Figure 1

Figure 1. Geographic and temporal distribution of laboratory-confirmed Lassa fever cases, Nigeria, January 1–May 6, 2018. A) Geographic distribution of laboratory-confirmed cases by state. Gray shading indicates states reporting no laboratory-confirmed cases. Locations...

During the study period (January 1–May 6, 2018), a total of 1,893 suspected LF cases were reported, including 10 probable cases and 423 laboratory-confirmed cases. The laboratory-confirmed cases were reported from 20 states and the Federal Capital Territory. Most (80.6%) of the laboratory-confirmed cases were reported from the 3 states with a dedicated LF treatment center: Edo (178 cases), Ondo (99 cases), and Ebonyi (64 cases) (Figure 1, panel A). LF incidence was highest in these 3 states (Table 1). Edo and Ondo also had the largest number of laboratory-negative suspected cases. The positive rate (i.e., the proportion of the number of laboratory-confirmed cases among all persons with suspected cases tested) was 22.5% (423/1,883) nationally, ranging from 3.4% (Lagos) to 70.0% (Delta). Among the 3 states with the highest number of cases, the positive rates were 16.5% (Edo), 31.6% (Ondo), and 69.6% (Ebonyi).

CFR among the laboratory-confirmed cases was 25.1% (106/423 [95% CI 20.9%–29.2%]). Among the 3 most affected states, CFR was 14.6% (Edo), 24.2% (Ondo), and 23.4% (Ebonyi) (Table 1). Among the 423 cases, a total of 414 CIFs with detailed information, including demographic information, onset date, symptoms, exposure history, and ribavirin administration, were collected (collection rate 97.9%). However, the data in some fields of the CIFs were incomplete. For example, the onset date was unknown in 2.4% (10/414) of cases, and symptom information was missing in 12.8% (53/414). An epidemic curve based on the onset date for laboratory-confirmed cases peaked at epidemiologic week 5 in 2018 (Figure 1, panel B). No statistically significant change was observed in the national CFR throughout the outbreak (p value for trend = 0.41).

We analyzed the age and sex distribution of the 414 patients with laboratory-confirmed cases (Table 2). Median age was 32 years (interquartile range 20–44 years); 157 (37.9%) were female and 257 (62.1%) male. CFR was lowest in children <10 years of age (11.1%) and highest in adults >61 years of age (38.2%); the aOR of fatal outcomes in the elderly group was 4.9 (95% CI 1.5–15.6) compared with children <10 years of age. Adults 41–60 years of age also had statistically higher CFRs compared with children <10 years of age (Table 2). CFR was higher for male patients (26.6%) than female patients (21.8%) but was not significantly different; the aOR of fatal outcomes in male patients compared with female patients was 1.3 (95% CI 0.81–2.1).

The most common signs and symptoms among patients with laboratory-confirmed LF were fever (96.4%, 348/361), headache (58.7%, 210/358), vomiting (49.4%, 177/358), fatigue (43.3%, 155/358), and abdominal pain (40.2%, 144/358) (Table 3). Hemorrhaging was observed in 17.0% (61/358) of these patients. Cough (p = 0.0050), hemorrhaging (p<0.001), and unconsciousness (p = 0.0018) were significantly more prevalent in fatal than nonfatal cases.

Figure 2

Figure 2. Exposure history and case-fatality rate among patients with laboratory-confirmed Lassa fever cases, Nigeria, January 1–May 6, 2018. A) Proportion of persons reporting Lassa fever exposure risks for all case-patients, HCWs, and...

During the 3 weeks before symptom onset, 17.7% (62/350) of case-patients reported contact with patients who had known suspected or confirmed LF, 17.0% (56/330) reported contact with rodents or their urine or feces, and 2.9% (9/315) reported attendance at a burial ceremony (Figure 2, panel A). During the study period, 5,012 persons were determined to have had contact with confirmed case-patients; follow-up was conducted for 5,001 of them. During the 21-day follow-up period, 81 contacts experienced onset of symptoms, and 28 were found to have laboratory-confirmed LF. The positive rate among symptomatic contacts was 34.6% (28/81), and the secondary attack rate was 0.56% (28/5,001; 95% CI 0.35%–0.77%).

During the study period, 37 HCWs were infected, resulting in 8 deaths (CFR 21.6%). The incidence of HCW infections was distributed throughout the outbreak period (Figure 1, panel B). A significantly high proportion of HCWs (55.9%, 19/34) reported contact with patients with known suspected or confirmed LF compared with non-HCWs (p<0.001) (Figure 2, panel A).

