The Changing Epidemiology of Invasive Haemophilus influenzae Disease, Especially in Persons ⩾65 Years Old
- 1Division of Infectious Diseases, Illinois Department of Public Health, Chicago, Illinois
- 2University of Michigan Medical School, Ann Arbor, Michigan
- 3Fargo Department of Veterans Affairs Medical Center, Fargo, North Dakota
- Correspondence: Dr. Mark S. Dworkin, University of Illinois at Chicago, School of Public Health, Div. of Epidemiology and Biostatistics, 1603 W. Taylor St. (MC923), Chicago, IL 60612 (mdworkin{at}uic.edu).
Abstract
Background. Few studies have reported the epidemiological characteristics of Haemophilus influenzae disease among adults.
Methods. Public health surveillance and hospital discharge data from Illinois were examined to determine the descriptive epidemiological characteristics and trends of invasive H. influenzae disease, and mortality data from Illinois were compared with data from several other states.
Results. During January 1996–December 2004, 770 cases of invasive H. influenzae disease were reported to the Illinois Department of Public Health (Springfield). The incidence of disease increased from 0.4 to 1.0 cases per 100,000 persons, including an increase of incidence in adults aged ⩾65 years from 1.1 to 3.9 cases per 100,000 persons. Nontypeable H. influenzae disease accounted for the greatest proportion of cases (35.8%–61.5%) in all but 1 age group. The number of cases of invasive nontypeable H. influenzae disease increased by 657%, from a low of 7 cases in 1996 to a high of 53 cases in 2004; as a proportion of annual cases, nontypeable H. influenzae disease increased from 17.5% in 1996 to 70.7% in 2004. Overall, the case-fatality rate was 12.7%, with the highest rate observed in persons aged ⩾65 years (20.6%). The case-fatality rate was similar for the hospital discharge database and for Indiana, Maryland, Oregon, and Wisconsin (range, 12.9%–18.2%).
Conclusions. In Illinois, the incidence of invasive H. influenzae disease increased from 1996 to 2004, and its epidemiological characteristics changed from a disease predominantly found in children and dominated by serotype b to a disease predominantly found in adults and dominated by nontypeable strains.
Before the introduction of the Haemophilus influenzae type b conjugate vaccine for young infants in 1988, H. influenzae type b was the most common cause of bacterial meningitis in children aged <5 years in the United States [1,2–3]. Consequently, the incidence of invasive H. influenzae disease in children aged <5 years has decreased by 98% [4, 5]. For adults, the incidence of invasive H. influenzae disease during the prevaccine era was estimated to be 1.7 cases per 100,000 persons, mortality was 28%, and H. influenzae type b was responsible for 50% of the cases [6]. During the past decade, few data have been reported about the epidemiological characteristics of H. influenzae disease in adults, although it was reported in 1993 that the incidence of invasive H. influenzae disease in persons aged ⩾12 years did not change substantially after the introduction of the vaccine for children [7]. More recently, the number of cases in children and adults have increased in Brazil [8] and England [9]. To examine the epidemiological changes that have occurred since the introduction of H. influenzae type b vaccine for children in the United States, we analyzed data that were reported to the invasive H. influenzae surveillance system in Illinois during 1996–2004 and compared the overall trend and mortality information with that from several other states.
Methods
Cases of H. influenzae disease were defined according to the Centers for Disease Control and Prevention (CDC) definition [10]: a clinically compatible case that is laboratory confirmed (isolation of H. influenzae from a normally sterile site—for example, blood, CSF, or, less commonly, joint, pleural, or pericardial fluid). Clinical syndromes of invasive H. influenzae disease included primary bacteremia, meningitis, otitis media, pneumonia, cellulitis, epiglottitis, peritonitis, pericarditis, septic arthritis, and conjunctivitis. The data were obtained from passive surveillance. In Illinois, the occurrence of any case of H. influenzae disease meeting the case definition is reportable within 24 h to the local health department and within an additional 24 h to the Illinois Department of Public Health (IDPH). Cases in persons who were not residents of Illinois were excluded from our analysis. All laboratories are required to report cases in which H. influenzae is isolated from a normally sterile site, and all hospital laboratories are required to forward isolates to the IDPH Division of Laboratories for serotype analysis, using a slide agglutination test.
