Clinical Spectrum, Etiology, and Outcome of Neurological Disorders in the Rural Hospital of Mosango, the Democratic Republic of Congo

Study design.

This study was hospital based and had a prospective observational design, with an outcome assessment at 3 and 6 months after discharge. We present our observations in accordance with the “Strengthening the Reporting of Observational Studies in Epidemiology” guideline and its recent adaptation to the field of neuroepidemiology (Standards of Reporting of Neurological Disorders).¹⁸

Study setting and population.

This study was conducted at the “Hôpital Général de Référence” (HGR) of Mosango, a 350-bed rural hospital located in the province of Kwilu, in the western part of DRC, at about 400 km from Kinshasa. The hospital is run by general practitioners and provides general referral care to about 110,000 inhabitants in a catchment area of approximately 3,350 km². Basic laboratory tests, conventional X-rays, and abdominal ultrasonography can be done locally, but with very limited electricity supply. There are neither imaging nor electrophysiological facilities available for advanced neurological care.

The study population consisted of patients older than 5 years presenting spontaneously, or referred from primary health care facilities, to the HGR of Mosango with neurological disorders. Inclusion criteria were purposively broad, to capture the main endemic neurological infections at an early stage. The precise study definitions and the rationale of all inclusion and exclusion criteria are reported in Supplemental Table 1. Briefly, for study eligibility, at least one of the following symptoms or signs had to be present: 1) altered state of consciousness, 2) changes in sleep pattern, 3) cognitive decline, 4) changes in personality/behavior, 5) recent epileptic seizure (within less than 2 weeks), 6) recent, severe and progressive headache, 7) meningism, 8) new onset cranial nerve lesion(s), 9) new onset sensory-motor focal deficits, and 10) new onset gait/walking disorders. Symptoms/signs had to be either of recent onset or of longer duration (weeks or even months) provided that they were still ongoing at study inclusion. From a clinical perspective, they were severe enough to warrant admission and further evaluation including a diagnostic lumbar puncture (in a setting where the decision cannot be supported by neuroimaging). Exclusion criteria were age of 5 years or below, neurological symptoms/signs related to an obvious recent physical or psychological trauma, chronic sequelae of a past neurological event (e.g., stroke), first seizure having occurred at the age of 5 years or below, and unwillingness or inability to comply with the study requirements (Supplemental Table 1).

Case definitions of priority conditions and other diseases.

We established a list of priority conditions corresponding to severe and treatable infections of the CNS of proven or suspected epidemiological importance in Central Africa, and for which a robust diagnosis and specific treatment were feasible without advanced facilities.¹⁴ The priority conditions were 1) second-stage HAT, 2) cerebral malaria, 3) bacterial meningitis and unspecified meningoencephalitis, 4) tuberculosis of the CNS, 5) neurosyphilis, and 6) HIV-related neurological disorders. The diagnosis was considered as confirmed or probable according to stringent composite case definitions based on pathogen demonstration either at the study site or in external reference laboratories in Kinshasa, DRC, or in Antwerp, Belgium (detailed description in Supplemental Table 2).

For neurological and systemic disorders other than the priority conditions, diagnosis was first made on site by the study neurologist (DM), according to predefined field-adapted case definitions (Supplemental Tables 3 and 4), in the absence of confirmatory tests for most diseases. Whenever possible, an etiological diagnosis was proposed if a combination of presenting features was considered sufficiently specific (e.g., tetanus and cerebrovascular accident). If not, a syndromic diagnosis, either neurological or systemic, was allocated, relying on criteria that best described the case under study. All case record forms were then reviewed for consistency by the study supervisor, specialized in internal medicine (EB), and the conclusions obtained jointly by both the study neurologist and the supervisor were subsequently critically revised by another two experts blinded for the provisory diagnoses, one infectious disease specialist (CY), and one senior neurologist (ASW). The final diagnostic ascertainment was obtained by consensus. In case of coinfection, the most severe disease was kept as the final diagnosis.

Patient management.

During the study period, all general practitioners of the hospital called the study neurologist, or his deputy (LM) if absent, for any patient presenting with neurological symptoms or signs in the emergency ward or in the outpatient clinic. This strategy was set up to maximize the capture of eligible patients. After careful assessment for eligibility, the investigator explained the study purpose and procedures to the potential participant with the help of a translator if necessary. Only patients for whom a written informed consent was obtained (from themselves or their legal representatives) were consecutively enrolled in the study. After inclusion, the study neurologist undertook the full clinical evaluation and the initial data collection. A standardized set of laboratory tests was then systematically performed at the study site by a dedicated and specifically trained laboratory technician (JLRK) according to preestablished standard operating procedures. It consisted of basic tests (blood cell count and differentiation, liver and kidney function tests, etc.), screening and confirmatory assays as defined per protocol (Supplemental Table 2) and sampling for blood cultures. Conventional radiography (e.g., chest and spine) and abdominal ultrasound could be requested whenever clinically indicated.

