ARTICLES
Interferon alfa-2a in Japanese encephalitis: a randomised double-blind placebo-controlled trial
Tom Solomon, Nguyen Minh Dung, Bridget Wills, Rachel Kneen, Mary Gainsborough, Tran Vinh Diet, Tran Thi Nhu Thuy, Ha Thi Loan, Vo Cong Khanh, David W Vaughn, Nicholas J White, Jeremy J Farrar Summary Background Japanese encephalitis virus (JEV), although confined to Asia, causes about 35 000–50 000 cases and 10 000 deaths every year, and is the most important cause of encephalitis worldwide. There is no known antiviral treatment for any flavivirus. Results from in-vitro studies and work in animals have shown inteferon alfa has antiviral activity on Japanese encephalitis and other flaviviruses; therefore, we aimed to assess the efficacy of inteferon alfa2a in Japanese encephalitis. Methods We did a randomised double-blind placebocontrolled trial of interferon alfa-2a (10 million units/m2, daily for 7 days) in 112 Vietnamese children with suspected Japanese encephalitis, 87 of whom had serologically confirmed infections. Our primary endpoints were hospital death or severe sequelae at discharge. Analysis was by intention to treat. Findings Overall, 21 children (19%) died, and 17 (15%) had severe sequelae. Outcome at discharge and 3 months did not differ between the two treatment groups; 20 children in the interferon group had a poor outcome (death or severe sequelae), compared with 18 in the placebo group (p=0·85, difference 0·1%, 95% CI –17·5 to 17·6%), there were no longterm side effects of interferon. Interpretation The doses of interferon alfa-2a given in this regimen did not improve the outcome of patients with Japanese encephalitis. Lancet 2003; 361: 821–26 Published online February 11, 2003 http://image.thelancet.com/extras/02art9329web.pdf Department of Neurological Science, University of Liverpool, Walton Centre for Neurology and Neurosurgery, Liverpool, UK (T Solomon MRCP); Centre for Tropical Medicine, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, UK (B Wills MRCP, Prof N J White FRCP, J J Farrar FRCP); Roald Dahl EEG Unit, Royal Liverpool Children’s NHS Trust, Liverpool, UK (R Kneen, MRCP); Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam (N M Dung MD, T V Diet MD, H T Loan MD, V C Khanh MD, T T N Thuy MD); University of Oxford-Wellcome Trust Clinical Research Unit, Hospital for Tropical Diseases, Ho Chi Minh City (T Solomon, BWills, R Kneen, M Gainsborough MRCP, J J Farrar); Department of Virology, US Army Medical Component, Armed Forces Research Institute of Medical Sciences, Bangkok, Thailand (D W Vaughn MD); Wellcome Trust-Mahidol University-Oxford Tropical Medicine Research Programme, Faculty of Tropical Medicine, Mahidol University, Bangkok (N J White). Correspondence to: Dr Tom Solomon, Department of Neurological Science, University of Liverpool, Walton Centre for Neurology and Neurosurgery, Lower Lane, Fazakerley, Liverpool L9 7LJ, UK. (e-mail:
[email protected])
THE LANCET • Vol 361 • March 8, 2003 • www.thelancet.com
Introduction Encephalitis caused by arthropod-borne viruses (arboviruses) is recognised as an increasingly important problem.1 For example during 2002, West Nile virus, a member of the genus Flavivirus family Flaviviridae, caused its largest ever encephalitis outbreak with more than 3800 cases and 225 deaths, in North America.2 However the most important member of this serogroup wordwide is the closely related flavivirus, Japanese encephalitis virus (JEV) which causes about 35 000–50 000 cases and 10 000 deaths per year.3 JEV is a small single-stranded positive-sense RNA virus transmitted between birds, pigs, and other vertebrate hosts by mosquitoes, mainly Culex tritaeniorhynchus. JEV occurs in southeast Asia, China, the pacific rim, and Asian subcontinent, but its geographical range is expanding, with recent outbreaks in Nepal and northern Australia.4 In rural Asian countries most people are infected during childhood, but few infections in human beings result in symptoms. If disease does occur, patients usually present with severe meningoencephalitis, which is often associated with seizures.5 A smaller proportion present with aseptic meningitis, or a poliomyelitis-like acute flaccid paralysis.6 Roughly 30% of patients with Japanese encephalitis die, and half of survivors have severe neurological sequelae, which imposes a large socioeconomic burden in the poor rural settings where Japanese encephalitis occurs. In