Nosocomial Transmission of Human Granulocytic Anaplasmosis in China FREE

Human granulocytic anaplasmosis (HGA) is an emerging tick-borne infectious disease that was recognized in the United States in 1990¹ and in Europe in 1997.² The disease name was changed from human granulocytic ehrlichiosis to HGA in 2001 when the causative rickettsia was reclassified from the genus ehrlichia as Anaplasma phagocytophilum.³ Although the clinical presentation of HGA is variable and although it may be difficult to diagnose, the annual number of infections reported in the United States since 1990 has steadily increased.^4- 5 Seroepidemiological data suggest that infection rates in endemic areas are as high as 15% to 36%,^6- 8 implying that the diagnosis is often missed or that infection is mild or asymptomatic. Because epidemiological, clinical, and microbiological information about HGA is limited, the disease is likely underrecognized and underreported worldwide.⁷

Despite the pathogen's global distribution, only a limited number of laboratory-confirmed cases have been reported from countries in Europe, where the median seroprevalence rate is 6.2%, similar to that in North America.⁹ Serological and molecular evidence also suggests that human infection exists in Korea, Japan, and China.^10- 14 Herein, we report the first cases of HGA acquired in China, as well as the unusual finding of nosocomial human-to-human transmission.

Patients suspected of HGA exposure were tested for serum IgG to A phagocytophilum using the IgG IFA kit (Focus Diagnostics, Cypress, California), screening at a 1:64 dilution and titrating if reactive.¹⁵ Nested polymerase chain reaction (PCR) using blood DNA (QIAamp DNA Mini Kit, QIAGEN, Hilden, Germany) was used to detect A phagocytophilum DNA with Anaplasma and Ehrlichia genus-common and A phagocytophilum species-specific rrs primers (16S rRNA gene),¹⁶ and A phagocytophilum species-specific groEL primers.¹⁷ An A phagocytophilum rrs plasmid and DNA from healthy people or distilled water were used as controls. Positive reactions were confirmed by direct sequencing. Polymerase chain reaction was conducted in 2 independent laboratories, the National Institute for Communicable Disease Control and Prevention in Beijing, and at the Anhui Province Center for Disease Prevention and Control in Hefei city. Each laboratory used its own primers, reagents, and patient blood DNA. All samples were tested concurrently with negative and no template controls (water) under the same conditions. Polymerase chain reaction samples from healthy people and negative controls consistently had negative results.

To exclude other infections, serological, antigen detection, and PCR diagnostic tests were conducted. These included tests on blood from the first 3 to 5 days after onset for reverse transcription (RT)–PCR of PCR for nucleic acids of Lassa fever virus, Ebola virus, Marburg virus, Hantaan virus, Junin virus, yellow fever virus, Crimean-Congo hemorrhagic fever virus, coxsackievirus, respiratory syncytial virus, adenovirus, Mycoplasma pneumoniae, Chlamydia species, Ehrlichia species, Rickettsia species, and Orientia tsutsugamushi.

Tests were also conducted on oropharygeal swabs from the first 3 to 5 days after onset for influenza A virus antigens, and by PCR for influenza A viruses, influenza B virus, and influenza virus subtype H5 nucleic acids. Tests for acute-phase serum were conducted to detect IgM and IgG to severe acute respiratory syndrome virus, as well as to detect IgM or IgM plus IgG antibodies by capture enzyme-linked immunosorbent assay against Bunyaviridae, Filoviridae, Lassa fever virus, Ebola virus, Marburg virus, Hantaan virus, Junin virus, yellow fever virus, and Crimean-Congo hemorrhagic fever virus.

All contacts of the index patient, including patients with similar clinical presentations and healthy persons, were interviewed before laboratory diagnostic results were obtained. A possible case of HGA was defined as a patient with a clinically compatible illness (fever, headache, chills) and laboratory findings including thrombocytopenia and leukopenia but who lacked serological or molecular tests for A phagocytophilum. A confirmed case was defined as a patient with a clinically compatible illness (as above) and in keeping with the US Centers for Disease Control and Prevention (CDC) criteria (http://www.cdc.gov/ncphi/disss/nndss/casedef/ehrlichiosis_2008.htm) by either seroconversion, a 4-fold increase in A phagocytophilum IgG antibody titer in acute and convalescent sera, or a positive PCR result for both A phagocytophilum rrs and groEL confirmed by direct sequence analysis.¹⁵

All contacts of the index patient were asked to complete a questionnaire about their health status and profession; experience with tick bites; exposure to the index patient—where, when, and how they had contact; exposure to wild animals; extent of outdoor activity; exposure to the index patient's blood and respiratory secretions or to grossly bloody oropharyngeal secretions; presence of skin lesions during exposure; whether skin surfaces were washed after exposure; whether they were exposed to the patient's stool or urine; and the timing of these events. Health care workers were asked about their use of masks and gloves.