When there was no clinical contraindication, cerebrospinal fluid (CSF) was systematically obtained by lumbar puncture and submitted to a set of basic and confirmatory analyses, either on site or in external laboratories (Supplemental Table 2). Lumbar puncture was absolutely contraindicated in case of unarousable coma or rapid deterioration of consciousness, hemodynamic or respiratory instability, papilledema on ophthalmoscopy, bleeding diathesis, skin infection at puncture site, or severe vertebral deformities. In case of responsive and nonrapidly progressive altered state of consciousness or focal neurological deficits, lumbar puncture was performed on a case-by-case decision basis, whenever an underlying CNS infection was suspected.

Blood and CSF samples were transported to Kinshasa to perform bacterial cultures in the Unit of Bacteriology of the “Institut National de Recherche Biomédicale” (INRB) and molecular testing for tuberculosis in the Médecins Sans Frontières-supported “Centre Hospitalier de Kabinda”, as well as shipped to the Institute of Tropical Medicine (ITMA), Antwerp, Belgium, for additional investigations and quality control, as per protocol (Supplemental Table 2). Results of the tests performed in external laboratories were not available for case management, except those realized in Kinshasa, which were communicated by phone.

Based on the clinical and laboratory findings immediately available or secondarily obtained, all study participants were treated according to local therapeutic protocols, with advice from expert clinical consultants when needed. The clinical course and response to specific treatment were followed up prospectively during the hospitalization and later on. The final outcome was assessed 3 months after hospital discharge, and again after 6 months for patients with chronic conditions (tuberculosis, HIV, HAT, epilepsy, and psychosis) or with neurological sequelae at 3 months. Possible outcomes at 6 months were death, cure, stabilized chronic disease, moderate or severe neurological sequelae, and loss to follow-up. Chronic diseases were defined as conditions stabilized at 6 months, but for which a complete cure could not be ascertained at that moment. Neurological sequelae were considered as moderate if daily activities were possible but as impaired and severe if the patient could not perform them anymore or remained bedridden.

Ethics statement.

The study protocol, the informed consent forms and all key study-related documents were approved by the Institutional Review Board of the ITMA and by the Ethical Committees of the University of Antwerp, Belgium, and of the Public Health School of Kinshasa, DRC. Written informed consent was obtained from each participant or from his/her legal representative for those of minor age (< 18 years) or in case the neurological condition would not allow for adequate decision. For minors from 12 to 18 years, in addition to the parental consent, informed assent was also checked, according to the DRC law. In case of illiterate patient/guardian, an independent witness was present during the consent interview and countersigned the consent form. Patients with neurological disorders who refused to participate in the study were treated in the same way as the study participants and at the cost of the study (to avoid that free treatment of study participants only would influence the consent). The study was monitored according to the standards of the Good Clinical (Laboratory) Practice guidelines, under supervision of the ITMA Clinical Trials and Tropical Laboratory Units.¹⁹ The study was registered at clinicaltrials.gov under the identifier NCT01589289.

Data management and statistical analysis.

Clinical and laboratory data collected at the study site were copied and double-entered at the INRB, Kinshasa, under the ITMA supervision for data cleaning and accuracy check. Statistical analyses were performed using Stata 14 software. Frequencies of the main etiologies (primary study endpoints) were reported as proportions of the whole study cohort. A bivariate analysis was performed to determine the factors associated with mortality. Those found with a statistically significant association (P value < 0.05) were entered in the multivariate model to identify the independent predictors of death.

Baseline characteristics and outcome.

A total of 351 patients were enrolled from September 2012 to June 2014, and the follow-up period lasted to January 2015. This number corresponds to approximately 20% of all cases referred by the general practitioners for evaluation by the study neurologist. The number of screened patients could not be precisely retrieved because the referral load was overwhelming at some moments. The main reasons for noneligibility were age of 5 years and below, recent physical trauma, and sequelae of a past event.

Baseline epidemiological and clinical characteristics of the study participants are presented in Table 1. Males accounted for 46%, and 51.5% belonged to the age group 19–49 years. The median distance from patients’ homes to the study hospital of Mosango was 36 km, and 33% of study participants came from outside the catchment area (another health district).