addition to JEV and West Nile virus, other flaviviruses that cause encephalitis include: St Louis encephalitis virus, which is endemic in the Americas; Murray Valley encephalitis virus, present in Australasia; and tick-borne encephalitis virus that occurs across Europe and the former Soviet Union.7 There is no known effective antiviral treatment for these or any of the other diseases caused by flaviviruses, such as dengue and yellow fever. Indeed, until now, none has been assessed in a controlled trial. For many years, corticosteroids have been used for Japanese encephalitis, but a randomised placebocontrolled trial of dexamethasone in 40 patients showed no benefit of this treatment.8 Interferon alfa is a glycoprotein cytokine that is produced naturally in response to viral infections, including Japanese encaphalitis.9 Interferons are not directly antiviral, but induce production of effector proteins in cells, which inhibit various stages of viral replication, assembly, or release.10 Recombinant interferon alfa has become the standard treatment for chronic hepatitis B and C. In tissue culture, recombinant interferon is effective against JEV and other arboviruses, including West Nile virus.11,12 This treatment has been given to 14 Thai patients with Japanese encephalitis, and although there has been a suggestion of benefit,13,14 there have been no randomised double-blind trials of interferon alfa in any viral encephalitis. We aimed to assess the efficacy of interferon alfa-2a for Japanese encaphalitis.
821
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Methods Patients Patients were recruited from the paediatric intensive-care unit at the Centre for Tropical Diseases, an infectious diseases referral centre for southern Vietnam in Ho Chi Minh city. The study protocol was approved by the hospital’s scientific and ethical committee, and informed consent was obtained from the accompanying relative. Between October, 1996 and October, 1999, we recruited children aged between 1 and 15 years who had clinically diagnosed viral encephalitis. On the basis of results from previous studies,5,6,15 encephalitis was diagnosed clinically in children who had a history of fever that lasted less than 14 days, and who had either convulsions or altered consciousness (Glasgow coma score ⬍15,16 or for children younger than 6 years, Blantyre coma score ⬍517); and a cerebrospinal fluid (CSF) cell count less than 1000 cells per L, a ratio of cerebrospinal fluid to plasma glucose greater than 40%, and a negative CSF gram stain. JEV was confirmed by virological tests. We excluded children aged between 6 months and 5 years who had had simple febrile convulsion.18,19 We also excluded those with epilepsy, a positive blood slide for asexual Plasmodium falciparum parasites or malaria pigment,20 and those referred from a peripheral health centre who had features suggestive of bacterial meningitis and for whom antibiotic treatment was documented. Procedures Because patients with severe neurological sequelae make an important contribution to the disease burden of Japanese encephalitis, we decided that to have death as the only primary endpoint was not appropriate. Therefore, primary endpoints were hospital deaths or severe sequelae at discharge. We assessed patients at discharge for disability in a range of activities (speech, comprehension, feeding, sitting, standing, walking, urinary continence, and faecal continence) and did neurological examinations. Patient outcomes were then classified into one of five predetermined outcome categories:5,21 (1) death; (2) severe sequelae (at least one disability that would make the child unable to function independently); (3) moderate sequelae (disability that affects function, but would not render the child dependent, or epilepsy not controlled with drugs); (4) minor sequelae (abnormality detected by neurological examination only, personality change as reported by parents, or epilepsy controlled with drugs); and (5) full recovery. 