutoffs for posi- jor considerations included a variety of after admission. Blood tests performed tive results, suggests that subclinical in- viral illnesses that might have led to de- 1 week after discharge (day 16 in Table 1) fections are common.14 hydration and prerenal azotemia, and a showed rapid resolution of laboratory The distribution throughout areas of group A streptococcal infection, which abnormalities. Serial echocardiograms at the southeast, south-central, and mid- might have caused the decline in renal 1 week, 6 weeks, and 4 months after dis- Atlantic regions corresponds to areas function via poststreptococcal glomeru- charge showed ejection fractions of 35%, where americanum (Lone lonephritis.6 On further inquiry, how- 45%, and 55%, respectively. Star) ticks, the vectors of E chaffeensis,

©2004 American Medical Association. All rights reserved. (Reprinted) JAMA, November 10, 2004—Vol 292, No. 18 2265 HUMAN MONOCYTIC EHRLICHIOSIS are abundant15,16 (FIGURE 3). White- known to bite between April the scapularis tick, as opposed tailed deer (Odocoileus virginianus), the and August, is thought to to A americanum) is caused by inflam- major animal reservoir of E chaffeen- occur most often after bites of adult mation at the site of B burgdorferi in- sis, are persistently bacteremic with the ticks.15 The tick bites themselves are oculation, followed by radial exten- but do not develop clinical generally and may go un- sion as the spirochete spreads through signs of infection.15,17 As with our pa- noticed. The skin lesions that result the . Human monocytic ehrli- tient, who became ill in May, most in- from differing of ticks, how- chiosis is not associated with a similar fections occur between May and July, ever, may differ based on the patho- skin lesion. corresponding to the peak feeding times gen inoculated. For example, the clas- Men are more frequently diagnosed of Lone Star ticks. Although ticks in sic lesion associated with HME than women by a ratio of both the adult and nymphal stages are with Lyme disease (transmitted by more than 2 to 1, presumably because

Figure 1. Peripheral Blood Smear and Replication of Ehrlichia chaffeensis and Cellular Response

A Peripheral Blood Smear BReplication of E chaffeensis and Inhibition of Immune C Activation of Immune Response in Uninfected Response in Infected Mononuclear Mononuclear

E chaffeensis Dense Core Cell Receptor-Based Recognition Antibodies TLR2, Fc Receptors Endocytosis CD14 TLR4

Infected E- or L-Selectin Monocyte INFECTED UNINFECTED Endosome MONONUCLEAR MONONUCLEAR PHAGOCYTE PHAGOCYTE Phagolysosome Reticulate Cell Inhibition of Lysosomal Fusion Lysosomes Unfused Activation of Replication Lysosomes Intracellular Signaling Morula Degradation

Direct Alteration Inhibition of of Gene Intracellular Transcription Signaling ? Alteration of Alteration of Gene Expression Release and Alteration of Gene Expression Dissemination Membrane Trafficking Antigen Presentation Induction of Cell Survival Pathways Release of Proinflammatory Reduced Expression of and Chemokines Recognition Receptors CD14 TLR2, TLR4 Activation of Innate and Adaptive Immune Responses Attenuation of Inflammation Innate and Adaptive Immune Responses

A, Peripheral blood smear obtained on hospital day 1 showing a monocyte with 2 vacuoles (morulae) that contain E chaffeensis clusters (arrowheads) (Wright stain; original magnification ϫ240). B, E chaffeensis infects mononuclear phagocytes in blood and in tissues. In the absence of an immune response, the bacterium is inter- nalized via endocytosis. After entry, dense core cells differentiate into reticulate cells that are suspected to be the metabolically active and replicative stage. By poorly understood mechanisms, E chaffeensis directs early endosomes containing the bacteria into a receptor-salvage pathway that precludes lysosomal fusion. Likewise, infection alters signal transduction pathways and is associated with significant changes in cell gene transcription. These events lead to global down-regulation of innate immune response pathways and receptor expression, intracellular vesicular trafficking, and activation of antiapoptotic survival mechanisms. The net effect is a more hospitable environment for intracellular replication. Replicated bacteria are released by cell lysis or exocytosis to infect other mononuclear phagocytes nearby or at disseminated sites. C, With emerging immune response, the bacteria are opsonized by anti–E chaffeensis antibodies, bound to Fc receptors, and phagocytosed. Ligation of -bound E chaffeensis to Fc receptors triggers G-protein–coupled intracellular signaling that results in destruction of the bacterium, degradation and presentation of antigens for maturing adaptive immune response, and activation of innate immune pathways. This process leads to the production of proinflammatory cytokines and chemokines. Such responses are ultimately beneficial to the host by activation of crucial CD4- and CD8-mediated immunity dominated by IFN-␥. How- ever, induction of proinflammatory mediators, including TNF-␣, has the potential to cause transient local and organ injury or systemic manifestations.

