We report a case of dengue fever in a Boston-area health care worker with no recent history of travel but with mucocutaneous exposure to infected blood from a febrile traveler who had recently returned from Peru. Serologic tests confirmed acute dengue virus infection in both the traveler and the health care worker. We believe that this is the first documented case of dengue virus transmission via the mucocutaneous route. We present case reports and review other ways that dengue virus has been transmitted without a mosquito vector.
Dengue fever, a mosquito-borne viral infection caused by 4 antigenically distinct dengue virus serotypes in the family Flaviviridae, is widespread in tropical and subtropical regions. Infection is characterized by the abrupt onset of fever, myalgia, fatigue, and headache. Diffuse erythema may be present early in the course of infection, and maculopapular eruption may be present later. Associated complications include dengue hemorrhagic fever and dengue shock syndrome, which are diagnosed on the basis of the presence of fever, thrombocytopenia, hemorrhage, and excessive vascular permeability. Laboratory findings commonly include leukopenia, thrombocytopenia, and abnormalities in the results of liver function tests.
We report a case of dengue fever in a health care worker with no recent history of travel outside of the northeastern United States. The source of her infection was a traveler who had recently returned from Peru; the presumed mechanism of transmission was mucocutaneous exposure that occurred during medical care. We describe both cases of dengue fever and review the ways that dengue virus can be transmitted without mosquito vectors.
Case reports. Patient 1, a 48-year-old female traveler, presented in November 2002 with a 5-day history of fever, myalgia, and headache. She had recently returned from Iquitos, Peru (on a trip that lasted from 1 through 10 November), where she worked as a nurse on a medical mission and traveled through the Amazon jungles. She stayed in a hotel that had unscreened, open windows. The patient had received hepatitis A virus, hepatitis B virus, and oral typhoid vaccines prior to travel, and she took mefloquine hydrochloride for malaria prophylaxis. She had received yellow fever vaccine in 1997.
At examination, the patient's temperature was 37.9°C. Abnormal findings included marked erythroderma, especially on the patient's back, and a maculopapular rash along her hairline. She had generalized erythema and a faint maculopapular confluent rash on her arms, legs, and back. Laboratory results obtained on 18 November (i.e., day 2 of illness) are summarized in table 1, along with the results of subsequent studies.
Patient 2, a 37-year-old health care worker, was seen on 19 December 2002 for residual fatigue, myalgia, headache, and low-grade fever, which occurred after an acute illness that began on 28 November 2002 and that included eye pain, nosebleeds, and decreased appetite. Ten days prior to the onset of her symptoms, the patient had contact with blood from patient 1. While patient 2 was transferring blood from a syringe to a blood culture bottle, the needle dislodged from the syringe, and she felt blood splash onto her face, including her eye, nose, and mouth. She had a history of 2 previous occupational needlestick exposures (in 1995 and 1999) and had been vaccinated against hepatitis B virus in 1995. Results of tests for HIV were negative after those incidents. Her only international travel was to the Bahamas 20 years before the exposure occurred; she denied recent travel to Texas or Florida.
Serum samples obtained from both patients were submitted to the Dengue Branch of the Centers for Disease Control and Prevention (CDC) in Puerto Rico. No virus was isolated from patient 1 on day 6 of illness, but dengue virus IgM titers were positive, and IgG titers were positive at 1:163,840. The convalescent-phase serum sample (obtained on day 23 after the onset of illness) was IgM positive. IgG titers were positive at 1:655,360, which is consistent with a secondary flavivirus infection. A serum sample from patient 2 (obtained on day 8 of illness) was IgM positive and IgG negative. Virus isolation was not attempted. A convalescent-phase serum sample from patient 2 (obtained on day 22 of illness) was positive for IgM and IgG at titers of 1:2560. Neutralization antibody tests showed antibodies against dengue serotype 2 and dengue serotype 3 in patient 1 and antibodies against dengue serotype 3 in patient 2 (figure 1).
Discussion. The usual mosquito vectors for dengue virus are Aedes aegypti and, less frequently, Aedes albopictus. Vector control programs initiated after World War II eliminated A. aegypti from most parts of the Americas, but, after the cessation of the programs in the 1960s, A. aegypti reinfestation occurred in most countries, including their urban areas. Dengue has become widespread in the Americas, and its incidence has been increasing .
Although dengue epidemics occurred in the United States decades ago, most of the recent cases of dengue have occurred among international travelers. Autochthonous transmission can occur in areas that have competent vectors, such as Texas and Florida. In 2001–2002, local transmission of dengue virus in Hawaii resulted in >100 cases of infection .
The diagnosis of dengue is confirmed by isolating the virus from serum either by inoculating live mosquitoes or by performing cell cultures [3, 4]. Although these methods are specific, the sensitivity may be as low as 50%, and the procedures take at least 1 week to perform. Viremia typically begins 2–3 days before the onset of symptoms, and it continues for 4–5 days during acute illness . Hence, there is a limited period during which the virus can be isolated. Among Thai children, viral RNA levels peaked 2 days before defervescence . PCR detects viral RNA [5, 6], can be performed more rapidly than can virus isolation but can detect viral RNA only during the period of viremia, and has a sensitivity similar to that of viral culture.
