Increased incidence of pneumococcal bloodstream infections

Pneumococcal bloodstream infections in pediatric patients rose from 3.6 to 14.8 episodes per 100 000 from 1986 to the year 2000 (P < 0.001) in a population of ∼650 000 inhabitants (16% <16 years of age). Incidence in the group <24 months of age increased from 7.4 to 78.6 episodes per 100 000 per year (P < 0.001). Overall 75% of the serogroups are included in the new heptavalent vaccine. In the pediatric population, invasive pneumococcal infections are well-known as an important cause of morbidity and mortality, but the evolution of the incidence of this problem in recent years is not well-known. This is particularly important when considering the potentially widespread introduction of a new heptavalent antipneumococcal conjugated vaccine in children. 1, 2 We retrospectively assessed the incidence, clinical presentation, serogroups, antimicrobial resistance pattern and clinical evolution of pneumococcal bloodstream infections in a large pediatric population during 15 years. Patients and methods. Hospital General Universitario “Gregorio Marañón” is a large teaching institution in Madrid, Spain [1750 beds (160 pediatric)] serving a mainly urban population of ∼650 000 inhabitants of whom 16% are <16 years of age (pediatric population). We are the main center for pediatric care in our catchment and referral area of pediatric patients to our institution, and this did not change during the study period. Our study included all patients <16 years of age with one or more positive blood cultures in which Streptococcus pneumoniae was isolated during the 15-year period from January 1986 through December 2000. S. pneumoniae were identified by standard and well-recognized procedures. 3 Capsular serotyping of isolates was conducted at a National Reference Laboratory (Centro Nacional de Microbiología, Instituto de Salud Carlos III, Majadahonda, Madrid, Spain). Antimicrobial susceptibility tests were conducted with a microdilution technique (SENSITITRE; Trek Diagnostic System, Ltd., East Grinstead, UK) 4 and interpreted according to National Committee for Clinical Laboratory Standards recommendations. Susceptibilities were interpreted as follows: penicillin: susceptible strains (MIC ≤ 0.06 μg/ml), intermediate strains (MIC = 0.12 to 1 μg/ml) and resistant strains (MIC ≥ 2 μg/ml); cefotaxime: susceptible strains ≤ 0.5 μg/ml, intermediate 1 μg/ml and resistant ≥2 μg/ml; erythromycin: susceptible ≤0.25 μg/ml, intermediate 0.5 μg/ml and resistant ≥1 μg/ml. Evolution of incidence of bloodstream infections in pediatric patients was also calculated for Escherichia coli and Haemophilus influenzae to compare the trends in evolution with those of S. pneumoniae. Definitions of community-acquired infections, meningitis, primary bacteremia and occult bacteremia are those accepted in general text books of infectious diseases. 5 Statistics. The significance of the increasing number of cases of pneumococcal bacteremia was assessed with the chi square and Mantel-Haenszel tests. Calculations of the annual increase rate were conducted by linear regression analysis. Percentage rates of the antimicrobial resistance of isolates in different periods were compared using Pearson’s chi square, except when one of the groups was <5, when we applied Fisher’s exact test. In all circumstances statistical significance was considered as P ≤ 0.05. All P values were two-tailed. Results. The population served by our institution (from 1986 to 2000) remained stable overall (606 369 and 644 611 inhabitants) with a substantial increase in adults (439 902 to 550 246) and a decrease in the pediatric population (166 467 in 1986 to 94 366). Between January 1986 and December 2000 we saw 1028 episodes of pneumococcal bloodstream infections in patients of all ages. Of these 106 (10.3%) occurred in the pediatric population. Overall the incidence of episodes per 100 000 inhabitants and year was 10.9 (12.3 episodes per 100 000 adults per year and 5.4 per 100 000 in the pediatric population-year). Figure 1 shows how the incidence of pneumococcal bloodstream infections evolved throughout the study. The progressive increase in pneumococcal bloodstream infections in pediatric patients rose from 3.6 episodes per 100 000 in 1986 to 14.8 episodes per 100 000 in the year 2000 (P < 0.001). During the study period the mean increase in incidence per year was 0.73 episode per 100 000.Fig. 1: Trend of evolution of the incidence of pneumococcal bloodstream infection throughout the study (per 100 000 inhabitants per year).The maximum incidence in pediatrics was found in the group <24 months of age (26.5 episodes per 100 000 per year). Incidence in this subgroup also increased, rising from 7.4 episodes per 100 000 per year in 1986 to 78.6 episodes per 100 000 per year in 2000 (P < 0.001). The increase in pneumococcal bloodstream infections in pediatric patients can also