2266 JAMA, November 10, 2004—Vol 292, No. 18 (Reprinted) ©2004 American Medical Association. All rights reserved. HUMAN MONOCYTIC EHRLICHIOSIS of increased recreational and occupa- also characteristic. The presence of hy- sponse to -mediated increases in tional exposure. All individuals are ponatremia, cytopenias (particularly vascular permeability. potentially susceptible to HME, and thrombocytopenia), More than 40% of recognized HME however, if they live, work in, or visit and elevated hepatic transaminase lev- infections require hospitalization.13,21 rural or suburban areas with deer. Al- els are typical of the ehrlichioses.13 Severe complications of HME include though the average age is about 48 Leukopenia, which involves both neu- , acute respira- years, severe and fatal infections have trophils and , usually re- tory distress , a septic or toxic occurred in children as well.13,15,18,19 In- solves after the first week of illness with –like syndrome, renal failure, co- fections recorded in regions without or without treatment.21 The frequency of agulopathy, and multiorgan fail- A americanum ticks, such as Califor- cytopenias in HME cannot be ex- ure.8,15 Myocardial involvement has nia, suggest alternate vectors.20 plained by direct bacterial lysis of in- been documented only rarely in fected cells because only a minority of HME.15,25,26 The case fatality rate is ap- Clinical Features leukocytes is infected and the bacteria proximately 2% to 3%.23 Fulminant The clinical manifestations of HME and infect neither platelets nor erythro- HME can occur in the setting of hu- HGA are similar, ranging from mild fe- cytes.8,15 Cytopenias are also not ex- man infec- brile illnesses to multisystem organ fail- plained by hematopoietic suppression or tions, organ transplantation, , or ure.15 The most typical presentation of hypoplasia, because most examina- immunosuppressive therapy.8,15,24,27-34 either infection is fever, often accom- tions reveal hypercellular or normocel- Severe infections and death, however, panied by , rigors, myalgias, lular bone marrow.22 The frequent de- may also occur in nonimmunocompro- and malaise. None of these features dis- tection of macrophage-rich inflammatory mised individuals.15,19,35-37 tinguishes HME or HGA from other infiltrates within the liver, often accom- An association between trimetho- nonspecific febrile illnesses.8,15 Evi- panied by hemophagocytic cells, impli- prim-sulfamethoxazole treatment and dence of systemic involvement is help- cates macrophage activation as the fulminant infection has been demon- ful in diagnosis: up to 50% of patients mechanism for both the cytopenias and strated with rickettsial , al- have nausea, , diarrhea, or ab- the frequency of increased serum trans- though the mechanism by which this dominal pain. Twenty-five percent of aminase levels.23 Inflammatory foci are occurs is not known. Two cases of HME HME patients have cough or other evi- found within the livers of patients with in healthy young adults taking trimeth- dence of respiratory tract involve- HME, with lobular , apoptotic oprim-sulfamethoxazole38,39 were as- ment, and 20% have central nervous , necrosis, and a paucity of sociated with fulminant infections, sug- system involvement without obvious infected cells.24 The etiology of hypona- gesting that our patient’s disease severity involvement of other organ systems.21 tremia in HME is uncertain but may be may have been heightened by his tri- Certain laboratory abnormalities are related to compensatory changes in re- methoprim-sulfamethoxazole regimen.

