Of the 76 personnel who were defined as being at possible risk, we included 53 in the study. The remaining 23 personnel were not included because we were unable to contact them despite several telephone calls or because they were unable to report for blood sampling. We obtained demographic characteristics for the 53 personnel (Table 2). The included personnel provided blood samples for a median of 77 days (range 69–110 days), and samples were analyzed for LASV IgG. Antibodies were not detected in any samples from the 53 personnel.

Of the 53 personnel in the study, 15 had different symptoms of illness during the incubation period (influenza-like symptoms, including sore throat, cough, and rhinitis, n = 12; stomachache, nausea, or vomiting, n = 3; and symptoms of urinary infection, n = 1). Five of these 15 personnel had a temperature >38°C. The 5 personnel who reported fever in conjunction with symptoms possibly suggestive of Lassa fever also provided blood samples during the incubation period. Test results for LASV RNA were negative.

Twelve of the 53 personnel had obtained blood (n = 11), cerebrospinal fluid (n = 1), urine (n = 1), or feces (n = 1) samples from the index patient. Twenty personnel had assisted the patient during lavatory visits and dealt with stool, urine, or both from the patient as a result. Seventeen laboratory personnel had handled blood, cerebrospinal, stool, and urine samples from the patient.

Five personnel and 1 medical student reported that they were not wearing gloves when in close contact with or handling specimens from the patient. Of these 6 persons, 2 doctors, 1 physiotherapist, 1 occupational therapist, and 1 medical student performed physical examinations, and 1 of the laboratory staff did not wear gloves when handling agar plates containing blood from the patient. Two of these 6 persons reported influenza-like symptoms, and 1 of these 2 persons reported a temperature >38°C during the incubation period.

In addition, 5 personnel who were wearing gloves and plastic aprons were exposed to body fluids from the index patient on unprotected undamaged skin or mucous membranes. Of these 5 personnel, 1 was possibly exposed to profuse vomitus on mucous membranes, 1 was possibly exposed to droplets of saliva on her face after the index patient coughed, and 3 were exposed to urine on undamaged unprotected skin. Four of these 5 personnel reported influenza-like symptoms (n = 3) or stomach ache (n = 1); 3 of the 4 reported a temperature >38°C during the incubation period.

Of the 53 personnel, 2 were categorized as being at high risk; all others were categorized as being at low risk after interviews had been conducted. These findings were in contrast to categorization during the acute phase, when all personnel were categorized as being at low risk. During the acute phase, information regarding high-risk exposures was misjudged or inappropriate probably because of less structured questions during the acute phase and a stressful situation. The 2 nurses who had been categorized as being at high risk after the study interview was conducted were the 1 possibly exposed to profuse vomitus on mucous membranes and the 1 possibly exposed to droplets of saliva on her face after the index patient coughed.

The lack of LASV IgG in blood samples from all included personnel who had been exposed to the index patient supports the suggestion that risk is probably low for hospital transmission of LASV when standard barrier nursing methods are used, at least when the symptoms are mild with few occasions of vomiting and diarrhea of the index patient. These results are consistent with those of previous studies, which showed a low risk for person-to-person transmission in hospital settings (6,9–11,13,17). In countries to which LASV is not endemic, ≈40 cases of imported Lassa fever have been reported since LASV was first identified in 1969, and only 1 case of secondary transmission in this type of country has been reported (18). The secondary case was a funeral home employee who had been handling a body before a diagnosis of Lassa fever was made at postmortem analyses.

In addition, 1 case of probable nosocomial transmission was reported in Germany in 2000 (5). A physician who had been examining a patient with Lassa fever and not using any barrier nursing methods was found to be reactive for LASV IgG when high-risk and medium-risk contacts were screened for LASV IgG. These antibodies were specific for the LASV strain isolated from the index patient. However, no specific increase in antibody titer was observed, and secondary transmission could not be confirmed. In the index patient in this study, virus RNA concentration in serum increased concomitant with disease progression. Other than the report from Germany (5), 3 studies have been conducted in countries to which LASV is not endemic in which serologic testing showed no asymptomatic infections in contact persons, including medium-risk and high-risk contacts (9,10,13). Our study strengthens the argument that the risk for asymptomatic infections in nosocomial settings is modest, at least if the disease manifests with relatively mild symptoms.