The study was approved by the ethics committee of China CDC, according to the medical research regulations of Ministry of Health, China. Oral informed consent was obtained from all study participants.

All statistical calculations were performed using Epi Info 6.04d (http://www.cdc.gov/epiinfo). To identify specific exposure risk factors, retrospective cohort comparisons were evaluated by calculating attack rates, relative risk, and 95% confidence intervals and by Fisher exact test; significance was defined as a 2-tailed P <.05.

A 50-year-old woman with a 1-day abrupt onset of sudden fever (39.2°C), headache, myalgia, arthralgia, dizziness, and malaise presented to the village clinic on October 31, 2006, and was treated with ribavirin, cephalothin, dexamethasone, and amidopyrine for 4 days. At 9 PM on November 3, she was admitted to the local hospital because of gum bleeding, facial edema, nausea, vomiting, and oliguria, a temperature of 39.7°C, blood pressure of 85/60 mm Hg, and pulse rate of 96/min; a rash was noted over her trunk. Laboratory testing showed leukopenia (white blood cell count, 3300/μL), thrombocytopenia (platelet count, 18 × 10³/μL), elevated serum aspartate aminotransferase (629 U/L) and alanine aminotransferase (69 U/L), elevated creatinine (2.6 mg/dL), and elevated blood urea nitrogen (48 mg/dL) levels. Dipstick urinalysis revealed 3⁺ hematuria and 3⁺ proteinuria (protein, 3 g/L). (To convert aspartate aminotransferase and alanine aminotransferase to microkat per liter, multiply by 0.0167; creatinine to micromole per liter, by 88.4; and urea nitrogen to millimole per liter, by 0.357.)

Her condition progressively deteriorated, so she was transferred to a regional hospital at 11 AM, November 4. By 7 PM, the patient became obtunded, cyanotic, and purpuric and was bleeding from her nose and mouth. This extensive mouth and nose bleeding required frequent aspiration and contaminated the working area surfaces, health care workers, and family members who were with her. Family members assisted with patient care by wiping blood from the patient's mouth and nose, rinsing and reusing the same towels. By 7:38 PM the patient developed rapidly progressive dyspnea and worsening oxygen saturation and required endotracheal intubation. The patient remained hypoxic and hypotensive with multiorgan failure and copious bleeding from the nose and mouth. Despite all efforts, the patient died at 6:45 AM, November 5, 2006. The final diagnosis was hemorrhagic fever with renal syndrome, even though no IgG antibodies to Hantaan virus were detected. A postmortem examination was not performed, and no blood or tissue samples remained for retrospective laboratory testing.

Retrospective questioning of the patient's family revealed that she was bitten by a tick 12 days before onset of fever: she had killed several mice in her home 9 days before onset, and her husband had hunted and brought home “wild animal carcasses” 3 days before onset of illness. A timeline of events is shown in the Figure.

Between November 9 and 17, 2006, 9 patients were identified at the regional hospital with fever higher than 38.0°C (9 of 9 patients), myalgia (5 of 9), diarrhea (7 of 9), leukopenia (white blood cell count, 1200-3700/μL in 9 of 9), thrombocytopenia (platelets, 39-115 × 10³/μL in all 9), and elevated serum aspartate aminotransferase and alanine aminotransferase (7 of 9) (Table 1). All patients had contact with the index patient, including 5 family members, 2 physicians, and 2 nurses who had accompanied or treated her between November 4 and 5 (Figure).

The initial secondary case experienced fever on November 9, 4 days after death of the index case, followed on November 11 by another patient, on November 12 by 3 patients, and on November 14 by 3 more patients. The last patient reported illness on November 17, 12 days after the death of the index patient. The patients were between 25 and 67 years (mean, 36.2 years), and 6 were men. All were previously healthy. The average incubation period was 7.8 days (range, 4-12 days). All had fever of at least 38.5°C for 1 to 6 days (mean, 4 days). Diarrhea was characterized as 1 to 3 loose stools per day persisting for 1 to 2 days. All patients had relative bradycardia. One patient developed acute respiratory distress syndrome as a complication of Aspergillus pneumonia and tuberculosis during his hospitalization. The other 8 patients were mildly affected, recovered, and were discharged in good health.