Table 2. Selected Epidemiological, Ecological, and Serological Features of Ehrlichioses8-15 Human Anaplasmosis Human Monocytic Ehrlichiosis (Granulocytic Ehrlichiosis) Ehrlichiosis “Ewingii” Organism Ehrlichia chaffeensis Anaplasma phagocytophilum Lone Star tick (Amblyomma Deer tick/black-legged tick (Ixodes Lone Star tick () scapularis), western americanum) black-legged tick () Animal host White-tailed deer, , foxes, White-footed mouse, other White-tailed deer, dogs, foxes, , coyotes mammals wolves, coyotes Cells infected Monocytes Neutrophils Neutrophils Time of year Spring/summer Spring/summer Spring/summer Location Southeastern, south-central, and Upper Midwest and northeast Southeastern, south-central, and mid-Atlantic United States United States, California, mid-Atlantic United States* Europe Reported cases Ͼ1500 Ͼ1700 ~8 Known tick exposure, % 80-90 45-85 90 Mortality rate, % 2.7 0.5-1.0 None reported Reported cases, male, % 67 56 100 Laboratory diagnosis Acute phase PCR, blood smear examination PCR, blood smear examination PCR, blood smear examination Convalescent phase (seroconversion) Serology (seroconversion) No specific serology Abbreviation: PCR, polymerase chain reaction. *There have been approximately 20 cases of E ewingii in humans identified in the mid-Atlantic states (J.S.D., unpublished data). The presence of E ewingii in ticks and infected dogs in the mid-Atlantic has been documented.9,10

©2004 American Medical Association. All rights reserved. (Reprinted) JAMA, November 10, 2004—Vol 292, No. 18 2267 HUMAN MONOCYTIC EHRLICHIOSIS

To our knowledge, HME has not pre- presentation herein.42 In a murine model (Figure 1). Host cells are altered by the viously been reported in any patient with of HME,43 a synergistic role for TNF-␣ intracellular infection in various ways WG. In the in which our pa- and interferon-␥ has been demon- (Figure 1B). For example, initially E chaf- tient was a participant,1,2 he had been as- strated in protection against mortality. feensis diminishes the proinflammatory signed to receive etanercept. Thus, our Use of a TNF-␣ antagonist by our pa- response of the host by altering the cell patient was receiving an immunosup- tient may have contributed to the in- surface expression of Toll-like recep- pressive regimen that included both creased severity by diminishing the in- tors (TLR2 and TLR4) and CD14.46 How- etanercept, a soluble TNF inhibitor duction of innate and adaptive immunity. ever, once opsonized by antibody, E chaf- known to be associated with an in- Loss of TNF-␣ alone does not increase feensis interacts with mononuclear creased risk of atypical infections,40 and mortality in mice, however, and a role phagocytes via Fc receptors and trig- azathioprine, an inhibitor of T and B lym- for this cytokine in human E chaffeensis gers an intense proinflammatory cyto- phocyte function. Although TNF is an infections has not yet been investigated. kine response (Figure 1C).42,47 Exces- important cytokine in the immuno- After attachment via surface recep- sive cytokine production contributes to logic response to Rickettsia species tors and internalization into monocytes/ the toxic and septic shock–like presen- and other infectious pathogens,41 its , E chaffeensis evades lyso- tations, respiratory failure, and acute res- specific role in the host response to somal fusion of the parasitophorous piratory distress syndrome observed in Ehrlichia species remains unclear. Re- vacuoles, thereby escaping destruction HME.15,35,42 cent animal models of HME suggest a and surviving within the cell.44,45 The bac- The pathophysiologic mechanisms of detrimental role for excessive TNF-␣,a teria propagate within vacuoles, form- myocarditis in HME appear to differ finding not entirely consistent with the ing inclusions known as morulae from those of Rocky Mountain spot-

Figure 2. Transmission of Ehrlichia chaffeensis by Tick Vector (Amblyomma americanum)

Uninfected Uninfected Larva Lone Star Tick White-Tailed Deer (Amblyomma Blood Meal 1 (Odocoileus americanum) virginianus)

Uninfected Infected Infected Deer or Deer 1 mm Adult

Blood Meal 3 Uninfected Larva Infected Larva

Infected Uninfected Uninfected Infected or Human Human Deer Deer Uninfected Nymph Infected Nymph

Blood Meal 2 Infected Adult

Infected Nymph Infected Uninfected Uninfected Infected Deer Deer or Human Human

Noninfected larvae obtain blood from a bacteremic vertebrate reservoir host (eg, white-tailed deer), become infected, and maintain ehrlichiae when the tick molts into the nymphal stage (trans-stadial [stage-to-stage] transmission). Infected nymphs may transmit E chaffeensis to susceptible reservoir hosts or incidentally to humans during acquisition of blood. Infected adult ticks, having acquired ehrlichiae either by trans-stadial transmission from the infected nymphal stage or during blood meal as nonin- fected nymphs on infected deer, may also pass E chaffeensis to other susceptible reservoirs or humans. Transovarial transmission (transmission to the and into larvae) has not been demonstrated, and eggs and unfed larvae are presumably not infected. White-tailed deer (Odocoileus virginianus) are the predominant reservoir for E chaf- feensis, although canids (eg, foxes, coyotes) are alternate reservoirs. White-tailed deer can maintain a subclinical E chaffeensis bacteremia for months and are hosts of all 3 stages of Lone Star ticks, its predominant vector. Photo provided by James Gathany/Centers for Disease Control and Prevention.

