?, Piaras O’Lorcain, Joanne Moran, Patricia Garvey, Paul McKeown, Jeff Connell, and Suzanne CotterViral encephalitis (VE) and viral meningitis (VM) have been notifiable infectious diseases under surveillance in the Republic of Ireland since 1981. Laboratories have reported confirmed cases by detection of viral nucleic acid in cerebrospinal fluid since 2004. To determine the prevalence of these diseases in Ireland during 2005–2008, we analyzed 3 data sources: Hospital In-patient Enquiry data (from patients discharged from the hospital) accessed through Health Intelligence Ireland, laboratory confirmations from the National Virus Reference Laboratory, and events from the Computerised Infectious Disease Reporting surveillance system. We found that the national surveillance system underestimates the incidence of these diseases in Ireland with a 10-fold higher VE rate and 3-fold higher VM hospitalization rate than the reporting rate. Herpesviruses were responsible for most specified VE and enteroviruses for most specified VM from all 3 sources. Recommendations from this study have been implemented to improve the surveillance of these diseases in Ireland.
Encephalitis and meningitis are serious inflammatory diseases of the brain that require hospitalization for many patients and are a substantial cause of illness. Although the etiologic agent is not identified formost cases (1), viral infection has been reported as a major cause (2).
Acute encephalitis is characterized by a triad of fever, headache, and altered mental status (3). Common features include disorientation/depressed level of consciousness; disturbances of behavior, speech, or executive function; and diffuse or focal neurologic signs such as cranial nerve dysfunction, hemiparesis, or seizures (3). Capillary and endothelial inflammation of cortical vessels is a striking pathologic finding, occurring primarily in the gray matter or the gray–white junction (4). These features distinguish encephalitis from the more commonly encountered meningitis. The most common agents that cause acute viral encephalitis (VE) are herpes simplex virus (HSV) and varicella-zoster virus (VZV) (5).
A distinction must be made between acute VE and autoimmune/postinfectious encephalitis, which can occur with a variable latent phase between acute illness and the onset of neurologic symptoms (6,7). This distinction is critical because the management and prognosis are often quite different (4). Evaluation of cerebrospinal fluid (CSF) following lumbar puncture is essential for accurately diagnosing disease, unless its collection is contraindicated because of high intracranial pressure (4). In this study, we did not attempt to ascertain the prevalence of autoimmune/postinfectious encephalitis in Ireland.
Aseptic meningitis refers to a disease with acute onset of symptoms and obvious signs of meningeal involvement, in which an etiologic agent is not apparent after bacterial culture of CSF (8). The disease is often associated with lymphocytic pleocytosis without other cause. These patients usually lack altered sensorium or abnormal global or focal neurologic signs (3,4). Viruses are most commonly associated with these clinical manifestations, most frequently, enteroviruses, HSV-2, and arboviruses (9). Under the 2003 case definitions covering this study period, laboratory evidence involving CSF analysis or immune response, in addition to clinical diagnosis, was necessary for the reporting of VE and VM to public health departments (10,11).
In this study, we examined 3 different available data sources to estimate how well data reported to public health authorities and captured by the Computerised Infectious Disease Reporting (CIDR) passive surveillance system, during 2005–2008, reflected the incidence of VE and VM in Ireland. CIDR is the real-time Internet-based surveillance system for 93.9% of all notifiable infectious diseases reportable by statute in Ireland. We compared cases reported by CIDR with laboratory detection of cases from the National Virus Reference Laboratory (NVRL) and cases identified from hospitalized patient discharge information in the Hospital In-Patient Enquiry (HIPE) scheme. In 2005, eight (24%) of the 34 public hospital laboratories, in addition to the national reference laboratories, were connected directly to CIDR. This connection increased to 16 (47%) in 2008. Hospital/community physicians who manage VE or VM patients using laboratories which were not directly connected to CIDR, were obligated to report notifiable diseases to CIDR through the local department of public health.
The data sources and the process of reporting notifiable diseases in Ireland are shown in Figure 1. HIPE is an active system that monitors hospital activity and is independent of either NVRL or CIDR. The diagnoses and procedures recorded on the patient’s chart are coded by hospital administration staff according to the International Classification of Diseases and Related Health Problems, 10th Revision (ICD-10). A copy of this database, which includes only publicly funded hospitals, was accessed by using Health Intelligence Ireland (HII), an open source software program developed by the Irish Health Services Executive (www.healthatlasireland.ie/f1live).
A previous study suggested an underestimation of the incidence of acute encephalitis and VM in the northeastern part of Ireland (12,13). Only hospital activity and CIDR data were compared, and no distinction was made between laboratory or clinical criteria for diagnosis of these diseases at that time.
During 2005–2008, only 9.6% of cases in hospitalized patients with VE and 29.2% of cases in hospitalized patients with VM were reported to CIDR. Attempting to ascertain the proportion of the difference due to underreporting, we looked at the difference in the percentage of patients hospitalized with a diagnosis for VE and VM and events of these diseases created in CIDR in 2011. By then, 100% of public hospital laboratories were connected to CIDR, which facilitated the reporting mechanism. We found that hospitalized patients reported to CIDR with a diagnosis of VE increased from 9.6% to 20.2% in 2011 and that patients with VM increased from 29.2% to 69.6% (T. Kelly, unpub. data). The laboratories conducting the testing probably contributed to the underreporting. Underreporting or misclassification of diagnoses by hospital physicians without confirmatory tests may also have occurred.
