Invasive Meningococcal Disease in Adolescents and Young Adults FREE

Neisseria meningitidis is a leading cause of meningitis and other serious infections in the United States and throughout the world.^1- 2 In recent years, the overall annual incidence of meningococcal infection in the United States has ranged from 0.8 to 1.0 per 100 000 population, with infants having the highest risk.³ Adolescents and young adults have generally had a low risk of invasive meningococcal infection.^3- 4

During the 1990s, the number of meningococcal outbreaks increased in the community and on college campuses.^5- 9 Several recent studies identified subgroups of college students that were at increased risk of meningococcal infection, such as freshmen living on-campus.^9- 11 This finding led to recent national recommendations about the use of meningococcal vaccine among college students.^11- 12

However, the incidence of meningococcal infection has increased overall in persons aged 15 through 24 years, including persons of high school age.⁹ The features of meningococcal infection in 15 through 24 year olds during the period of this increase have not been well characterized. This is an important public health issue because of the devastating nature of meningococcal infection and because many of these cases are potentially vaccine preventable. The purpose of this study was to characterize the problem of meningococcal infection in adolescents and young adults in an entire state during the 1990s.

This study was approved by institutional review boards of the Johns Hopkins University and Maryland Department of Health and Mental Hygiene. Active surveillance for invasive meningococcal infection was initiated in Maryland on November 1, 1991, as part of the Maryland Bacterial Invasive Disease Surveillance (BIDS) project.^9,13 BIDS is the Active Bacterial Core Surveillance (ABCs) component of the Emerging Infections Program Network that is funded by the Centers for Disease Control and Prevention (further information is available at http://www.cdc.gov/ncidod/dbmd/abcs). The case definition is the isolation of N meningitidis from a normally sterile body fluid from a Maryland resident of any age. All acute care hospitals in Maryland participate, as do large hospitals in Washington, DC, where Maryland residents frequently seek medical care. Nonhospital microbiology laboratories that receive blood cultures are also included. For eligible cases, the hospital infection control professional completes a case report form, which includes demographic and brief clinical information, and the laboratory submits the bacterial isolate for species confirmation and further testing. Microbiology laboratory audits to identify unreported cases are performed by reviewing the laboratory records.

For this study, case report form data were supplemented with reviews of the medical and health department records, using a standardized chart abstraction form. For clinical variables of disease severity, the most abnormal observation during the first 3 days of the hospital stay was recorded.

For most of the analyses in this study, we used BIDS data for 1992 (the first full calendar year of the BIDS project) through 1999. For trends analysis, we included passive Maryland Department of Health and Mental Hygiene surveillance data for the years 1990 through 1991 because, based on previous studies, there was evidence that the increase in meningococcal infection in adolescents and young adults began around this time.^3,8- 9 Combining the passive and active surveillance data was justified because there were features of meningococcal infection surveillance in Maryland that made substantial underreporting unlikely. These included the legal requirement that both health care providers and microbiology laboratories report cases (ie, dual reporting), that providers were given an incentive to report because of the assistance provided by health authorities in administering chemoprophylaxis to prevent secondary cases and because of the frequent reporting of meningococcal cases in the local media. Seventy-five cases were reported during the 1990 through 1991 passive surveillance years vs 72 cases in the years 1992 through 1993, suggesting that passive reporting in Maryland was reasonably complete. In addition to analyzing trends in meningococcal incidence, we examined trends in the proportion of all meningococcal disease that occurred among persons aged 15 through 24 years because it is unlikely that this proportion would be influenced by incomplete reporting, if it occurred, in passive surveillance. The cases obtained through the passive surveillance system were excluded from all other analyses because the meningococcal isolates were not available for these cases.

Species identification and serogrouping were performed using standard methods.¹⁴

Population estimates based on the 1990 census for the years 1990 through 1996 were obtained from the Centers for Disease Control and Prevention Wonder database (available at http://wonder.cdc.gov) and for the years 1997 through 1999 were obtained from the US Census Bureau. A Washington, DC, suburb was defined as the Maryland counties of Prince Georges and Montgomery. A vaccine serogroup was defined as a serogroup included in the licensed serogroup A, C, W-135, and Y polysaccharide vaccine.¹⁵ Meningitis was defined as either a physician diagnosis of meningitis and/or a white cell count of higher than 50 × 10³/µL in the cerebrospinal fluid. Shock was defined as having either a systolic or a diastolic blood pressure of lower than 90 mm Hg or 60 mm Hg, respectively, for those aged 13 years or older, lower than 80 mm Hg or 50 mm Hg for ages 6 through 12 years, and lower than 70 mm Hg or 40 mm Hg from birth through age 5 years. Meningococcemia without meningitis, included because of evidence that meningococcemia alone has a worse prognosis than meningitis with or without meningococcemia,⁴ was defined as the presence of ecchymoses and/or petechiae, shock, and/or a physician diagnosis of purpura fulminans, and the absence of meningitis.