2268 JAMA, November 10, 2004—Vol 292, No. 18 (Reprinted) ©2004 American Medical Association. All rights reserved. HUMAN MONOCYTIC EHRLICHIOSIS ted fever (caused by R rickettsii) and Figure 3. Lone Star Tick and Its Geographic Distribution16 other rickettsioses. Ehrlichia chaffeen- sis, in contrast with R rickettsii, does not infect endothelial cells in vivo.23 The common link between cardiac involve- ment in these 2 infections is infiltra- tion of the myocardium by inflamma- tory cells. Although true vasculitis with endothelial cell loss, vascular wall ne- crosis, and thrombosis can occur in Rocky Mountain spotted fever, the more common histopathologic manifesta- Approximate Distribution of Amblyomma americanum tion is myocardial inflammatory cell in- filtrates with ,48,49 as observed with E chaffeensis infections.23 It is pos- sible that E chaffeensis–activated mac- Approximate geographic distribution of Amblyomma americanum in the United States. rophages infiltrating the myocardium produce proinflammatory cytokines or in only about 3% of patients with Treatment induce their production by other HME.8,23,53 In our patient, the identifica- No controlled prospective evaluations of cells.42,47 Activated macrophages (eg, via tion of E chaffeensis morulae within therapy for HME or HGA have been con- interferon-␥) can lead to hepatitis and peripheral blood monocytes confirmed ducted. However, empirical data, retro- hepatic necrosis.24,42,50 A similar mecha- the results of PCR and validated the spective analyses, and in vitro studies in- nism could also occur within myocar- empirical use of doxycycline. Examina- dicate that doxycycline is the drug of dial inflammatory infiltrates contain- tion of buffy-coat smears might increase choice.21,54 Rifampin, effective in vitro, ing E chaffeensis–infected monocytes. the likelihood of morulae detection, but has been associated with some success no comparative studies of buffy-coat in treatment of HGA in pregnant women. Diagnosis examinations vs routine peripheral blood Because the outcome of HME is depen- The diagnosis of HME requires a com- smears have been performed. dent on early identification and early ef- patible clinical history and laboratory fective therapy, a high degree of clinical evidence of infection. Supportive labo- Differential Diagnosis suspicion is critical to selecting the most ratory evidence includes a single IgG an- Because HME presents with nonspe- appropriate antimicrobial regimen. tibody titer of at least 256; seroconver- cific findings, the differential diagnosis Therapy is usually advocated for at least sion from negative to positive antibody is broad. Presentations with multiorgan 3 to 5 days after defervescence or ap- status (with a minimum titer of 64), a disease invoke , toxic shock syn- proximately 10 days total. However, in 4-fold rise in titer during convales- drome, meningococcemia, typhoid fe- states such as Maryland, where the preva- cence, recovery of E chaffeensis in cul- ver, and as potential diag- lence of Lyme disease is high, clinicians ture, or detection of E chaffeensis nucleic noses. Noninfectious mimics of HME often use a 2-week course of doxycy- acids in blood or by include , thrombotic cline because of the possibility of simul- PCR.15,51 Although our patient had a late thrombocytopenic , hemato- taneous infection with B burgdorferi, acute-phase E chaffeensis antibody titer logic malignancies, collagen vascular dis- which requires a second tick bite be- of 1024, the majority of patients (60%- eases, and, as demonstrated herein, sys- cause this organism infects the I scapu- 97%) are seronegative during the acute temic vasculitis. Given a history of laris tick (Figure 3), not A americanum. phase.8,13,21 Thus, therapeutic deci- animal, tick, or vector exposure, the dif- sions must often be based on a high in- ferential diagnosis must also include Prognosis dex of clinical suspicion and labora- Rocky Mountain spotted fever, murine Most patients with HME respond favor- tory evidence of the infection, such as , , Lyme disease, tu- ably to treatment, especially if therapy is PCR assays and peripheral blood smears. laremia, , Colorado tick administered early in the course of the Diagnosis by PCR is faster than enzyme fever, , , , infection.13,21 The typical response to immunoassay and has a sensitivity of 60% and, in Europe, tick-borne encephali- doxycycline is rapid defervescence within to 85%,13,52 but the potential for false- tis. Our patient illustrates the impor- 24 to 48 hours. After 1 dose of doxycy- positive results is high.15 Examination of tance of considering ehrlichiosis in pa- cline, our patient had no more fever. the peripheral blood smear for the typi- tients who present with fever, generalized Most laboratory changes also resolve rap- cal intracytoplasmic morulae in mono- aches, cytopenias (particularly leukope- idly, with return of the white blood cell cytes is not informative in most cases. nia and thrombocytopenia), and abnor- and platelet counts to normal within sev- Morulae are detected within monocytes mal liver function. eral days.21 Although the objective clini-

