Alkhurma Hemorrhagic Fever in Humans, Najran, Saudi Arabia

TOC summary: Infection was associated with tick bites and contact with farm animals.

A lkhurma virus (ALKV) was discovered in Saudi Arabia in 1995 in a butcher with suspected Crimean-Congo hemorrhagic fever. His fever developed after he had slaughtered a sheep from the city of Alkhurma. Diagnostic testing identifi ed a fl avivirus as the etiologic agent (1,2). Subsequently, ALKV was isolated from the blood of 6 male butchers in Jeddah, and another 4 cases were diagnosed serologically. This disease was named Alkhurma hemorrhagic fever (ALKHF) because the fi rst case was reported from the Alkhurma governorate (1).
After initial virus identifi cation, from 2001 through 2003, another 37 suspected ALKHF cases, of which 20 were laboratory confi rmed, were reported in Alkhumra district, south of Jeddah (3). Among the 20 patients with confi rmed cases, 11 had hemorrhagic manifestations and 5 died.
Full genome sequencing has indicated that ALKV is a distinct variant of Kyasanur Forest disease virus, a virus endemic to the state of Karnataka, India (4). Recently, ALKV was found by reverse transcription-PCR in Ornithodoros savignyi ticks collected from camels and camel resting places in 3 locations in western Saudi Arabia (5). ALKHF is thought to be a zoonotic disease, and reservoir hosts may include camels and sheep. Suggested routes of transmission are contamination of a skin wound with blood of an infected vertebrate, bite of an infected tick, or drinking of unpasteurized, contaminated milk (6).
Several studies have been conducted to describe the characteristics and determinants of ALKHF (1,3,5,6). We conducted a case-control study to assess associated risk factors.

Study Area
The study was conducted in the city of Najran, which is in the southern part of Saudi Arabia on the border with Yemen. It is the capital of Najran region and has a population of ≈250,000. It is an agricultural city in which residents commonly raise domestic animals in their backyards. Cases of ALKHF were found in 6 districts, which were close to each other (within ≈30 km) and rural and in which hygiene was poor.

Case Identifi cation
From 2006 through 2009, laboratory testing for ALKV was performed for Najran residents who sought medical care and whose illnesses met the case defi nition for suspected ALKHF. Infection with ALKV was suspected if a patient had acute febrile illness for at least 2 days; nega-Alkhurma Hemorrhagic Fever in Humans, Najran, Saudi Arabia tive Rift Valley fever, Crimean-Congo hemorrhagic fever, and dengue confi rmatory test results; and >2 of the following: 1) at least 3-fold elevation of alanine transferase or aspartate transferase or clinical jaundice; 2) signs of encephalitis such as confusion, disorientation, drowsiness, coma, neck stiffness, hemiparesis, paraparesis, or convulsions; 3) signs of hemorrhage such as ecchymosis, purpura, petechiae, gastrointestinal bleeding (hematemesis, melena, hematochesia), epistaxis, bleeding from puncture sites, or menorrhagia; and 4) platelet count <100 × 10 9 /L, or lactate dehydrogenase or creatine phosphokinase 2× upper reference level.
In addition, as part of public health surveillance, blood samples were collected from household contacts of patients with laboratory-confi rmed ALKHF. Samples from persons seeking medical care were tested by ELISA for ALKVspecifi c immunoglobulin (Ig) M and IgG by using ALKV antigen as described (7,8) and for viral-specifi c sequence by reverse transcription-PCR (TiBMolbiol, LightMix kit; Roche Applied Science, Basel, Switzerland). Samples from follow-up testing of household contacts were tested by ELISA for ALKV-specifi c IgG. A total of 11 cases were identifi ed through persons seeking medical care whose illnesses met the case defi nition for ALKHF, and another 17 cases were identifi ed through follow-up testing of household contacts.

Case-Control Study
A case of ALKHF was defi ned as illness in any person who lived in the catchment area of Najran General Directorate of Health Affairs and who had serologic evidence of ALKV infection during January 1, 2006-April 30, 2009. All 28 case-patients identifi ed were included in the study. For each case-patient, 2 controls selected from the same house or the nearest neighboring house. A total of 65 controls were enrolled, each of whom was serologically negative for ALKV-specifi c antibodies.
A structured questionnaire was based on information collected during the initial review of the outbreak and asked for demographic data (name, age, gender, nationality, educational status, occupation, marital status, and place of residency), clinical features, and possible risk factors such as exposure to domestic animals. Exposure history for case-patients was limited to the 30 days before onset of illness and for controls to the past 30 days. If the case-patient or control was a child and unable to respond, the child's mother, father, or an older family member helped answer the questions.

Data Analysis
After the samples (frequency and percentage distributions, means, and standard deviations) were described, cross-tabulations were constructed to compare risk factors. To estimate the strength of association, we calculated odds ratios (ORs) with 95% confi dence intervals (CIs). For the multivariate analysis, the most signifi cant variables from the bivariate analysis (at p<0.05) and relevant variables thought to be associated with ALKHF from other studies were selected for backward stepwise inclusion in the model.

Outbreak Description
Of 28 case-patients, 11 (39%) were seeking medical advice and hospitalized for ALKHF, and the others were identifi ed through surveillance among household contacts. Among the 11 hospitalized case-patients, 7 diagnoses of ALKHF were made by PCR and 4 by serologic testing. Among the hospitalized case-patients, only 1 had severe symptoms and was admitted to the intensive care unit. The remaining 10 stayed in the hospital for 5-15 days (mean duration 9.3 ± 3.3 days) and received supportive care, including intravenous fl uid administration and antimicrobial drugs when indicated.