Surveillance data are routinely used to quantify the incidence of disease and to identify outbreaks or emerging viral diseases, and this underidentification of VE and VM in Ireland is of concern. With the emergence of West Nile fever in Italy (16), dengue fever in France (17) and other European countries, and changes in the mosquito populations of different countries, CIDR must be able to detect possible emerging pathogens.
VE was shown to have greater effects on the Irish health service than VM in terms of deaths, length of hospital stay, repeat hospitalizations, ongoing hospital care, financial costs, and residual damage, although VM caused a greater number of hospital admissions. Similar to the situation in the United States, those seeking treatment for VE and VM were more likely to be male (9,18). However, a UK study found no difference in VE by sex and showed a lower hospitalization rate for patients with VE (19).
VE hospitalizations and events were most prominent in the youngest and oldest age groups. VM followed the same pattern for all 3 data sources (without statistical difference), with peaks in the 0–4 and 15–19 year age groups, which had also been found in Denmark (20). The peak in adolescents may reflect a large mumps outbreak that occurred during this period which primarily affected adolescent boys (21). The mumps outbreak in this age group may be attributable to a combination of factors, including social and environmental exposures and waning immunity to mumps (22). A previous seroepidemiologic study had identified inadequate immunity to measles, mumps, and rubella in school-aged children in Ireland (23–25).
Similar to results of studies in other countries (15,19,26), herpes and varicella-zoster viruses were the most commonly specified causes of acute aseptic encephalitis. As for VE, most (84.6%) VM hospital activity had an unspecified cause, which is similar to the 92% reported in the United States by the Centers for Disease Control and Prevention (9), which also reported that the highest rates of specified VM were due to enterovirus infection. The large number of enteroviral meningitis events, which are not reflected in hospital activity, suggests that many enteroviral meningitis cases are classified as unspecified.
We did not identify seasonal trends in VE incidence but did identify a higher incidence of VM during the summer months as found in other studies (3,9). Across all 3 data sources, the highest rates of VM were found in July 2006 and July 2008, correlating to periods when community outbreaks of enteroviral meningitis were reported to CIDR. This seasonality for VM continued in subsequent years (27).
Because lumbar puncture has a key role to play in accurate diagnosis of VE and VM, we evaluated the number of patients with VE and VM who had undergone this procedure and found that a higher percentage of patients with suspected VM had undergone lumber puncture than did patients with suspected VE. We also found that a higher percentage of specimens tested positive for a virus causing meningitis. The percentage positive for VE was remarkably low. This finding may reflect either a higher contraindication rate to lumbar puncture in VE patients, a problem confirming the diagnosis of encephalitis, or an increase in the rate of intrahospital transfer of severely ill VE patients to other hospitals with expertise in VE case management (in which case, repeat lumbar punctures would not be usual). It is also possible that alternative noninvasive diagnostic tools were used in patients for whom lumbar punctures were contraindicated, such as magnetic resonance imaging or electroencephalographic testing (5).
As a result of this study, new case definitions (28) have been implemented that incorporate a case classification for “possible” VE (enabling the reporting of patients meeting clinical criteria without laboratory confirmation). Only 52% of NVRL confirmations were reported to CIDR, but the NVRL Laboratory Information Management System extract used to report positive test results to CIDR was updated to capture omitted results. HIPE coding errors had also been identified in this study, and improved training and data entry validation measures were put in place. Clinical guidelines are being prepared by the Health Protection Surveillance Centre, which aim to improve both the investigation of VE and VM and reporting by physicians. Additional modifications to the list of notifiable diseases (29) include new case definitions for dengue fever, West Nile fever, and chikungunya disease, specifically, as well as new case definitions for hospitalized patients with chickenpox caused by VZV. The enhanced surveillance for influenza will now capture encephalitis.
Potential limitations of this study include the presentation of hospitalizations as numbers of patients hospitalized rather than numbers of discharges. Patients obtain a new medical records number when admitted to a different hospital, and patients who are transferred could be counted twice, but these differences in presentation are considered more likely to affect numbers of VE cases than VM, because of disease severity. We do not have a unique patient identifier in Ireland or a single health information system, as do some Scandinavian countries. HIPE data are collected with the sole purpose of being a source of hospital activity information and, under the Data Protection Act, cannot be used for any other purpose; as such, hospital information cannot be linked with surveillance data (30).
This study underscores a key disconnect between health care providers, diagnostic laboratories, and disease reporting. Further investigation of the perspective of attending physicians, laboratory directors, and public health officials may identify approaches for establishing more effective communication between each group on the essential issues of expeditious linking of pathogen identification with clinically apparent disease and reporting to public health. We recommend a follow-up study comparing rates with those when the new case definition (implemented March 2012), improved laboratory reporting, and clinical guidelines have been put in place, to evaluate whether there has been an effect by these changes. Analysis of the referral source of clinical diagnoses and laboratory results to CIDR, specimen type, and laboratory testing performed would facilitate better understanding of the proportion of difference between HIPE and CIDR because of underreporting. Such analysis would also provide feedback and education to the partners involved in health protection as an aid in highlighting the value of the surveillance of these diseases in Ireland and the detection of possible threats to public health.
No comments:
Post a Comment