9 months after the study began, because we became increasingly aware that flaccid paralysis makes an important contribution to the morbidity of Japanese encephalitis,6 we decided to distinguish between the patients with moderate sequelae who could walk, and those who could not. For analysis, children who died or who had severe sequelae were defined as having poor outcome, whereas those who made a full recovery, or had moderate or mild sequelae were the better outcome group.5 Results from a preliminary prospective pilot study of 50 patients with Japanese encephalitis seen over 12 months showed that 44% had poor outcome.21 Our study was designed to show a two-thirds reduction in this proportion to 15%, with 95% confidence and 80% power, and allowance for a dropout rate of 10%. Time to death, to recover from coma, to sit independently, to stand independently, to walk at least 5 m independently, and to leave hospital were chosen as secondary outcome measures. Furthermore, we modified the Denver developmental assessment test22 to a simple score by awarding 1–4 points according to whether a child was normal or abnormal for
822
their age for each of the following categories: gross motor skills, fine motor skills and vision, hearing and speech, and social skills. An interim analysis was done by the study monitor after 40 patients. We randomly allocated patients to a treatment using blocks of 20, and the code for each patient was kept in a sealed envelope in a locked cupboard on a separate ward. Once a patient was enrolled, technicians who were not otherwise connected with the study, opened the envelope, drew up the study drug, and delivered it to the study ward, where nurses gave the treatment. Children received either interferon alfa-2a (Roche Pharmaceuticals, Basel, Switzerland) at a dose of 10 million units/m2 of body surface area, or the same volume of sterile water as an intramuscular injection daily for 7 days. We used a normogram to calculate body surface area from patient height and weight. The two substances used in the study were identical in colour, viscosity, and volume. A second sealed envelope was kept at the back of each patient’s study notes incase a physician urgently needed to know which drug a patient had received. We took a detailed history and a member of the study team did daily clinical and neurological examinations (or more frequently if needed) until death, stabilisation, or discharge. Information was recorded on standardised forms. On admission, we did a lumbar puncture with patients in the left lateral position, and if the patient was calm the opening pressure was measured with a spinal-fluid manometer. This procedure was delayed in patients who were convulsing, or those with clinical signs of raised intracranial pressure.5 If a patient died before lumbar puncture could be done, cerebrospinal fluid was taken immediately after death for diagnosis. We did cell counts and differentials, measured protein, glucose, lactate, did gram stain, and bacterial culture on the CSF. Blood was tested for packed cell volume, malaria parasites, platelet and differential white cell counts, and we did glucose, lactate, and biochemical screens, and viral serological tests. To diagnose JEV we measured anti-JEV IgM and IgG antibodies in serum, and CSF with a double sandwich capture ELISA that distinguishes between JEV antibodies and dengue antibodies.23,24 To monitor side-effects of interferon treatment, we did a full blood count daily, and liver function test twice a week during the admission. We planned to stop the study drug if a patient became neutropenic, and give these patients broad spectrum antibiotics. Other adverse events were graded according to WHO recommendations.25 Electroencephalography was done on admission, day 2, day 7, and as indicated clinically. We took a second serum and cerobrospinal fluid sample on day 7 to increase the chance of detection of JEV IgM antibodies.23 There were no onsite facilities for acute CT or MRI, but some patients were referred to a local centre for such imaging once their condition had stabilised. Prolonged or repeated seizures were treated with diazepam (0·25 mg/kg), followed if necessary by intravenous phenobarbital (10 mg/kg), but we did not have the facilities to paralyse and electively ventilate patients in status epilepticus. Mannitol (0·5–1 g/kg) was given at the discretion of the admitting physician, and suspected septicaemia was treated with broad-spectrum antibiotics. At discharge and at 3-months’ follow up, we did a physical and neurological examination, and assessed development. Patients who did not return for follow-up were reminded by letter, and if they were unable to return to the hospital, we did the examination at their home. At follow-up, we took particular note of seizures, progress at school, and changes in personality.