©2004 American Medical Association. All rights reserved. (Reprinted) JAMA, November 10, 2004—Vol 292, No. 18 2269 HUMAN MONOCYTIC EHRLICHIOSIS cal signs respond rapidly, symptoms such Granulocytic ehrlichiosis in dogs from North Carolina 33. Sadikot R, Shaver MJ, Reeves WB. Ehrlichia chaf- and Virginia. J Vet Intern Med. 1998;12:61-70. feensis in a renal transplant recipient. Am J Nephrol. as may persist for weeks or even 10. L, McPherson T, Harrison B, Engber B, Ander- 1999;19:674-676. months. The medical basis for the per- son A, Whitt P. Prevalence of Ehrlichia ewingii in Am- 34. Talbot TR, Comer JA, Block KC. Ehrlichia chaffeen- blyomma americanum in North Carolina. J Clin sis infections among HIV-infected patients in a human sistent manifestations is unclear. There Microbiol. 2000;38:2795. monocytic ehrlichiosis-endemic area. Emerg Infect Dis. is no evidence of persistent E chaffeen- 11. Gardner SL, Holman RC, Krebs JW, et al. National 2003;9:1123-1127. sis infection in humans, and treatment surveillance for the human ehrlichioses in the United States, 35. Patel RG, Byrd MA. Near fatal acute respiratory dis- 1997-2001, and proposed methods for evaluation of data tress syndrome in a patient with human ehrlichiosis. South for disease beyond the acute infection is quality. AnnNYAcadSci. 2003;990:80-89. Med J. 1999;92:333-335. not indicated.8,13,15 12. McQuiston JH, Paddock CD, Holman RC, et al. 36. Smith Sehdev AE, Sehdev PS, et al. Human mono- The human ehrlichioses in the United States. Emerg cytic ehrlichiosis presenting as acute dur- Infect Dis. 1999;5:635-642. ing pregnancy. Clin Infect Dis. 2002;35:e99-e102. CONCLUSION 13. Olano JP, Masters E, Hogrefe W, et al. Human 37. Dunn BE, Monson TP, Dumler JS, et al. Identifica- monocytotropic ehrlichiosis, Missouri. Emerg Infect Dis. tion of Ehrlichia chaffeensis morulae in cerebrospinal fluid Human monocytic ehrlichiosis is a dis- 2003;9:1579-1586. mononuclear cells. J Clin Microbiol. 1992;30:2207-2210. ease that can cause severe infections or 14. Marshall GS, Jacobs RF, Schutze GE, et al; Tick- 38. Peters TR. Severe ehrlichiosis in an adolescent tak- Borne Infections in Children Study Group. Ehrlichia ing trimethoprim-sulfamethoxazole. Pediatr Infect Dis death yet provides few specific clinical chaffeensis seroprevalence among children in the J. 2000;19:170-172. clues to diagnosis. Awareness of this dis- southeast and south-central regions of the United 39. Brantley RK. Trimethoprim-sulfamethoxazole and States. Arch Pediatr Adolesc Med. 2002;156:166-170. fulminant ehrlichiosis. Pediatr Infect Dis J. 2001;20:231. order, recognition of its particular risk 15. Paddock CD, Childs JE. Ehrlichia chaffeensis. Clin 40. Ellerin T, Rubin RH, Weinblatt ME. Infections and factors (eg, travel to endemic regions, Microbiol Rev. 2003;16:37-64. anti--␣ therapy. Arthritis Rheum. tick bites), and understanding the spec- 16. Childs JE, Paddock CD. The ascendancy of Am- 2003;48:3013-3022. blyomma americanum as a vector of pathogens af- 41. Walker DH, Popov VL, Crocquet-Valdes PA, et al. trum of known presentations and com- fecting humans in the United States. Annu Rev Cytokine-induced, nitric oxide-dependent, intracellu- plications of HME are all critical to mak- Entomol. 