Most participants (71.4%) had more than one neurological symptoms/signs at presentation, in various combinations (Table 1). Severe headache was reported by the majority of patients (199, 57.3%). Other frequent clinical presentations were gait/walking disorders (97, 27.6%), epileptic seizure (87, 24.8%), sensory-motor focal deficit (86, 24.5%), and cognitive decline/abnormal behavior (74, 21%); 15.4% of the cases had an alteration of consciousness at admission. Associated nonneurological symptoms and signs were present in 38.5% of the recruited patients, including fever in 28.8%.

The delay before admission was rather long, with a median duration of symptoms/signs of 3 weeks. Forty-seven percent of the study participants had attended another health facility before admission (but less than half of them were directly referred with an accompanying document). Antibiotics or antimalarials had been reportedly administered to 42.4% patients, 17% having received both.

The most frequent past medical histories were tuberculosis (28, 7.9%), psychiatric illness (19, 5.4%), and HAT (12, 3.4%). Arterial hypertension and/or insulin-dependent diabetes mellitus were present in about 15% of the patients.

Main diagnoses.

The CNS infections listed as the priority conditions accounted for 63 (17.9%) of the 351 cases (Table 2). Unspecified meningoencephalitis and bacterial meningitis were the most frequent diagnoses in this group (33, 9.4% together), followed by second stage HAT and HIV-related neurological disorders (10, 2.8% for each diagnosis). Diagnostic investigations for this subset of etiologies are detailed in Table 2.

Other focal/systemic infections with neurological manifestations were confirmed or strongly suspected in an additional 60 cases (17.1%) with respiratory tract infection (sinusitis, otitis, and pneumonia), undifferentiated febrile illness, uncomplicated malaria, and spinal tuberculosis as leading diagnoses in this group. No cases of rabies, poliomyelitis, or leprosy were observed during the study period.

Neurological noncommunicable syndromes were diagnosed in approximately half of all patients, with an important burden of epilepsy (61, 17.3%) and cerebrovascular accidents (23, 6.6%). Neuropathic and myelopathic syndromes as well as degenerative neurological diseases were observed with frequencies ranging from 3% to 5%. Other noncommunicable disorders are listed in Table 2.

Psychiatric disorders were seen in 56 cases (16% of the total load). Of note, several metabolic and vascular diseases (diabetes and arterial hypertension) were occasionally diagnosed in the study participants as the likely etiology of their neurological manifestations.

Frequencies of the priority and other leading conditions according to neurological symptom/sign of presentation.

Table 3 summarizes the frequencies of the main neurological diagnoses according to the different clinical presentations, in other words the pretest probability of a given neurological disease according to the main presenting symptom/sign in this setting. Bacterial meningitis, unspecified meningoencephalitis, HIV-related neurological disorders, cerebrovascular accident, and epilepsy accounted all together for more than half of the patients presenting with altered consciousness, whereas psychiatric disorders and HAT did so for change in sleep pattern. Two third of patients admitted for epileptic seizures were eventually diagnosed with epilepsy. Cognitive decline and behavior disturbance were mainly seen in people with psychiatric disorders and epilepsy. In patients with focal deficits, cerebrovascular accident was by far the main clinical diagnosis, but unspecified meningoencephalitis and HIV-related neurological disorders were rather frequent as well. The leading causes of gait/walking disorders included cerebrovascular accident, spinal tuberculosis, and unspecified meningoencephalitis.

Outcome.

The rate of loss to follow-up was low (11%) taking into consideration the difficult study conditions. A total of 29 enrolled patients died, corresponding to an overall fatality rate of 8.2%, including 16/29 (55%) within 3 days after admission (Table 1). The case fatality rates were 23.8% (15/63), 11.7% (7/60), and 3.1% (7/228) for the priority conditions, the other focal/systemic infections, and the remaining noncommunicable diseases, respectively. At least 76 patients (21.6% of the total inclusions) had moderate or severe sequelae after 6 months of follow-up, whereas the remaining was considered as clinically cured or stabilized.

The risk factors associated with fatal outcome could be analyzed for 311 patients (88.6%) with complete initial and follow-up data (28 deaths and 283 survivors). Features associated with death in the bivariate analysis were symptom duration < 1 week, prior contact with primary health facilities, prior treatment, and several clinical features at presentation (Table 4). When the multivariate model was applied, the only independent predictors of death were the following clinical features: cachexia (defined as body mass index < 20) with an adjusted odds ratio (aOR) of 8.2 (95% confidence interval [CI]: 3–25.2); neck stiffness, aOR 6.0 (95% CI: 1.5–24); altered consciousness, aOR 4.9 (95% CI: 1.7–14.0); and fever, aOR 3.5 (95% CI: 1.3–9.5).