Incidence rates were based on Maryland census data for each year of the study. The 2-sided Fisher exact test was used for the analysis of dichotomous variables. Trends in meningococcal incidence and the proportion of all cases among 15 through 24 year olds were analyzed by the χ² test for trend. SAS version 6.12 (SAS Institute, Cary, NC) was used for the logistic regression analyses to identify demographic and vaccine-related factors, which could be used to distinguish patients aged 15 through 24 years from those in other age groups. Models were constructed using stepwise model-building procedures with only demographic and vaccine-related variables that were associated with having a 15 through 24-year-old case in univariate analysis being eligible for entry. Entry and stay criteria for the models were set at a significance level of .10. Another pair of stepwise models was constructed to determine whether having a 15 through 24-year-old case was independently associated with death.

A total of 295 meningococcal cases of all ages were identified during the years 1992 through 1999. An additional 75 cases of all ages were identified through the passive surveillance during the years 1990 through 1991. Of the 295 cases identified through active surveillance, the N meningitidis isolates were available for serogrouping for 257 cases (87.1%) and the medical and/or health department records for 272 cases (92.2%). One hundred nine cases (36.9%) were younger than 15 years, 71 (24.1%) were aged 15 through 24 years, and 115 (39.0%) were 25 years or older. The median age among those younger than 15 years was 12 months, 18 years among cases aged 15 through 24 years, and 51 years among those 25 years or older. Among those aged 15 through 24 years, 21 were high school and 21 were college students. There were 2 small meningococcal clusters during the study period: 2 serogroup C cases on a college campus in 1997 and a party-associated cluster of 3 cases among adults (ages 18, 20, and 21 years) in 1999^9,16; both outbreaks were caused by identical serogroup C strains.

The average annual meningococcal incidence from 1992 through 1999 was 0.73 per 100 000 population and varied considerably by age (Figure 1). The annual incidences in those aged 15 through 19 years and in those aged 20 through 24 years were 1.8 and 1.0 per 100 000, respectively. The highest incidence in 15 through 24 year olds occurred during the years 1996 through 1997, when it was 2.1 (Figure 2). During this period, the incidence was 2.6 in 15 through 19 year olds and 1.5 in 20 through 24 year olds. When analyzed by each year of age, the highest rates were among 17 and 18 year olds, with annual incidences of 2.8 and 3.0, respectively.

From 1990 to 1997, the annual incidence in persons aged 15 through 24 years increased from 0.9 to 2.1 cases per 100 000 (P = .01, χ² for trend) (Figure 2). Likewise, the proportion of all disease that occurred in this age group increased from 16.0% to 28.9% (P = .03, χ² for trend). The incidence and proportion of cases subsequently decreased to 1.0 per 100 000 and to 16.4% in the years 1998 through 1999 essentially to what they were in the years 1990 through 1991. During the period 1990 through 1999 the overall annual incidence varied from a low of 0.6 in the years 1994 through 1995 to a high of 0.9 in the years 1996 through 1997. The annual incidence in children younger than age 15 years declined from 2.2 to 1.2 (P = .01, χ² for trend), mostly due to a decline among children younger than 5 years, and the annual incidence in adults aged 25 years or older increased from 0.3 to 0.5 (P = .03, χ² for trend).

Compared with those younger than 15 years, 15 through 24 year olds were more likely to live in a suburb of Washington, DC, to have ecchymoses, meningococcemia without meningitis, shock, coma, thrombocytopenia, an elevated serum creatinine level, a fatal outcome, and infection caused by a vaccine serogroup strain (Table 1). Compared with those 25 years or older, 15 through 24 year olds were more likely to be male; be a smoker; have meningitis, ecchymoses, shock, and serogroup C infection; and less likely to have a chronic medical condition. Of those aged 15 through 24 years, 22.5% died vs 4.6% of those younger than 15 years and 16.5% of those 25 years or older. Among those younger than 15 years, the case fatality ratio was 3.8% (3 of 79) for children younger than 5 years and 11.8% (2 of 17) for children aged 5 through 9 years.