Age groupings of the CDC's Active Bacterial Core Surveillance (ABCs) were used (<5, 5–17, 18–34, 35–49, 50–64, and ⩾65); however, for our analysis, some of the age groups were combined because of the small numbers in our data set. Additionally, we mainly compared children aged <5 years and adults aged ⩾65 years, because the most substantial differences were found in these 2 groups.
Because a large proportion of mortality data were missing in the passive surveillance data set, to better estimate mortality, we also queried a hospital discharge database (the IDPH Facility Discharge Data) for the frequency of hospital discharges and deaths with International Classification of Diseases, 9th Edition, Clinical Modification codes for septicemia (038.41) and meningitis (320.0) secondary to H. influenzae disease. We assumed that each patient was hospitalized only once as a result of invasive disease. Because patients with HIV infection have been reported to be at an increased risk for bacteremia from pneumonia due to H. influenzae disease, we also queried the same database for patients with a hospital discharge or death International Classification of Diseases, 9th Edition, Clinical Modification code for HIV infection and AIDS (042) who concurrently had a diagnosis of invasive H. influenzae disease. Hospitals and local health departments that reported cases of invasive H. influenzae disease during 2004 were contacted to obtain information on survival when survival data were missing. For comparison, the state health departments of Indiana, Maryland, Wisconsin, and Oregon were queried to determine the frequency of missing survival data. Data from the CDC's ABCs project were also examined and compared with data from Illinois [11]. Illinois is not an ABCs site. Population estimates were based on United States 2000 census data [12]. EpiInfo, version 3.3 (CDC), was used for all statistical analyses.
Results
During January 1996–December 2004, 770 cases of invasive H. influenzae disease were reported to IDPH. During 1996–2004, the number of cases increased from 49 to 132 (increase of 169%; linear test for trend, χ2 = 62.0; P < .0001) (figure 1 and table 1). The incidence of H. influenzae disease in 2004 overall and for adults (age, ⩾18 years) was 1.0 cases per 100,000 persons (based on United States 2000 census estimates [12]). A substantial increase in incidence was observed among adults, especially persons aged ⩾65 years (figure 1). Among adults who were aged 18–64 years, the number of cases increased from 13 to 37 (185%), whereas among adults aged ⩾65 years, the number of cases increased from 16 to 58 (263%). Persons aged 18–64 years and ⩾65 years were more likely to have a case of invasive H. influenzae disease during 2004 than during 1996 (OR, 2.9 [95% CI, 1.5–5.6] and 3.6 [95% CI, 2.0–6.6], respectively). The incidence of H. influenzae disease among children aged <5 years increased during 2000–2004 (figure 1). Persons at the extremes of age (⩾65 years and <5 years) accounted for 310 cases (44.0%) and 148 cases (21.0%), respectively. The mean number of cases that was reported during the winter months (December–February) was 28% higher than the mean number that was reported during the summer months (June–August; 74.0 vs. 57.7 cases; OR, 1.3; 95% CI, 0.9–1.8).
Figure 1
Incidence of cases of invasive Haemophilus influenzae disease in Illinois per 100,000 persons during 1996–2004 (total and by age group)
Figure 2
Percentage of each Haemophilus influenzae serotype, by year of onset, for 522 isolates with serotype information in Illinois during 1996–2004.
Table 1
Age, sex, and case-fatality rate among 770 patients with invasive Haemophilus influenzae disease.
The overall ratio of male to female subjects infected with H. influenzae was 0.9:1, but the ratio showed some trends within age groups. For persons aged ⩾65 years, the ratio of men to women was 0.7:1, and for persons aged <5 years, the ratio of male to female subjects was 1.6:1. There was also a peculiar pattern of H. influenzae disease and sex that was noted in persons aged 35–49 years, with a ratio of men to women that was 0.4:1. Racial characteristics among the patients infected with H. influenzae disease were as follows: white, 488 subjects (78.5%); black, 123 subjects (19.4%); Asian, 11 subjects (1.7%); and other races, 2 subjects (0.3%; OR for white vs. black, 0.8; 95% CI, 0.7–1.0).