be demonstrated when the data were calculated against the denominator of 1000 hospital admissions. We had an overall incidence of 1.1 episodes per 1000 admissions with a minimum of 0.41 in the year 1987 and a maximum of 2.52 in the year 2000. To rule out the hypothesis that the increase in incidence could be caused by a higher use of blood culturing in the pediatric population, we compared the evolution of bloodstream infections caused by S. pneumoniae with those caused by E. coli and H. influenzae. The increase in E. coli and the decrease in H. influenzae during the study period are not statistically significant (P = 0.425, P = 0.973, respectively). The H. influenzae conjugate vaccine was not regularly applied in this population during the study period. Regarding clinical presentation 60.5% of the episodes of pneumococcal bacteremia in pediatric patients occurred in male patients and 39.5% in female patients. Regarding age distribution of the episodes, 66.6% occurred in patients <2 years of age. More than one-half of our pediatric patients (57%) did not have any underlying disease. The remaining patients had, in order of frequency, congenital heart diseases (11%), HIV infection (9.7%), malignancies (5.5%) and a miscellany of other diseases (14%). Pneumococcal bloodstream infections were mostly community-acquired (85%) and the main sites of infection were: pneumonia (43%); upper respiratory tract infections (27.5%); and meningitis (8.3%). Pneumococcal bloodstream infections presented as primary bacteremia in 8% of the pediatric population and as occult bacteremia in 9.7%. All patients had high fever (>39°C), 65% had a high white blood cell count (> 15 000/μl) and 67% had >300 000 platelets/μl. There was no significant change in serotype distribution over time. The four most common serogroups were 19, 23, 6 and 14. Overall 75% of the responsible serogroups obtained in our series are included in the new heptavalent vaccine. The antimicrobial susceptibility of the bloodstream isolates was available for all the strains, and susceptible, intermediate and resistant strains were, respectively, as follows: penicillin (51, 25 and 24%); cefotaxime (73, 23 and 4%) and erythromycin (78, 0 and 22%). The antimicrobial susceptibility pattern remained basically the same during the study period without significant variations in the antimicrobial activity of penicillins, cephalosporins and macrolides. Two patients died of pneumococcal bloodstream infections. One of them was a 9-year-old patient with full blown AIDS who was admitted with bacteremic pneumococcal pneumonia without meningitis. He was in shock and died 14 days after admission. He was treated with iv cefotaxime from the moment of admission. Necropsy examination showed also Pneumocystis carinii pneumonia, disseminated cytomegalovirus disease, acute bronchopneumonia and herpetic esophagitis. The second patient was a 22-month-old infant with congenital heart disease who was admitted with septic shock, disseminated intravascular coagulation and multiorganic failure but without meningitis. He died after 24 h of admission and was treated with iv cefotaxime from the beginning. Discussion. Our study shows that the incidence of invasive bloodstream pneumococcal infection in our pediatric population is increasing yearly. This fact cannot be attributed to an increase in our pediatric population, which is decreasing at present, or to a deterioration of the standards of living or public health, the HIV epidemic or improved performance of microbiologic techniques for detection. Our data agree with those reported by Fraser et al. 6 in Israel. They demonstrated a significant increase in the incidence of invasive pneumococcal infections in children during the period 1989 through 1998. With regard to portals of entry, it is remarkable that >50% of our patients with bacteremia caused by S. pneumoniae did not have associated pneumonia. Series of bacteremic pneumococcal infections in developing countries report higher proportions of pediatric patients with pneumonia (>50%), whereas reports from well-developed nations show incidences of <50%. 7–11 The upper respiratory tract is a common source of pneumococcal bacteremia in children. 9, 12 Our serotype distribution is similar to that of other European countries, and 75% of serogroups responsible for pediatric bacteremia in our study are included in the currently available heptavalent vaccine. 6, 13–17 The worldwide scale of antimicrobial resistance in S. pneumoniae prompts the introduction of vaccines in this field. 18, 19 Serogroups 6, 14, 23, 19 and 9, the most prevalent serogroups in our series, are particularly associated with antibiotic resistance. 19, 20 Acknowledgments. We thank Thomas O’Boyle for his help with the English version of the manuscript. The manuscript was partially produced with public funds of the Spanish Pneumococcal Infection Study Network (G03/103).