Temporary Increase in Incidence of Invasive Infection Due to Streptococcus pneumoniae in the Netherlands

In 1996 and 1997, the Netherlands Reference Laboratory for Bacterial Meningitis (Amsterdam) noted an increase in Streptococcus pneumoniae isolates from blood but not from CSF [1]. To find an explanation for this increase, we determined the incidence of invasive pneumococcal disease detected in the period 1991–1998 by 6 regional public health laboratories covering 18% of the Dutch population (15.4 million inhabitants in 1994). Subsequently, we related this incidence to the inci-dence of influenza and the severity of the winters in the study period [2]. Invasive pneumococci were defined as those isolated from blood or CSF. Repeated isolation of invasive S. pneumoniae from the same patient within 1 month was ignored. Incidences were calculated with 1 August as the start of a “pneumococcal year” (i.e., the year 1997 was from 1 August 1996 to 31 July 1997). The Royal Meteorological Institute (KNMI, De Bilt, Netherlands) provided a winter severity index, calculated as the negative sum of the negative 24-h average temperatures (°C) in the period 1 November to 31 March. The incidence of influenza-like illnesses was derived from a countrywide network of general practitioners (NIVEL, Utrecht, Netherlands). Age-specific incidences were calculated by use of yearly updated demographic data and the postal codes of the patients in the areas served by the laboratories. Differences in incidence between years and contributing laboratories were tested simultaneously by means of PROC GENMOD of SAS Version 6.12 (SAS Institute, Cary, NC), with year and laboratory as classes, the number of blood cultures as a covariate, and the catchment population as the offset variable. During the study period, 2182 S. pneumoniae isolates from blood and 113 S. pneumoniae isolates from CSF, summarized as invasive isolates, were reported by the participating laboratories. A second invasive isolate of S. pneumoniae was recovered only from 9 patients. Therefore, the number of confirmed cases of invasive pneumococcal infection and the number of invasive pneumococcal isolates were equal. The incidence of pneumococcal bacteremia or meningitis in the period 1991–1994 was 8.3–9.2 cases per 100,000 person-years (table 1). In 1996, the incidence increased significantly to 14.8 cases per 100,000 person-years, and in 1997 to 13.3 (P = .0001); but in 1998 it returned to 8.9 cases per 100,000 person-years. The incidence for the age group ≥65 years was approximately 4-fold higher and showed the same trend. The average number of pneumococcal isolates per 1000 blood cultures was 8.7 in the 2 years with the highest incidence of invasive pneumococcal disease, in contrast to 6.1 in the other years (table 1). The gradual increase in the number of blood cultures in our study did not explain the sudden increase in invasive pneumococcal isolates in 1996 and 1997 [3]. Winter severity index, incidence of influenza-like illnesses, no. of cases of invasive pneumococcal disease (IPD) per 1000 blood cultures, and incidence of IPD in the Netherlands by year. The mean incidence of 23 cases per 100,000 person-years for the age category 0–1 year (table 1) was much lower than the incidence of 136 cases per 100,000 person-years reported by Pastor et al. [4]. An explanation may be that children with severe illness in the Netherlands are preferably admitted to specialized children's hospitals that are not served by the laboratories participating in our surveillance. No trends in the distribution of pneumococcal serotypes [1], changes in the occurrence of influenza viruses (table 1), or other respiratory viruses were noted. However, the higher incidence of invasive pneumococcal infections in 1996 and 1997 for the age group ≥65 years coincided with relatively severe winters, as shown by the winter severity index (table 1).