Outbreak Setting
Cases were identifi ed within 6 different districts; most were in the city of Najran. They were in agricultural areas with livestock, some of which lived with the people in the houses.

Patient Demographics and Risk Factors
The mean age of case-patients was 22.3 years ± 11.2 years, and mean age of controls was 25.2 years ± 15.4 years; the difference was not signifi cant (t test, p = 0.657). The age group 20-39 years contained 14 (50%) case-patients and 31 (48%) controls; the age group <20 years contained 13 (46%) case-patients and 24 (37%) controls. No signifi cant difference was noted between case-patients and controls in terms of gender, nationality, education level, marital status, type of home (modern or not), occupation, or district of residence (Table 2).
Among the case-patients, 14 (50%) gave a history of owning or raising domestic animals, compared with 17 (26%) controls (OR 2.82, 95% CI 1.02-7.91). Among those who reported owning or raising domestic animals, all case-patients and controls had sheep on their farms; thus, owning sheep was a signifi cant risk factor (OR 2.82, 95% CI 1.02-7.91). Owning cows and camels was less common; no signifi cant association was found (OR 1.37, 95% CI 0.30-5.86 and OR 10.29, 0.84-279.2, respectively). Risk for infection was signifi cantly higher for those who lived <100 meters from farms (OR 4.00, 95% CI 1.40-11.75) than for those who lived farther from farms (Table 3).

Discussion
Our fi ndings that some patients had subclinical illness and that no deaths were documented among the 28 casepatients suggest that previous studies may not have characterized the full spectrum of ALKV-associated illness and that case-fatality rates as high as 25% may have resulted from detecting only severe cases of ALKHF. In addition, our identifi cation of multiple seropositive members within families suggests the occurrence of family-based clusters of ALKV infection.
ALKV-positive persons with subclinical disease, identifi ed by the active surveillance system, did not undergo thorough clinical and laboratory investigations and were not directly observed by a clinician. Because of this lack of observation and because of the inability of some casepatients to recall and report low-grade fever within 1 month before onset of illness, only 53% of case-patients reported fever, but all case-patients who were hospitalized had fever. Our fi ndings underscore the role of the high percentage of case-patients with subclinical infection in the epidemiology of this disease; seroprevalence studies should be encouraged. A similar seasonal pattern of disease (March-July) was found in western provinces (Jeddah and Makkah) among 11 case-patients who recovered during 1994-1999. This fi nding might support evidence of disease association with the peak activity of ticks in the beginning of March (9,10).
Similar to other hemorrhagic fevers, ALKHF showed no predilection for patient age, gender, or nationality (11). Risk factors identifi ed by this study included a broad array of activities associated with animal exposures but most significantly with direct contact with animals. Similarly, a higher proportion of case-patients than controls owned or raised animals; however, no difference was noted in the proportion reporting abnormalities in their animals. This fi nding might suggest the low virulence of the virus in animals and highlights the need for animal studies. Among the animals raised, sheep were signifi cantly associated with the disease.
Although we found no signifi cant association between ALKHF infection and livestock-related occupations such as butchering, we found a high association for history of slaughtering livestock. These fi ndings agree with those of other studies conducted in the cities of Jeddah and Makkah (1,3). Furthermore, we found that another major risk factor for human infection was direct contact with blood or body fl uids from animals. Ingestion of unpasteurized milk has been noted as a risk factor in previous studies; the mode of transmission is not yet clear but has been suggested to result from contamination of the milk (6,12). However, our multivariate analysis found no association.
Only 10 of the 28 interviewed case-patients had a history of a tick bite within 1 month before symptom onset; however, tick exposure was signifi cantly more common among case-patients than among controls. The association of tick bites and ALKV was supported by Charrel et al., who detected ALKV RNA in ticks (Ornithodoros spp.) collected from camels and camel resting places in western Saudi Arabia (5).
ALKV has been identifi ed only in the southern and western parts of Saudi Arabia. Given our current evidence that subclinical ALKV infections occur in humans, the virus may be more widespread in Saudi Arabia than previously realized. We are conducting studies to further characterize the distribution of ALKV in Saudi Arabia. In addition, the history of the reported disease in Makkah during the Hajj, when thousands of livestock are imported to Saudi Arabia, and the existence of the outbreak in Najran, which is at the border with Yemen, necessitate further studies in adjacent countries (3,5). Additionally, ALKV is closely related to Kyasanur Forest disease virus, which has been well characterized in India. The possibility remains that this virus has a wider geographic range in the Middle East and central Asia than previously realized.
Because our investigations were retrospective, we cannot exclude recall bias about exposure, but as long as the controls were from the same households, they were exposed to some of the questions during the surveillance done by the preventive department. Furthermore, the investigators who administered the questionnaires based the questions on the month before the interview, which should minimize recall bias.
ALKHF is a zoonotic disease with clinical features ranging from subclinical asymptomatic to severe complications. This study highlights the different activities related to exposure to animals and tick bites in the transmission of ALKV to humans; it found no signifi cant association with mosquitoes. Further studies are needed to understand the role of livestock, wildlife, and ticks in the maintenance of the virus and the risk factors so that public health measures can be developed to reduce the extent of the disease in humans.
Funding for this research was provided by the Ministry of Health, Kingdom of Saudi Arabia.
Dr Alzahrani is an epidemiologist with the Field Epidemiology Training Program, Ministry of Health, Kingdom of Saudi Arabia. His interests include the epidemiology of emerging infectious diseases. †aOR for risk factors (contact with domestic animals, tick bites, adjacent farm distance) after elimination of nonsignificant variables (drinking unpasteurized milk and owning or raising domestic animals) calculated by using backward stepwise strategy.