In our study, the index patient was provided care for 15 days with standard barrier nursing methods. A delay in diagnosis is not uncommon for imported cases (4,11,19) and is probably at least partly dependent on often nonspecific symptoms for Lassa fever. During these 15 days, 5 personnel and 1 medical student were in close contact with the index patient or with specimens from the patient, without use of barrier methods. In addition, 5 other personnel were exposed to body fluids from the index patient on unprotected skin, 1 of them possibly on mucous membranes. Indeed, risk for nosocomial transmission can be even higher before diagnosis when use of barrier nursing methods might be inappropriate, as in the case in Germany (5). Other situations with high risk for nosocomial transmission include profuse excretion of body fluids, such as vomitus, watery stool, or blood with high amounts of virus shedding, or invasive care in an intensive care unit. To avoid nosocomial transmission, barrier nursing methods must always be used by healthcare workers when caring for patients with potentially contagious diseases, according to national recommendations.

Moreover, it is not evident how to categorize the contacts, and there are different suggestions in the literature (11,18,20,21). It is also useful to ask structured questions and to follow criteria strictly when categorizing contacts to avoid incorrect categorization, as was seen in our study. In addition, it is not evident that all viral hemorrhagic fevers should be categorized the same way because there are considerable differences in illness and mortality rates for various viral hemorrhagic fevers (e.g., between Lassa fever and Ebola). In our study, all contacts except for 2 were categorized as being at low risk despite exposure of unprotected skin to body fluids or close physical contact with the index patient. This type of categorization is in contrast to that of the study in Germany, in which these types of contacts were categorized as high risk and prophylaxis with oral ribavirin was administered (5). However, ribavirin is associated with side effects, including pancreatitis and liver injury (22). In addition, prophylactic efficacy has not been demonstrated for humans. Because no secondary transmission of LASV has been proven in contacts with exposure of body fluids to unprotected skin or close physical contact, including in our study, we regard the categorization of Kitching et al. (11) as reasonable. The criteria for high-risk contacts in that study included exposure of body fluids to mucous membranes or damaged skin, such as by needle injury. Of the exposed personnel in our study, no one received ribavirin after risk classification and categorization. Thus, we agree with Kitching et al. that prophylaxis with ribavirin should only be considered in the instance of confirmed, extensive exposure to potentially infected body fluids.

Another aspect of transmission is the degree of illness. A patient with severe symptoms, including profuse vomiting, diarrhea, and bleeding, implies a high risk for virus transmission, and one could also presume that the degree of illness reflects the viremia (23), although data on the association between degree of illness and amount of infectious virus in different body fluids are scarce (5,16,24). However, despite the absence of fever 7 days before admission, the index patient in the present study had moderately high concentrations of LASV RNA in serum at admission and traces of viral RNA (<300 copies/mL) up to 32 days after admission. In addition, she had detectable virus RNA in feces and urine, and virus might be shed from the urine for a long time after recovery (25).

Whether detectable LASV RNA in different body fluids represents living virus or only the viral genome incorporated into dead and dying cells from necrotic tissue that might gain direct access to the circulation is unknown. It is also not clear how transmissible LASV is at different concentrations in different body fluids. Only virus cultivation can determine whether body fluids contain replication-competent virus. In a recent study, results for LASV cultivation in blood were positive for up to 11 days but were negative after fever resolution, although LASV RNA was detectable for a longer period in different body fluids (24). In our study, results of virus cultivation in serum were positive for up to 16 days, even after fever resolution. However, cultivation is difficult to perform, especially from materials such as urine or feces, and is not sufficiently sensitive. Thus, there is no useful method available for distinguishing virus RNA/DNA in different body fluids from living virus or dead and dying cells from necrotic tissue, a prerequisite for evaluating the extent of contagiousness.

Last, we have verified that barrier nursing methods are not consistently defined. In our study, we used basic hygiene procedures, including use of gloves and plastic aprons when persons were at risk for direct contact with body fluids, before diagnosis. Definition of barrier nursing methods of WHO and the Centers for Disease Control and Prevention when caring for a patient with suspected or confirmed Lassa fever includes gloves, gowns, and facial shields or masks and goggles, although WHO emphasizes that these precautions are most necessary when being in close contact (<1 m) with the patient.

In summary, our study strengthens the argument for low risk of secondary transmission of LASV in humans when proper basic nursing methods are used and the disease manifests with relatively mild symptoms. The adequate safety level when caring for patients with suspected or confirmed Lassa fever in countries to which LASV is not endemic should be discussed. Further studies of how infectivity varies depending on severity of symptoms and route of transmission are essential.