Epidemiology of Meningococcal Disease, New York City, 1989-2000

Alexandre Sampaio Moura, Ariel Pablos-Méndez, Marcelle Layton, Don Weiss


Emerging Infectious Diseases. 2003;9(3) 

In This Article


Our study has shown that a significant decrease in meningococcal disease incidence rates occurred at the beginning of the 1990s in New York City, and low incidence rates were observed throughout the rest of the decade. The decline in incidence rates are unlikely to have occurred because of changes in surveillance; no modifications in the diagnostic criteria for meningococcal disease were made, and only passive surveillance was conducted throughout the entire study period.

Compared to national surveillance data, the overall incidence rate in New York City during the period was 34% lower (0.67 vs. 1.02/100,000).[6] Age-specific rates in children <5 years of age have declined both nationally and in New York City, although the magnitude of the decline in the city has been greater. Nationally, the rate for patients <1 year of age declined from 13.5/100,000 in 1989 to 6.79/100,000 in 2000. For children 1-4 years of age, the rate declined from 4.18/100,000 to 2.04/100,000 over the same period.[6] Much of the decline occurred during the years 1998-2000. For New York City, the rates in the <1-year age group declined from 18.6/100,000 in 1989 to 2.72/100,000 in 2000. For the 1-4 year age group, the rate declined from 4.95/100,000 in 1989 to 0.93/100,000 in 2000. A similar trend for New York State (excluding New York City) has occurred, with overall meningococcal rates dropping 44% (from 1.28/100,000 in 1989 to 0.72/100,000 in 2000) and the rate <5 years declining 85% (from 8.85/100,000 to 1.29/100,000, unpub. data, New York State Department of Health, Division of Epidemiology).

The median age of New York City case-patients was higher than that observed in epidemiologic reviews for the United States and the New England region.[2,6] The increase in the median age of cases from 1989 to 2000 observed in New York City is predominately due to a decrease in the incidence rates among young children (1-4 years of age) and infants (<1 year of age), along with slight increases in rates among adults (25-64 years of age). This finding of higher median age may be a result of the greater decline in meningococcal disease seen in children <5 years of age in New York City compared to the rest of the United States.

In accordance with trends observed in other areas of the United States,[2,3] a significant increase in the incidence of serogroup Y infection occurred in New York City. The median age of patients with serogroup Y infections in New York City (30 years) was comparable to that of Connecticut patients (29 years) but greater than that seen in Illinois patients (16 years) when comparison data from 1989-1996 were used.[3] Serogroup C was responsible for an increasing number of sporadic cases and outbreaks in the United States[7] and Canada[8] in the late 1980s and early 1990s. In New York City, the single cluster involving three people in 1997 was caused by serogroup C infection. The incidence of serogroup C infection did not change substantially in New York City throughout the 12 years of the study; serotype C infection accounted for 18.2% of the total number of cases (range per year group 15.6-19.4).

Serogroup W-135, important worldwide because of the cases associated with returning pilgrims from Saudi Arabia, accounted for 4.6% of all cases in New York City from 1989 to 2000. This association with the pilgrimage to Mecca accounted for the three cases in New York City caused by serogroup W-135 reported from January to April in 2000: One patient was a returning pilgrim, another was a household contact of a returning pilgrim, and the third patient reported having interacted with returning pilgrims or their families.[9]

The incidence rate of serogroup B declined threefold during the period of study, and the infection has nearly disappeared in children <5 years of age in New York City (one case in 1999-2001). The number of serogroup B meningococcal cases in New York State (excluding New York City) has also declined, from 10 in 1997 to 3 in 2000 (unpub. data, New York State Department of Health, Division of Epidemiology). Data from New Jersey indicate that the number of meningococcal infections in children <5 years of age has similarly declined, from 16 in 1995 to 3 in 2000, although the number of serogroup B cases has remained relatively constant (unpub. data, New Jersey Department of Health and Human Services). In Oregon, where an increase in serogroup B meningococcal disease occurred in the last decade,[10] no similar decline in serogroup B has been seen in children <5 years of age (Frederick Hoesly, pers. comm.). An interesting discovery is the coincident increase in the use of Haemophilus influenzae conjugate vaccine containing serogroup B meningococcal outer membrane protein. Private provider vaccine orders received by the Vaccine for Children program for NYC indicate that the proportion of H. influenzae vaccine containing serogroup B meningococcal outer membrane protein has risen steadily from 0% in 1994 to 52% in 2000 (unpub. data, Department of Health, Immunization Program). Comparison Vaccine for Children data for the public sector are incomplete. Studies conducted by the manufacturer found that immunity to serogroup B meningococcus was induced by the serogroup B meningococcal outer membrane protein vaccine in a primate animal model[11] and in children during phase III vaccine trials (Alan Shaw, pers. comm.). Further epidemiologic and immunologic research are needed to explore the protective immunity and potential use of this vaccine for meningococcal serogroup B disease.