The index patient had contact with 63 persons after onset of illness: 21 family members and 42 health care workers. Of the 42 health care workers, 18 were from the local hospital, including 2 from the village clinic, and 24 were from the regional hospital. Of the 21 family members, 4 had contact with the index patient in only the local hospital, 13 only in the regional hospital, and 4 in both. The 9 secondary cases occurred among the 39 health care workers and relatives with patient exposure at the regional hospital, representing an attack rate of 23%. All 9 cared for the index case in the final 12 hours of her life while she was in the critical care unit and during the endotracheal intubation procedure. No one whose only contact with the index patient was before these 12 hours was infected.

Anaplasma phagocytophilum IgG seroconversions were detected for all 9 patients, and a 4-fold IgG titer increase was observed in 7 of 9 patients (Table 1). Nested PCR using genus-common rrs and species-specific rrs and groEL primers identified A phagocytophilum DNA in the blood samples from all 9 patients when they were in the acute phase, whereas all healthy and template controls had negative test results. The identity of amplicons from each of the 9 patients was confirmed by sequencing; all rrs sequences (206 base pairs) were identical and all groEL sequences (446 nt) were identical (GenBank accession numbers: rrsEF211110-17 and EF473210; groELEF47320108 and EF473209). Although the rrs sequences were identical to most other human-derived strains globally, sequences from groEL were identical to some US strains (Wisconsin and New York) but differed from A phagocytophilum in China (93.6%; EU008083), Germany (99.4%; AY281850), and California (99.7%; U96727). These data support the premise that a single clone was responsible for all of the 9 secondary cases. Although peripheral blood smears were examined for all 9 patients at the time of illness, no convincing evidence of A phagocytophilum morulae was observed. All RT-PCR, PCR, antigen detection, and IgM antibody detection tests for microbial and viral etiologies were negative.

The exposure data implicate transmission at the regional hospital, permitting focus on risk factors in 39 individuals, including 24 health care workers and 15 family members (Table 2). Two family members who had contact with the index case after her death were not included.

None of the 9 secondary cases reported tick bites, exposure to wild animals, or participation in hunting activity in the preceding 2 months, and only 1 reported recent outdoor activity. For all 9 secondary cases, culture serological, antigen detection, and nucleic acid detection studies for other infectious etiologies were negative.

Of 24 regional hospital health care workers who had contact with the index patient, 18 were on duty during the final 12 hours, and 4 of the 18 who were involved in the endotracheal intubation were infected. Of these 4, 3 were involved in endotracheal intubation and care during times of hemorrhage. Sixteen of 24 health care workers (67%) from the regional hospital wore masks and 9 of 24 (38%) wore gloves.

Of 17 family members who reported contact with the index patient at the regional hospital, 13 were present during endotracheal intubation, 5 of whom were infected. Of these 5 individuals, 3 reported blood contamination of skin and possible mucocutaneous exposures, suggesting direct contact with blood or respiratory secretions as the mechanism of transmission.

Among the 28 individuals who reported close contact (≤50 cm) with the index patient during the final 12 hours of her life, 9 were infected. In contrast, none of the 11 individuals who reported a physical distance of more than 50 cm from the index patient during the same time was infected. The index patient was exposed to 20 contacts for more than 2 hours, and 9 were infected, whereas none of 19 contacts exposed fewer than 2 hours was infected. All 9 infected patients reported contact with blood (P = .002) and 7 had contact with respiratory secretions (relative risk, 7.0; 95% confidence interval, 1.7-29.1; Table 2). Those persons with skin expsoure to blood (P < .001) or respiratory secretions (P = .004), or those with preexisting skin lesions or injuries followed by exposure to blood (relative risk, 3.6; 95% confidence interval, 1.1-7.6; P = .02) were significantly more likely to be infected (Table 3). Neither exposure to stool nor exposure to urine from the index case resulted in increased risk (0.6 and 1.1, respectively).