THE LANCET • Vol 361 • March 8, 2003 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Statistical analysies The analysis was by intention to treat. Normally distributed data were compared with Student’s t test; those that were not normally distributed were compared with the MannWhitney U test. We tested differences between proportions using the 2 test with Yates’ correction or Fisher’s exact test. Time to death and recovery times were analysed by survival analysis with the Peto-Peto-Wilcoxon test (Epi Info version 6 and StatView 4.02). Role of the funding source The sponsors of this study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Results Between Oct 1, 1996, and Oct 15, 1999, 117 patients met the entry criteria. 61 (52%) were randomly allocated to receive interferon and 56 (48%) to receive placebo (figure 1). Five patients (two in the interferon group) were transferred to other hospitals when other diagnoses became apparent, and thus data from these patients could not be included in analyses. Three of these patients had tuberculous meningitis, one had a cerebral haematoma, and one had cerebral effusion revealed by CT. Thus, 59 patients in the interferon group, and 53 in the placebo group could be assessed. Our analyses included one patient who died, and two who had fully recovered consciousness before the study drug could be given, and one for whom the drug was not available. There was one breach of the randomisation protocol when the wrong envelope was opened for a patient. However, because treatment allocation remained concealed from the patient and those treating her, we did not exclude her results; she was given placebo and analysed in the placebo group. The code did not need to be broken for any patient, but the study drug was discontinued in one patient (in the placebo group) when we noted that his creatinine concentration at admission was 10 g/L. The daily median dose of interferon in the treatment group was 7·6 (range 4·8–14·2) million units, which is equivalent to 10·9 (10·4–11·2) million units/m2 daily. 117 patients randomised
61 assigned interferon alfa 2a
56 assigned placebo
2 transferred to another hospital 1 with haematoma 1 with tuberculous meningitis 59 analysed at discharge
3 transferred to another hospital 2 with tuberculous meningitis 1 with effusion
53 analysed at discharge 10 died 2 lost to follow-up
47 alive and followed-up at 3 months
11 died 3 lost to follow-up
39 alive and followed-up at 3 months
Figure 1: Trial profile
THE LANCET • Vol 361 • March 8, 2003 • www.thelancet.com
Interferon (n=59) Demographics and history Boys Age (years) From rural location Length of illness (days) ⭓7 days illness Headache Rigor Vomiting Neck stiffness Rigidity spasm Convulsion >one convulsion Signs on admission Temperature Pulse (beats per minute) Respiration (breaths per minute) Response to menace Coma* Able to localise pain Seizure Rigidity spasm Meningism Facial nerve palsy Hemiparesis Flaccid limbs Absent abdominal reflex Increased limb tone Clonus Opisthotonus Decerebrate/decorticate posturing Investigations Positive for Japanese encephalitis virus Admission WCC⫻109/L Admission AST (U/L) CSF opening pressure (cm CSF)† CSF WCC/mL
Placebo (n=53)
32 (54%) 6 (1–14) 53 (90%) 5 (2–11) 17 (29%) 35 (59%) 21 (36%) 33 (57%) 26 (45%) 17 (29%) 34 (59%) 17 (29%) 38·5º (37-40) 120 (80-200) 28 (20-60) 21 (35%) 46 (78%) 24 (41%) 8 (14%) 15 (25%) 32 (54%) 7 (12%) 8 (14%) 13 (22%) 45 (76%) 26 (44%) 12 (20%) 3 (5%) 14 (24%)
33 (62%) 8 (1.9–14) 45 (85%) 5 (1–13) 11 (21%) 37 (70%) 18 (34%) 35 (67%) 26 (49%) 14 (27%) 25 (47%) 17 (32%) 38.0º (37-41) 120 (60-160) 28 (20-65) 23 (43%) 34 (64%) 27 (51%) 3 (6%) 12 (23%) 36 (68%) 2 (4%) 4 (8%) 10 (19%) 38 (72%) 25 (47%) 4 (8%) 4 (8%) 12 (23%)
45 (76%)
42 (79%)
12·5 (5.0–37.5) 420 (160–2000) 17 (5–33·5) 90 (0–540)
13.8 (1.6–33.8) 460 (150–1790) 17 (6–33·0) 40 (0–492)
*Coma defined as Glasgow coma score