In the logistic regression analysis, fatal infection was associated with being aged 15 through 24 years vs being younger than 15 years (odds ratio [OR], 5.2; 95% confidence interval [CI], 1.8-15.6; P = .003). Infected individuals from a Washington, DC, suburb had a higher odds of being aged 15 through 24 years old vs being younger than 15 years (OR, 2.3; 95% CI, 1.1-5.1; P = .03). Males (OR, 4.3; 95% CI, 2.1-8.8; P<.001) and those with serogroup C infection (OR, 3.9; 95% CI, 1.9-8.4; P<.001) were at higher odds of being aged 15 through 24 years than being aged 25 years or older. In the separate logistic regression analysis using death as the outcome, being aged 15 through 24 years was associated with a fatal outcome compared with being younger than age 15 years (OR, 6.1; P = .002) but not when compared with patients 25 years or older (OR, 1.2; P = .70).

There were 3 (2.9%), 5 (7.2%), and 56 (50.0%) patients with chronic conditions among the persons younger than 15 years, aged 15 through 24 years, and 25 years or older, respectively. Among the children younger than 15 years, 2 had human immunodeficiency virus infection and 1 was receiving short-term steroid therapy at the time of the onset of meningococcal infection. Among the 15 through 24 year olds, 1 abused alcohol and 1 each had diabetes mellitus, ulcerative colitis, and a cerebrospinal fluid leak. Among the patients 25 years or older, the most common chronic medical conditions were diabetes mellitus (17 patients), malignancy (11 patients), alcohol abuse and HIV/AIDS (7 patients each), congestive heart failure (6 patients), and organ transplantation (3 patients). Many patients 25 years or older had more than 1 chronic condition.

Among the fatal cases of those aged 15 through 24 years, there were 2 deaths each in 1992 and 1993, 1 in 1994 and 1996, 3 each in 1995 and 1997, and 4 in 1999. Three (18.8%) lived in Baltimore County, 5 (31.3%) lived in a Washington, DC, suburb, and 8 (50.0%) lived in other Maryland counties. Ten (62.5%) fatal cases were male, and 9 (56.3%) were white. Of the 14 fatal cases for which serogroup was known, 8 (57.1%) were serogroup C, 4 (28.6%) serogroup Y, and 2 (14.3%) serogroup B. There were no significant differences in case fatality by year.

In the 55 survivors aged 15 through 24 years, long-term sequelae included 1 survivor with bilateral below-the-knee amputation and 1 with hearing loss; in the 104 survivors younger than 15 years, they included 1 survivor with bilateral below-the-knee amputation, 1 with above-the-knee amputation, and 9 with hearing loss; and in the 96 survivors aged 25 years or older, they included 1 survivor with multiple toe amputations and 3 with hearing loss.

In those aged 15 through 24 years, 46.9% had serogroup C, 31.3% had serogroup Y, 12.5% had the serogroup B, and 4.7% had serogroup W-135 strains. In analyzing infections by group, 82.8% of those aged 15 through 24 years had infections that were potentially vaccine preventable compared with 68.1% of children younger than 15 years and compared with 76.8% of those aged 25 years or older (P = .04, comparing 15 through 24 year olds vs <15 years). Among 15 through 24 year olds, the proportion of infections caused by the serogroup C and Y strains tended to increase and decrease, respectively, during the 1990s. Neither of these trends was statistically significant (Figure 3). The increase in the proportion of infections caused by serogroup C continued even during the years 1998 through 1999, when the overall incidence rate fell. Serogroup Y accounted for half of cases during the years 1992 through 1993 but only 16.7% of infections in the years 1998 through 1999. Serogroups B and W-135 accounted for a relatively small proportion of disease in this age group.

We found that the incidence of meningococcal infection in Maryland increased substantially among 15 through 24 year olds during the 1990s. During the period of 1994 through 1997, 15 through 24 year olds accounted for nearly 30% of all meningococcal infections. The emergence of meningococcal infection in this age group was not unique to Maryland. Similar increases were found in Oregon and other states.^8,17 Interestingly, the incidence dropped precipitously and returned to baseline during the last 2 years of the study. This downward trend has recently been noted elsewhere, albeit to a lesser extent. In several other ABCs sites, the incidence of meningococcal infection in those aged 15 through 24 years was 0.72 per 100 000 in 1992, peaked at 2.04 in 1996, and declined to 1.63 in 1999 (Kathleen Shutt, MS, written communication, September 15, 2000).