2003;48:307-337. lar antirickettsial activity of mouse endothelial cells. Lab 17. Varela AS, Stallknecht DE, Yabsley MJ, et al. Ex- Invest. 1997;76:129-138. ing the diagnosis promptly and achieving perimental infection of white-tailed deer (Odocoi- 42. Ismail N, Soong L, McBride JW, et al. Overproduc- a favorable outcome. The syndrome of leus virginianus) with Ehrlichia chaffeensis by differ- tion of TNF-␣ by CD8+ type 1 cells and down- ent routes. J Wildl Dis. 2003;39:881-886. regulation of IFN-␥ production by CD4+ Th1 cells con- fever, generalized aches, cytopenias, and 18. Fichtenbaum CJ, Peterson LR, Weil GJ. Ehrlichiosis tribute to toxic shock-like syndrome in an animal model elevated liver function measurements presenting as a life-threatening illness with features of of fatal monocytotropic ehrlichiosis. J Immunol. 2004; may be particularly useful in triggering the . Am J Med. 1993;95:351-357. 172:1786-1800. 19. Dumler JS, Dey C, Meier F, et al. Human mono- 43. Feng H-M, Walker DH. Mechanisms of immu- suspicion of HME. Even in the absence cytic ehrlichiosis: a potentially severe disease in children. nity to : a model of monocytotropic of an obvious vector exposure or travel Arch Pediatr Adolesc Med. 2000;154:847-849. ehrlichiosis. Infect Immun. 2004;72:966-971. 20. Fritz CL, Kjemtrup AM, Conrad PA, et al. Seroepi- 44. Zhang J, McBride J, Yu X. L-selectin and E- history, empirical doxycycline therapy demiology of emerging tickborne infectious diseases in selectin expressed on monocytes mediating Ehrlichia may be lifesaving for patients in whom a northern California community. J Infect Dis. 1997;175: chaffeensis attachment onto host cells. FEMS Micro- 1432-1439. biol Lett. 2003;227:303-309. there is a reasonable clinical suspicion 21. Fishbein DB, Dawson JE, Robinson LE. Human ehrli- 45. Mott J, Barnewall RE, Rikihisa Y. Human granu- of HME. chiosis in the United States, 1985 to 1990. Ann Intern locytic ehrlichiosis agent and Ehrlichia chaffeensis re- Med. 1994;120:736-743. side in different cytoplasmic compartments in HL-60 22. Dumler JS, Dawson JE, Walker DH. Human ehrli- cells. Infect Immun. 1999;67:1368-1378. REFERENCES chiosis. Hum Pathol. 1993;24:391-396. 46. Lin M, Rikihisa Y. Ehrlichia chaffeensis down- 23. Walker DH, Dumler JS. Human monocytic and regulates surface Toll-like receptors 2/4, CD14 and 1. The WGET Research Group. Design of the Wege- granulocytic ehrlichioses. Arch Pathol Lab Med. 1997; transcription factors PU.1 and inhibits lipopolysaccha- ner’s Granulomatosis Etanercept Trial (WGET). Con- 121:785-791. ride activation of NF-␬B, ERK 1/2 and p38 MAPK in trol Clin Trials. 2002;23:450-468. 24. Sehdev AE, Dumler JS. Hepatic pathology in hu- host monocytes. Cell Microbiol. 2004;6:175-186. 2. The WGET Research Group. Etanercept in addi- man monocytic ehrlichiosis: Ehrlichia chaffeensis 47. Lee EH, Rikihisa Y. Anti-Ehrlichia chaffeensis an- tion to standard therapy in patients with Wegener infection. Am J Clin Pathol. 2003;119:859-865. tibody complexed with E. chaffeensis induces potent pro- granulomatosis. N Engl J Med. In press. 25. Williams J, Snow R, Arciniegas J. Myocardial in- inflammatory cytokine mRNA expression in human 3. Hoffman GS, Kerr GS, Leavitt RY, et al. Wegen- volvement in a patient with human ehrlichiosis. Am J monocytes through sustained reduction of I␬B-␣ and er’s granulomatosis: an analysis of 158 patients. Ann Med. 1995;98:414-415. activation of NF-␬B. Infect Immun. 