Among 522 H. influenzae isolates submitted to the IDPH laboratory, serotype b accounted for 78 cases (14.9%), encapsulated non-b (a, c, d, e, or f) for 161 cases (30.8%), and nontypeable for 283 cases (54.2%). Most of the 161 encapsulated non-b isolates were type e (44 cases; 27.3%) and type f (96 cases; 59.6%). Nontypeable H. influenzae accounted for the greatest proportion of cases (35.8%–61.5%) among all age groups, except in persons aged ⩾65 years, among whom type f accounted for the greatest proportion of cases (36.5%; table 2). The number of cases of invasive H. influenzae type b disease decreased from 15 cases per year during 1996 to ⩽6 cases per year during 2001–2004 (figure 2). The number of cases of invasive encapsulated H. influenzae non-b disease ranged from 6 cases during 1997 to 29 cases during 2002. The number of cases of invasive nontypeable H. influenzae disease increased from 7 cases during 1996 to 53 cases during 2004 (increase of 657%; incidence, 0.06 cases per 100,000 persons in 1996 vs. 0.43 cases per 100,000 persons in 2004). As a proportion of annual cases, nontypeable H. influenzae increased from 17.5% in 1996 to 70.7% in 2004.
Table 2
Serotype distribution for 475 isolates of Haemophilus influenzae by age group during 1996–2004.
Meningitis, bacteremia, and any other invasive disease due to H. influenzae accounted for 105 cases (13.6%), 620 cases (80.5%), and 45 cases (5.8%), respectively. Among the 503 case reports including information on survival, the case-fatality rate was 12.7%. Meningitis accounted for only 1 death during 1996–2004. However, the case-fatality rates for persons with bacteremia and other disease manifestations were 14.9% (59 deaths) and 11.4% (4 deaths), respectively. Persons with bacteremia had a greater risk of death than did persons without bacteremia (OR for bacteremia vs. no bacteremia, 3.1; 95% CI, 1.2–9.9). Survival information was available for 65.1% of the patients with serotype information. The case-fatality rate was 7.3% (4 of 55 persons) among persons with H. influenzae type b disease, 12.4% (12 of 97) among those infected with encapsulated H. influenzae non-b (including 7 [31.8%] of 22 patients infected with H. influenzae serotype e and 4 [6.3%] of 63 patients infected with H. influenzae serotype f), and 13.9% (21 of 151) among those infected with nontypeable H. influenzae. The case-fatality rates for persons aged ⩾65 years who were infected with H. influenzae serotypes e and f were 38.5% and 11.1%, respectively. Five (71.4%) of the 7 deaths due to H. influenzae serotype e disease and 3 (75%) of the 4 deaths due to H. influenzae serotype f disease occurred among persons aged ⩾65 years. Overall, the case-fatality rate increased in association with an increase in age (table 1). The case-fatality rate for persons aged ⩾60 years (19.4%) was similar to that for persons aged ⩾65 years. Mortality was highest during October (8 [23.5%] of 34 patients) and November (10 [22.7%] of 44 patients). The case-fatality rate ranged from 3.9% in 1998 to 21.8% in 2001; however, there was no overall trend during 1996–2004. The case-fatality rate did not differ between cases reported from the Chicago metropolitan area (including Cook, Lake, Kane, Will, and DuPage counties [504 cases]) and greater Illinois (266 cases; 43 [13.1%] of 329 cases from the Chicago metropolitan area vs. 21 [12%] of 175 cases from greater Illinois).
Missing data exceeded 10% for race (16.5%), serotype (32.2%), and survival (34.7%) but did not exceed 10% for age (8.7%), month of onset (2.5%), or sex (3.1%). Information was complete for county of residence and disease manifestation. Information on survival was missing for 54 (40.9%) of 132 of patients experiencing onset of H. influenzae disease during 2004. After survival information was obtained for 33 additional patients infected with H. influenzae (32 survived, and 1 died), the case-fatality rate for 2004 was calculated to be 7.8% (8 of 103 patients).