We identified epidemiologic and clinical features of meningococcal infection in 15 through 24 year olds that distinguished this group from meningococcal infection occurring at other ages. The most striking finding was that nearly a quarter of meningococcal infections in 15 through 24 year olds was fatal. This was unexpected because the case fatality caused by meningococcal infection in this age group typically is low; the highest case fatality tends to be among neonates and elderly persons.^4,18 The high case fatality in this group was particularly surprising because the vast majority of these 15 through 24 year olds were previously healthy; whereas, half of the those aged 25 years or older had chronic medical conditions that would be expected on average to lead to a less favorable outcome. In a study of 44 cases of meningococcal infection among adults (including elderly persons) in Atlanta, Ga, from 1988 through 1993, 3 (8%) of cases were fatal.¹⁹ In a study of cases that occurred in Spain from 1987 through 1992, the case fatality rates among patients aged 15 through 19 and 20 through 29 years were 1.4% and 3.8%, respectively.²⁰ Among children younger than 15 years in that study, the case fatality was between 2.7% and 6.1%, similar to fatality rates in Maryland, indicating that the association between death and being aged 15 through 24 years old was not due to a spuriously low case fatality among Maryland children younger than 15 years. The mechanism of the high case fatality is not known but the clinical data suggest that the increased mortality was because of a high frequency of the syndrome of meningococcemia.

The factors responsible for the increased incidence of meningococcal infection in 15 through 24 year olds and the subsequent decline are not known. The increase in Oregon was found to be caused largely by a serogroup B clone.^17,21 Although the increased incidence in Maryland was caused mainly by 2 different serogroups, most serogroup C disease from diverse regions of the United States has been found to belong to the electrophoretic type 37.⁸ Similarly, serogroup Y infection in the United States is fairly clonal, with the majority of strains belonging to electrophoretic types 501/508 or 516. If these or other clonal groups had been recently introduced into Maryland and were antigenically new to adolescents and young adults, this could provide a partial explanation for the observed increase. Ongoing studies will characterize the molecular epidemiology of meningococcal infection in Maryland and correlate the findings with the changing incidence of infection in 15 through 24 year olds.

Cigarette smoking is a known risk factor for meningococcal infection and other invasive bacterial diseases and was reported for nearly half of the 15 through 24-year-old cases in our study.^22- 26 Although what role smoking played, if any, in the increase in meningococcal incidence in 15 through 24 year olds is not known, national trends in youth smoking mirrored trends in meningococcal incidence during the 1990s.^27- 29 For example, the prevalence of cigarette smoking in the previous 30 days among 12th grade high school students was 28.3% in 1991, peaked at 36.5% in 1997, and decreased to 34.6% in 1999.²⁸ If smoking is in fact causally associated with meningococcal infection, then, by definition, changes in smoking prevalence would lead to changes in meningococcal incidence. That the decline in meningococcal incidence in Maryland was of larger magnitude than the decline in smoking prevalence, however, suggests that smoking is not responsible for all of the changes in meningococcal incidence in 15 through 24 year olds. The prevalence of binge drinking, also associated with meningococcal infection, also seems to have increased slightly during the 1990s and then decreased by 1999.^30- 31 We suspect that the changes in meningococcal incidence we observed were due to a combination of changes in population immunity to circulating strains and changes in behavioral risk factors.

What are the implications of this study for the prevention of meningococcal infection? Vaccine prevention in 15 through 24 year olds is attractive for several reasons. First, more than 80% of infections in our study were caused by serogroups that are included in the vaccine that is licensed in the United States.¹⁵ Second, vaccine efficacy in adolescents and young adults is high.^32- 34 Third, there is evidence that transmission of N meningitidis frequently occurs from young adults to children, suggesting that a vaccine that afforded herd immunity through a reduction in nasopharyngeal carriage could potentially lead to reduced transmission to susceptible children.^35- 36 Fourth, a relatively large proportion of all meningococcal infections in Maryland occurred in this age group. However, the relatively small number of cases, potential logistical difficulties in immunizing adolescents and young adults, questions about the cost-effectiveness of meningococcal vaccines,¹¹ and the decline in incidence in the years 1998 through 1999 and elsewhere must also be considered. In any case, these data and continued surveillance for meningococcal infection will be useful in deciding whether vaccination is appropriate for this age group.

There are several limitations of this study. The requirement for the isolation of N meningitidis in culture led to an underestimation of the burden of meningococcal infection in all age groups in our study.³⁷ Since we did not follow up cases prospectively after discharge from the hospital, we likely also underestimated the frequency of certain sequelae, such as hearing loss. Finally, the study design we used did not allow us to directly assess the contribution of smoking and other behavioral risk factors to the changing epidemiology of meningococcal infection in Maryland.

In summary, the incidence of meningococcal infection in Maryland residents aged 15 through 24 years increased during the 1990s but subsequently decreased. During some years, persons aged 15 through 24 years accounted for nearly 30% of all meningococcal infections in Maryland. Infection in this age group was associated with an unexpectedly high case fatality. The majority of infections were vaccine preventable. This study underscores how the epidemiology of meningococcal infection continuously changes and the need to monitor disease patterns with active, laboratory-based surveillance. Surveillance will provide crucial information for the development of optimal immunization strategies, once improved meningococcal vaccines become available.