A significant change occurred in the prevalence of vaccine-preventable strains during the period. During the years 1989-1991, only 29.7% (43/145) of the meningococcal infections that occurred in patients >2 years of age were caused by vaccine-preventable strains. The proportion of vaccine-preventable strains increased steadily in each 3-year interval reaching 66.4% (85/128) in 1998-2000 (chi square for trend; p<0.01). Currently, the quadrivalent meningococcal polysaccharide vaccine, the only licensed and approved vaccine in the United States, provides good efficacy against serogroups A, C, W-135, and Y infections in older children and adults. The vaccine is not routinely recommended for the general population because of its short duration of protection, poor efficacy in children <2 years of age, and the low incidence of meningococcal infections in the United States.[1,12] To overcome the problems of immunity in young children, conjugate vaccines have been recently developed and might dramatically improve the prevention of meningococcal disease because of their greater efficacy among infants and longer duration of immunity.[13] The conjugate vaccine that is currently licensed in the United Kingdom only protects against serogroup C infection; its addition to the routine childhood vaccine schedule in New York City would have limited impact based on current serogroup incidence.[13] Meningococcal serogroup C accounted for 12% (3/25) of infections in children <5 years of age and 2% (3/144) of all meningococcal infections in 1998-2000. During the entire 12-year period, only one serogroup A meningococcal infection occurred in a child <5 years of age. To make an impact on rates of meningococcal disease in New York City through routine childhood vaccination, a vaccine is needed that produces good, long-lasting immunity in young children to serogroups B, C, W-135, and Y.

In contrast with the overall lower incidence rates, the CFR for 1989-1998 was 16.6% for New York City, compared to 9.3% for the rest of the United States.[6] Possible explanations for this finding include differential reporting of severe cases, presence of virulent clones in the population, and timely access of medical care. Additionally, not all public health jurisdictions include probable cases in their surveillance reports to the Centers for Disease Control and Prevention, raising the possibility that the national number of deaths is low because of underreporting of culture-negative fatal cases. A small proportion of cases (5%) in the New York City surveillance database met the definition for probable cases, suggesting that such cases may be underreported; however, no statistically significant difference existed in deaths by case status (confirmed CFR = 16.7%; probable CFR = 21.1%; chi square =0.46; p=0.50). This proportion of probable cases in New York City is comparable to that found in a review of meningococcal disease in New England for 1993-1998, where 4% of the cases met the probable case definition and the CFR was 10%.[14] The proportion of probable cases and CFR for meningococcal disease in New Jersey in 1990-2000 were 10.5% and 11.6%, respectively (unpub. data, New Jersey Department of Health and Senior Services, Infectious and Zoonotic Diseases Program). An assessment of meningococcal surveillance in New York State (excluding New York City) found delays but relatively complete reporting.[15] The exceptionally high death rate in 1999 prompted a closer examination of these data. Median age for the year was 35, higher than median age for any other year cohort and significantly higher than median age for all other years (median age excluding 1999 = 21; p=0.013). The proportion of group Y disease was 41%, which also differed significantly from the years excluding 1999 (chi square = 19.0, p<0.01). Further epidemiologic investigation, including molecular typing, is necessary to explain the excess meningococcal deaths in New York City.

A limitation of surveillance-based studies is the bias introduced by underreporting. However, because of the severity of the disease and the need for intravenous antibiotic treatment, most meningococcal disease case-patients are hospitalized, and the local health department is usually notified in order to track down close contacts and ensure that they receive antibiotic prophylaxis. A study assessing the completeness of the New York State surveillance system or meningococcal disease by using hospital discharge data as the basis for comparison showed that 93% of estimated cases in hospitalized patients in 1991 were properly reported.[15] Hospital practices, such as antibiotic administration before acquisition of cultures, might render samples from case-patients culture-negative; however, no evidence suggests that a change in such practice has occurred in the study interval.

Because our inclusion criteria required a positive bacterial culture, positive cerebrospinal fluid antigen test, or purpura fulminansis, cases that were culture-negative where antigen testing was not available or nonmeningitis cases without purpura fulminans might have been missed. The use of these inclusion criteria was important to ensure the validity of the study and comparability to other jurisdictions but could have slightly inflated the CFR if fatal cases tend to be reported more often to the health departments.

Another limitation was the large proportion of missing information for outcome (22.1%) that may have underestimated the CFR. We minimized this problem by performing death certificate searches using multiple search criteria.

Approximately three quarters of the New York City isolates during the study period were identified by serogroup; this proportion was similar to that observed in other surveillance-based studies conducted in the United States.[2,3,9] Assuming that the lack of serogroup information for a proportion of the cases was not related to problems in identifying any specific serogroup (i.e., independent from serogroup), bias was unlikely to have been responsible for the observed trends in serogroup.

The time-trend analysis performed in this study assessed the presence of epidemiologic trends during the 1990s but not the factors responsible for them. Therefore, our study was important in identifying trends, but further studies need to be conducted to test specific hypotheses about the factors responsible for them.

National surveillance data used for comparison of rates and CFR were based on the same case definitions as used in our study; however, not all jurisdictions follow these definitions. For example, New York State Department of Health excludes probable cases, and this variation in surveillance methodology may affect the national CFR used for comparison.