The IDPH Facility Discharge Data database included a total of 744 hospitalizations for either septicemia (647 cases) or meningitis (97 cases) secondary to H. influenzae as either the primary or secondary diagnosis. The overall case-fatality rate was 12.9% (94 [14.5%] of 647 patients with septicemia and 2 [2.1%] of 97 patients with meningitis). Among 647 cases of septicemia due to H. influenzae, 6 patients (0.9%) were coinfected with HIV. No more than 1 case of septicemia due to H. influenzae among patients coinfected with HIV was identified per year. There were no cases of meningitis due to H. influenzae among persons coinfected with HIV.
The surveillance data from Illinois (population, 12.4 million [12]) included 770 cases, 65 deaths, and 233 case reports without information on survival. During the same period, data from Maryland (an ABCs site, with a population of 5.3 million [12]) included 734 cases, 97 deaths, and 21 case reports without information on survival (unpublished data), which resulted in an overall case-fatality rate of 13.6% and a missing data rate of 2.9%. In Oregon (an ABCs site with active surveillance in Portland, passive surveillance for the remainder of the state, and a population 4.3 million [12]), there were 375 cases, 46 deaths, and 94 case reports without information on survival (unpublished data), which resulted in a case-fatality rate of 16.4% and a missing data rate of 25.1%. Data from Wisconsin (a state with enhanced passive surveillance by a volunteer laboratory surveillance network, data available from 1998 to 2004, and a population of 5.4 million [12]) included 288 cases, 30 deaths, and 75 case reports without information on survival (unpublished data), which resulted in a case-fatality rate of 14.1% and a missing data rate of 26.0%. Finally, in Indiana (a state with passive surveillance, data available from 1999 to 2004, and a population 6.1 million [12]), there were 289 cases, 48 deaths, and 25 case reports without information on survival (unpublished data), which resulted in a case-fatality rate of 18.2% and a missing data rate of 8.6%.
Discussion
The epidemiological characteristics of invasive H. influenzae disease have changed dramatically since the introduction of the H. influenzae type b conjugate vaccine for children aged <5 years. Before the vaccine era, H. influenzae type b accounted for >95% of cases of invasive bacteremic disease in children [13, 14], and children accounted for 85% of patients with invasive H. influenzae type b disease [15]. Before the vaccine era, nontypeable H. influenzae was not a major cause of invasive disease [14, 16], although it has been noted previously to be an important pathogen in adults [6]. Our study of recent data showed that H. influenzae type b accounted for a small proportion of cases, even among children aged <5 years, but nontypeable H. influenzae accounted for the greatest proportion of cases. Data obtained during 2000–2004 from a large tertiary health care center in Manitoba have also called attention to invasive disease from encapsulated H. influenzae non-b (especially type a) and nontypeable strains, although most of the cases were among children aged <2 years [17]. It was reported previously that invasive disease due to H. influenzae type b decreased in children by 99% [4, 5] after introduction of the H. influenzae type b vaccine. We have observed a continued decrease of invasive H. influenzae type b disease. There was no apparent trend in encapsulated non-b invasive H. influenzae disease as a group during the study period.
Of concern is the increasing trend in the number of persons with invasive H. influenzae disease per year, which suggests that non-b H. influenzae is an emerging pathogen, especially among older adults. Adults aged ⩾65 years currently account for the greatest proportion of cases of disease (44.1%), followed by children aged <5 years (21.1%). Recent data from active population-based surveillance are consistent with our findings. The CDC's ABCs project reported that persons aged ⩾65 years accounted for 45.8% of cases, and children <5 years of age accounted for 11.8% of cases during 1997–2004 [11].
Overall, the case-fatality rate was 12.7%, with a much lower rate among children aged <5 years (3.7%) and a much higher rate among persons aged ⩾65 years (20.6%). Recent data from the United Kingdom support our finding, with an overall case-fatality rate of 11% [9]. It has been demonstrated previously that the case-fatality rate increases as age increases and is highest among persons aged ⩾60 years [5]. The case-fatality rate was previously estimated to be 5% among children aged <5 years, despite the administration of antibiotic treatment [1, 18]. Other US studies have reported mortality rates of 28% and 10% among adults aged ⩾18 years [6, 19], 13% among adults aged 30–59 years [2], and 24% and 32% among adults aged ⩾60 years [5, 6]. One study in Finland reported a mortality rate of 26% among persons aged ⩾16 years [20]. Comparisons between case-fatality rates reported in other studies and those reported in our study are limited by the age groups that were chosen.