Ribavirin was administered to 94.1% (334/355) of the patients with laboratory-confirmed cases. CFR for patients who received ribavirin was 20.7% (69/334), compared with 71.4% (15/21) for patients who did not receive ribavirin (Figure 2, panel B). We also analyzed the subset of patients who survived >7 days after symptom onset to account for the possible effect of the difference in clinical conditions. We further divided the patients who received ribavirin into 2 groups: patients who received the drug within 7 days after symptom onset and patients who received the drug after that point. Among case-patients who survived >7 days, CFR was significantly higher for patients who did not receive any ribavirin (66.7%, 12/18) compared with patients who received the drug (p<0.01), whether receipt of the drug occurred within 7 days of symptom onset or >7 days after onset (Figure 2, panel C). CFR was lower among patients who received the drug within 7 days of symptom onset (12.5%, 15/120) compared with patients who received the drug after that point (20.1%, 38/189), although this difference was not significant (p = 0.095). Because this reduction in CFR might have been attributable to not only ribavirin but also the other supportive treatments provided, the time between symptom onset date and the patient’s arrival at a health facility was included in addition to age and administration of ribavirin as a covariable in the binomial logistic regression analysis for fatal outcomes. Although absence of ribavirin administration (p<0.001) and advanced age (p = 0.025) remained significant factors in fatal outcomes, delay in visiting health facility did not (p = 0.19).

We describe the epidemiologic and clinical features of the LF outbreak in Nigeria during January 1–May 6, 2018. A total of 423 laboratory-confirmed cases were reported during the study period. Most of the laboratory-confirmed cases were reported from the 3 states (Edo, Ondo, and Ebonyi) where dedicated LF treatment centers are located; disease incidence was also highest in these areas. However, the positive rate among suspected cases was not especially high for these 3 states. In addition to the high prevalence of LF in the areas, HCWs in these 3 states had a high suspicion of LF in patients with high-grade fevers, which might have led to increased testing. Conversely, suspicion of LF might be low in some areas other than these 3 states. Lack of LF expertise and diagnostic capacity in these other areas might have discouraged active detection of LF patients, leading to underreporting of the disease. Also, LF incidence and prevalence might actually be lower in some areas for epidemiologic and environmental reasons, such as low prevalence of the virus in rodents or good hygiene practices that help reduce contact between humans and rodents. To clarify different LF prevalence by areas, sensitization and strengthening of the surveillance system to detect suspected cases and obtain test samples are required, particularly in states other than Edo, Ondo, and Ebonyi. In addition, seroprevalence surveys and ecologic studies of LF in humans and rodents should give further insight on the actual burden of the disease.

We used 2 protocols for molecular diagnosis of LF to cover the heterogeneous Lassa virus in Nigeria (20), and we tested >95% of the samples by using both protocols during the outbreak. Further study is needed to reveal sensitivity and specificity of each protocol in this country. That information would also give us further insight on genetic diversity of the virus in natural hosts in the country.

Although NCDC situation reports showed the peak of the outbreak at week 7 (17), our study found that the outbreak peaked at week 5. This difference occurred because the epidemic curve from the situation reports was based on the reporting date, whereas the epidemic curve in our study was based on symptom onset date. The difference can be explained by the time lag between symptom onset to health facility presentation and subsequent diagnosis and reporting. The median of the time lag between symptom onset to suspicion of LF was 7 days (interquartile range 4–11 days), and several additional days were required for a health facility to report through the DSNO for each Local Government Area and for the state epidemiologist in each state to report NCDC. Further strengthening of the surveillance system is required to shorten this time gap.

Girls and women accounted for a lower proportion of the laboratory-confirmed LF cases than boys and men (37.9% vs. 62.1%). Past studies have shown no or little difference in LF incidence between male and female patients (23–25). It is unclear whether the difference in this study came because men and boys were at higher risk for infection or more susceptible to the disease than women and girls or because ascertainment of cases in women and girls was low during this outbreak.

CFR among laboratory-confirmed cases during the study period was 25.1% and did not change substantially throughout the outbreak. CFR can reach 50% in hospitalized patients during epidemics (8), whereas the observed CFR among patients hospitalized with severe LF is 15% (7). CFR in this outbreak was especially high among elderly patients. Such difference in CFR among age groups was not observed in previous studies (26,27). The large number of cases in this outbreak might have increased the statistical power to detect the difference.