Because of the drawbacks of culture and PCR, testing paired acute-phase and convalescent-phase serum samples by use of ELISA has become the primary technique for the diagnosis of dengue [7, 8]. Detection of IgM in acute-phase serum samples usually supports the diagnosis of dengue, although IgM may be undetectable during the early stages of the infection. A 4-fold increase in the levels of antibodies in specimens obtained from the acute phase to the convalescent phase strongly supports the diagnosis of acute dengue virus infection. However, dengue virus antibodies cross-react with those of many other flaviviruses, including West Nile virus, Japanese encephalitis, and yellow fever viruses. Acute or past infections with other flaviviruses, or vaccination against them (in the case of yellow fever virus or Japanese encephalitis), can complicate the interpretation of dengue serologic findings.
For patient 1 (the traveler), the results of ELISA performed on the acute-phase serum samples were dengue virus IgG positive, which could be attributed to past flavivirus infection or to immunization against yellow fever . Although no virus was isolated from the serum samples obtained on day 6 of illness, a serum sample obtained on day 4 was IgM negative but became IgM positive on day 6, which is consistent with previous reports on the timing of seroconversion . In addition, convalescent-phase serum samples tested positive for both IgM and IgG, with a 4-fold increase in IgG titers, strongly supporting the diagnosis of acute dengue virus infection.
For patient 2 (the health care worker), the diagnosis was more straightforward because of the absence of past exposure to flaviviruses. Although transmission of West Nile virus has been occurring in the northeastern United States since 1999, the December onset of illness for patient 2 occurred after the transmission season ended in Boston. The results of ELISA were positive for dengue virus IgM on day 7, strongly suggesting the diagnosis of acute dengue virus infection. Although the 2 women were linked epidemiologically, positive results of viral cultures with genetically identical isolates from both would have been required for absolute proof of linked infections.
Analysis of neutralizing antibodies against dengue viruses identified serotypes 2 and 3 in patient 1 and serotype 3 in patient 2. The plaque reduction neutralization test uses reference viruses (expressed as plaque-forming units), which are mixed and incubated with diluted test sera in cell culture. Antibodies that neutralize specific serotypes of the virus reduce the number of plaques formed ; therefore, the specific dengue serotype can often be determined.
Although rarely documented, dengue virus transmission without a mosquito vector has been reported. The routes of transmission include needlestick injuries, bone marrow transplantation, and intrapartum and vertical transmission (table 2). A brief report on ProMed-mail described 2 suspected cases of dengue fever aquired through blood transfusion in Hong Kong; in both cases, the donor developed symptoms consistent with dengue 1 day after giving blood and tested positive for dengue infection .
Dengue virus presumably infected patient 2 via blood contact with mucous membranes. This is biologically plausible, given the well-documented nosocomial spread of multiple viruses (i.e., hepatitis B virus, hepatitis C virus, and HIV) after mucocutaneous contact with blood . The mean volume of blood delivered via a needlestick injury with a 22-gauge needle attached to a syringe containing 2 mL of blood has been found to be only 1.40 µL, yet transmission of many infections, including dengue, has occurred . The amount of blood associated with mucocutaneous transmission of pathogens has not been defined. Because the level of viremia can reach 109 RNA copies per milliliter of blood in acute dengue virus infections, it is plausible that blood splashed on broken skin or on a mucous membrane could deliver a sufficient amount of virus to cause infection .
Nosocomial transmission, including mucocutaneous transmission, may occur in areas of endemicity but is unlikely to be recognized in areas in which dengue virus circulates widely. Hemorrhage, a feature of dengue hemorrhagic fever, may increase the risk of nosocomial transmission. Assessing the magnitude of nosocomial dengue virus transmission in areas of endemicity is difficult because all health care workers are also potentially exposed to infective mosquitoes. It is not surprising that nosocomial transmission of dengue virus has been identified primarily in areas of nonendemicity, in settings in which no other exposures to the virus are plausible.
West Nile virus is transmissible via breastfeeding, as well as through blood transfusions, organ transplantations, stem cell transplantations, intrauterine exposure, and needlestick injuries [23,24,25,26–27]. It is possible that dengue virus could be transmitted also through breast milk, although no documented cases have been reported.
Health care workers have frequent exposures to blood. One recently published survey found that 43% of physicians, 38.5% of registered nurses, 26.4% of licensed practical nurses, and 24.5% of medical technologists reported at least 1 mucocutaneous blood exposure within the previous 3 months . Adherence to standard precautions was not ideal, and the underreporting of incidents was common. Phlebotomy equipment that operates as a closed system could minimize the number of blood-splash and needlestick exposures.
In summary, health care workers should be aware that nosocomial transmission of dengue virus can occur by mucocutaneous exposures, as well as by needlestick exposures. This may be especially relevant to health care workers who care for patients with dengue with hemorrhage in resource-poor areas in which gloves and access to good infection-control measures are limited.
We thank the Dengue Branch of the Centers for Disease Control and Prevention, for performing the dengue studies; Dr. Vance Vorndam, Dr. Timothy Brewer, and Kerrie Dirosario, for reviewing the manuscript; and Dr. Joycelyn Datu, Dr. Stanley Sagov, and Diane Boivin, for their helpful comments.
Conflict of interest. All authors: No conflict.