A limitation of our surveillance data was the relatively large proportion of cases that were missing information on survival. However, the overall case-fatality rate was within 6% of the rates that were reported by Maryland and Oregon (which have partial or complete active population-based surveillance, respectively), Wisconsin (which has enhanced passive surveillance), Indiana (which has passive surveillance), and the IDPH Discharge Database. Also, the retrospective investigation of records from 2004 demonstrated that such missing records are more likely to represent patients who survived. Only 1 death (3%) was identified among 33 cases that had previously lacked information on survival. Interpretation of mortality data from passive surveillance should consider a possible bias toward the reporting of patients who died.
The data from our analysis update knowledge of the temporal and sex distribution of infection due to this emerging pathogen. Regarding seasonality, although a bimodal peak in cases has been reported from data from >20 years ago (with one peak during the winter months of October through December and another peak in the summer months during April and June [2, 21]), our data demonstrated only 1 peak (during December through February), which is consistent with a study of US data from ∼10 years ago [5]. Also, it had been previously noted that the incidence of disease was higher for boys than for girls [2], but for adults aged ⩾18, the incidence of disease was higher for women [6]. Our data support this finding for children aged <5 years and for adults aged ⩾65 years.
Additionally, we considered that the increase in the number of cases of H. influenzae disease may have been a result of an increase in the number of patients living with HIV infection or AIDS because of increased survival with these conditions. However, the query of the IDPH Discharge Data identified few patients with HIV infection or AIDS and invasive H. influenzae disease. Therefore, the increase in the number of cases of invasive H. influenzae disease could not be explained from these data by an increase in persons living with HIV infection or AIDS.
Although our data were collected primarily by using passive surveillance, completeness of reporting of cases was high. Compared with 744 cases of bacteremia or meningitis due to H. influenzae reported to the Hospital Discharge Database during the study period, 770 cases were reported to IDPH. The difference in the number of cases may relate to the inability of the hospital database to reliably identify other manifestations of invasive H. influenzae. Therefore, it was not queried for such cases. Hospital discharge data may overestimate the number of incident cases, because patients may be transferred from one hospital to another or may be hospitalized >1 time. The true burden of invasive disease may be much higher than demonstrated in our data, because many patients receive antibiotic therapy before culture samples are obtained. Patients who benefit clinically from this empirical therapy may subsequently have negative blood culture results and, thus, not satisfy our case definition, which depends on isolation of H. influenzae from a normally sterile site [22]. Also, our surveillance system lacked information about underlying diseases that may influence the incidence and outcome of disease. Active surveillance systems are better suited for collecting such data.
The epidemiological characteristics of invasive H. influenzae disease have changed from a disease that predominantly affects children and is dominated by type b to a disease that predominantly affects adults and is dominated by nontypeable strains. The increasing trend in the number of cases, especially among older adults, deserves increased attention. Surveillance is information for action. In the case of invasive disease due to H. influenzae type b, such public health action includes respiratory isolation, investigation, prophylaxis, and follow-up of close contacts, such as household and day-care contacts. Surveillance data can also demonstrate geographic regions of continuing transmission and areas where vaccine coverage should be improved [5]. However, there is no vaccine for emerging invasive H. influenzae disease due to non-type b isolates. Because of the growing elderly and HIV-infected populations (both groups at an increased risk for this disease), it is worth considering whether invasive H. influenzae disease due to no-type b isolates should be a future focus for vaccine development.
Acknowledgments
We acknowledge assistance with the data sets from Roland Lucht and Rich Forshee and sharing of data by the Indiana State Department of Health, Maryland Department of Health and Mental Hygiene, Oregon Department of Human Services, and Wisconsin Department of Health and Family Services.
Potential conflicts of interest. All authors: no conflicts.
- Received August 4, 2006.
- Accepted October 19, 2006.
- © 2007 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
-
- Articles and Commentaries
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