Gastroenteric symptoms, including vomiting, abdominal pain, and anorexia, were frequently observed among patients with laboratory-confirmed cases (>30%), whereas hemorrhaging was only observed in 17.0% of case-patients (Table 3). This observation is consistent with previous reports (28,29). Because some symptoms, such as cough, hemorrhaging, and unconsciousness, were more frequently observed for fatal cases than nonfatal cases, these symptoms might be predictors for fatal outcomes. Patients with such symptoms would require more attention to achieve better clinical outcomes.

Lassa virus is primarily transmitted to humans from rodents. The virus is also occasionally transmitted through the body fluids of infected persons (3). In this outbreak, ≈17% of patients reported a history of contact with rodents. A similar percentage reported contact with patients with suspected or confirmed LF. Because this information was obtained by patient interview, recall bias might have influenced the accuracy of exposure history. Although the secondary attack rate was as low as 0.56%, the positive rate for LF among person with suspected cases was higher for those who had contact with confirmed case-patients than that in the general population (34.6% vs. 22.5%). Contact tracing did not account for 34 LF patients, although they had reported contact with other confirmed or suspected case-patients. Some case-patients might have been targeted in the contact tracing whereas others were not because they reported contact with persons with suspected (but not confirmed) cases. Strengthening and expanding contact tracing is required but posed a challenge in resource-limited settings during the outbreak. A high rate of contact with suspected or confirmed LF case-patients among HCWs suggests the possibility of nosocomial infections, although we could not rule out another source of infection, such as rodents (30). Nevertheless, good infection prevention and control practices and readily available personal protective equipment are important to protect HCWs from infection with the virus (5,6).

The findings in this study support the effectiveness of ribavirin in reducing mortality rates from LF. CFR was lowest among patients who received treatment with ribavirin. Patients in severe conditions might have not received ribavirin because they died before reaching healthcare facilities where treatment was available; that is, the difference in CFRs between patients who did or did not receive ribavirin might be attributable not only to the ribavirin treatment but also to the patient’s clinical condition before ribavirin administration (6,28). To explore this issue further, we analyzed a subset of patients who survived >7 days after symptom onset. The highest CFR was still observed among patients who did not receive ribavirin; CFR was lower even if provision of ribavirin was deferred to >7 days after symptom onset. Furthermore, the early commencement of ribavirin treatment reduced CFR compared with deferred administration of the drug, although this difference was not statistically significant. Therefore, the reduction in CFR was more likely attributable to ribavirin than to the other supportive treatments provided. These findings support Nigeria’s national guidelines for clinical management of the disease, which advise that patient outcomes are more favorable when ribavirin treatment is commenced earlier (9).

This study has several limitations. Although we collected CIFs from 414 of the 423 patients with laboratory-confirmed cases, the data in some fields were incomplete. Availability of CIFs for patients with suspected cases whose laboratory tests were negative would have enabled us to conduct a case–control study to better determine the risk factors for the disease; however, collection of this information for 1,460 laboratory-negative suspected cases would have been burdensome to outbreak response staff. Also, our study did not provide any insights into why the 2018 LF outbreak was larger than those in previous years. Studies using viral genomic data have suggested that the outbreak was not caused by ongoing human-to-human transmission from a single source but by multiple environmental sources (31). Better surveillance through increased availability of laboratory testing of suspected cases might have played a role in the larger number of laboratory-confirmed cases during this outbreak. Localized clusters, which were not investigated as part of this study, might have contributed to the inflation of case numbers in some areas.

LF is endemic in Nigeria. The LF surveillance data from Nigeria, even with their limitations, are arguably the best longitudinal data collected on LF globally. Although these data mostly relate to the 2018 outbreak, data collection has continued, and all aspects of surveillance are being continuously improved (e.g., educating HCWs on the case definition, increasing ease and efficiency of the transport of samples, and increasing capacity and quality of diagnosis). In this study, we described the epidemiologic and clinical features of the largest recorded LF outbreak, which had a high CFR. The investigation also revealed several risk factors for fatal outcomes and the contribution of early treatment in reduction of CFR. These findings should lead to further investigation of the disease. Our study also highlights the need for specific incidence and seroprevalence surveys to determine the actual burden of disease in Nigeria and West Africa. Although the emergency phase of this outbreak was declared over on May 10, 2018, a small number of cases in some areas continued to be reported thereafter (10). Ceaseless efforts to improve risk communication, surveillance, laboratory diagnosis, clinical management, infection prevention and control practices, logistics, and coordination could mitigate the impact of future outbreaks.