Nine cases of A phagocytophilum infection were confirmed at the regional hospital in the Anhui Province of China in a 9-day period. All presented with HGA as described in North America and Europe⁷ and fulfilled the US CDC laboratory criteria for the diagnosis of HGA.¹⁵ The most remarkable aspect of these cases was that transmission was very unlikely to be tick-borne, but was closely associated with blood or respiratory secretion exposure from an index patient who died of a fulminant febrile illness with hemorrhage. Although the index patient can only be categorized as a possible case, clinical and historical support for the diagnosis of HGA is strong. She had a tick bite within the known incubation period and had a clinical presentation compatible with severe HGA.⁴ Moreover, the epidemiological investigation of exposed individuals with HGA implicates her as the index case. Unfortunately, no tissue or serum sample is available to confirm retrospectively her diagnosis.

Human granulocytic anaplasmosis and human monocytic ehrlichiosis were initially identified with presentations now recognized as relatively uncommon for their natural histories.^18- 19 Infection can be severe, with intensive care unit admission required in 7% of patients and fatalities occurring in up to 1%, yet most infections are sporadic and probably self-limited.⁴ Based on the mild to moderate severity observed in 8 of the 9 secondarily infected patients, Chinese HGA conforms to the spectrum of clinical severity observed in North America.^4,7,15 The fatal outcome in the index case is clinically similar to that observed for other HGA fatalities, including exsanguination with sepsis syndrome possibly relating to cytokine overproduction, opportunistic infections, and increased HGA severity in the setting of preexisting medical conditions such as diabetes mellitus.^7,20

A phagocytophilum transmission in China and Asia is predicated on the presence of this zoonotic agent in vector ticks and vertebrate hosts. Although studies in Asia are limited, at least 8 have examined A phagocytophilum infection of ticks, including 2284 Ixodes persulcatus ticks, of which 4.4% carried A phagocytophilum DNA, a prevalence similar to that in European and North American Ixodes species ticks.^{12,14,21- 27} Likewise, 9% and 24% of Apodemus species field mice in northern China and Korea, respectively, and 64% of Crosidura lasiura shrews in Korea are infected.^{13,21,24,28- 29} Although no proven cases of HGA have been previously identified in China, at least 1 study describes A phagocytophilum DNA in the blood of 4 Chinese patients with tick bites,^14,30 and seroepidemiological investigations demonstrate that 2% to 9% of febrile patients in Korea,^10- 11 and between 0.5% and 6% of healthy Chinese residents have A phagocytophilum antibodies.³¹

Rare examples of nontick transmission of HGA exist in the literature and include direct exposure to deer blood,³² transfusion,³³ and transplacental transmission.³⁴ Similarly, under the proper circumstances other rickettsial infections are transmissible via aerosol, direct contact with mucous membranes or conjunctivae, or mechanical fomite transmission.^35- 38 Direct exposure to small blood volumes probably carries a low risk because experimental and natural infections of white-tailed deer result in only low-level bacteremia.³⁹ However, it is possible that this low risk may be offset by large volumes of animal blood and tissues, such as those to which butchers are exposed.

Another factor related to transmissibility is the blood burden of A phagocytophilum, which appears to increase with immunosuppression resulting in absolute infected neutrophil counts as high as 2.7 to 5.9 × 10⁹/L.^18,40 It is unclear to what degree the sustained dexamethasone treatment of the index case contributed to transmission. The final consideration is the likelihood of health care worker and family member exposure to sufficient volumes of infectious body fluids to account for transmission. It is not unusual for occupational blood exposure to occur among those caring for patients with hemorrhage or during procedures such as intubation or surgery, for which the relative risk is 3 to 4 times higher than for other medical specialties.⁴¹ In western societies, most family members are excluded from these events and health care workers are increasingly protected by training and barriers such as gloves, gowns, and masks.⁴² However, retrospective questioning of our cases clearly indicated that both family members and health care workers not only participated in these events but were unlikely to use gloves and so reported that body surfaces were contaminated by potentially infectious fluids. Moreover, many participants did not acknowledge use of postexposure precautions, such as hand and skin washing.

Although it is likely that routine blood and body fluid precautions will protect against such future events, strict adherence to protective protocols is mandatory even if communicability is deemed unlikely. The lessons of this study remain relevant to the daily hospital and health care unit operations to prevent any additional nosocomial outbreaks of HGA. Moreover, as China advances into its future, it must also now become prepared to deal with the increasing threat that tick-borne rickettsial pathogens have been already brought to the United States and Europe.