1997;65:2890-2897. Intern Med. 1992;116:488-498. 26. Vanek NN, Kazi S, Cepero NM, et al. Human ehrli- 48. Walker DH, Paletta CE, Cain BG. Pathogenesis of 4. Reinhold-Keller E, Beuge N, Latza U, et al. An in- chiosis causing left ventricular dilatation and myocarditis in Rocky Mountain spotted fever. Arch terdisciplinary approach to the care of patients with dysfunction. Clin Infect Dis. 1996;22:386-387. Pathol Lab Med. 1980;104:171-174. Wegener’s granulomatosis: long term outcome in 155 27. Paddock CD, Suchard DP, Grumbach KL, et al. 49. Marin-Garcia J. Left ventricular dysfunction in patients. Arthritis Rheum. 2000;43:1021-1032. Fatal seronegative ehrlichiosis in a patient with HIV Rocky Mountain spotted fever. Clin Cardiol. 1983;6: 5. The WGET Research Group. Baseline data on pa- infection. N Engl J Med. 1993;329:1164-1167. 501-506. tients in the Wegener’s Granulomatosis Etanercept Trial 28. Paddock CD, Folk SM, Shore GM, et al. Infec- 50. Sotomayor EA, Popov VL, Feng HM, et al. Ani- (WGET). Arthritis Rheum. 2003;48:2299-2309. tions with Ehrlichia chaffeensis and Ehrlichia ewingii mal model of fatal human monocytotropic ehrlichiosis. 6. Mitchell T. The pathogenesis of streptococcal in persons coinfected with human immunodeficiency Am J Pathol. 2001;158:757-769. infections. Nat Rev Microbiol. 2003;1:219-230. virus. Clin Infect Dis. 2001;33:1586-1594. 51. Walker DH. Diagnosing human ehrlichiosis. ASM 7. Dumler JS, Barbet AF, Bekker CPJ, et al. Reorga- 29. Marty AM, Dumler JS, Imes G, et al. Ehrlichiosis mim- News. 2000;66:287-290. nization of genera in the families and icking thrombotic thrombocytopenic purpura. Hum 52. Everett ED, Evans KA, Henry RB, McDonald G. in the order ; unifica- Pathol. 1995;26:920-925. Diagnosis using polymerase chain reaction. Ann In- tion of some species of Ehrlichia with Anaplasma, Cow- 30. Tan HP, Dumler JS, Maley WR, et al. Human mono- tern Med. 1994;120:730-735. dria with Ehrlichia, and Ehrlichia with ; cytic ehrlichiosis: an emerging in 53. Childs JE, Sumner JW, Nicholson WL, et al. Out- descriptions of six new species combinations; and des- transplantation. Transplantation. 2001;71:1678-1680. come of diagnostic tests using samples from patients ignation of Ehrlichia equi and “HGE agent” as sub- 31. Arav-Boger R, Knepp JH, Walls JJ, et al. Human with culture-proven human monocytic ehrlichiosis. jective synonyms of Ehrlichia phagocytophila. Int J Syst monocytic ehrlichiosis in a child with leukemia. Pediatr J Clin Microbiol. 1999;37:2997-3000. Evol Microbiol. 2001;51:2145-2165. Infect Dis J. 2000;19:173-175. 54. Brouqui P, Raoult D. In vitro suscepti- 8. Olano JP, Walker DH. Human ehrlichioses. Med 32. Liddell AM, Sumner JW, Paddock CD, et al. Rein- bility of the newly recognized agent of ehrlichiosis in Clin North Am. 2002;86:375-392. fection with Ehrlichia chaffeensis in a liver transplant humans, Ehrlichia chaffeensis. Antimicrob Agents 9. Goldman EE, Breitschwerdt EB, Grindem CB, et al. recipient. Clin Infect Dis. 2002;34:1644-1647. Chemother. 1992;36:2799-2803.

2270 JAMA, November 10, 2004—Vol 292, No. 18 (Reprinted) ©2004 American Medical Association. All rights reserved.