Secondary Logo

Journal Logo

Original Studies

Recurrent Systemic Pneumococcal Disease in Children

Mason, Edward O. Jr. PhD*; Wald, Ellen R. MD; Tan, Tina Q. MD; Schutze, Gordon E. MD§; Bradley, John S. MD; Barson, William J. MD; Givner, Laurence B. MD#; Hoffman, Jill MD**; Kaplan, Sheldon L. MD*

Author Information
The Pediatric Infectious Disease Journal: June 2007 - Volume 26 - Issue 6 - p 480-484
doi: 10.1097/INF.0b013e31805ce277
  • Free


Recurrent systemic pneumococcal infection is known to occur in immunocompromised patients and patients with underlying conditions such as sickle-cell disease, asplenia, HIV or intracranial structural abnormalities (CSF leaks).1–4 In contrast, recurrent systemic infection due to Streptococcus pneumoniae in nonimmunocompromised children is rare.5 As part of a multicenter study of pneumococcal disease in children in over 4000 patients during a 12-year period, we documented the clinical findings, antibiotic susceptibility, serotype distribution and DNA relatedness among pneumococcal isolates from children with relapsing or recurrent systemic pneumococcal infection over a 12.3-year period.


Between September 1, 1993 and December 31, 2005, investigators at 8 pediatric hospitals in Texas, Pennsylvania, Illinois, California, Ohio, Arkansas and North Carolina identified prospectively cases of invasive pneumococcal disease and completed retrospectively standard case report forms, which included demographics and clinical information regarding those infections.6,7 These forms along with the bacterial isolate(s) were forwarded to the Infectious Diseases Laboratory at Texas Children's Hospital, Houston, Texas. The data were entered into a central computer database and children diagnosed with >1 episode of invasive pneumococcal infection were identified. Isolates from these recurrent infections were tested for antibiotic susceptibility, serotype and DNA relatedness. This study was approved by the Institutional Review Board of Baylor College of Medicine and the participating centers.

Antibiotic susceptibility was determined for penicillin and ceftriaxone by microbroth dilution using the Clinical and Laboratory Standards Institute methodology.8 Broth medium consisted of cation-supplemented Mueller-Hinton II broth (BBL, Sparks, MD) with 3% lysed horse blood. Microdilution plates were prepared and used within 1 week of testing. The bacterial inoculum was prepared by diluting actively growing strains to a 0.5 McFarland standard with Mueller-Hinton broth to obtain a final dilution in each well of 5 × 105 cfu/mL. Inoculum consistency was verified by colony counts. S. pneumoniae ATCC No. 49619 was used as the quality control strain.

Isolates were serotyped or serogrouped by the capsular swelling method with antisera from Statens Seruminstitut (Copenhagen, Denmark). Strains grown overnight on sheep blood agar were suspended in phosphate-buffered saline and reacted with omni and pooled antisera (pools A–H). Strains showing evidence of capsular swelling or agglutination on light microscopy were tested with the individual specific antisera contained in the reactive pool and serotyped using the appropriate factor antisera when necessary.

Pulse field gel electrophoresis (PFGE) was performed using clamped homogeneous electrical field electrophoresis (CHEF-DR III (Bio-Rad Laboratories, Hercules, CA) equipment and a modification of the protocol by Murray et al.9 Agarose plugs were prepared and digested with SmaI and electrophoresed in 1% agarose in 0.35× TBE buffer using the parameters: pulse times between 1 and 17 seconds for 18.5 hours, at 14°C at 6 V/cm at a 120° angle. Relatedness was determined by the criteria of Tenover et al; isolates were considered unrelated if there was a difference of >4 bands.10

This study reviewed only prespecified clinical and laboratory data and systematically obtained information on the humoral and cellular immune competence of study subjects was not available.

Dichotomous variables were analyzed using the χ2 or χ2 test for trend (True Epistat – Epistat Services, Richardson, TX).


Four thousand sixty-seven children were diagnosed with systemic pneumococcal disease during the 12.3-year study period between September 1, 1993 and December 31, 2005. Either as a result of their pneumococcal infection or as a complication of comorbidity, 81 children (2%) died leaving 3986 patients who survived a 1st infection. One hundred and eight episodes of recurrent disease were seen in 90 children (2.6%) during that period; 75 children experienced 2 infections, 12 children experienced 3 infections and 3 children experienced 4 infections. The number of recurrent infections remained relatively constant until 2001 when the numbers declined dramatically (Fig. 1). Primary bacteremia accounted for 67 (62%) of the episodes. Other infections were pneumonia (25), meningitis (10), peritonitis (3), cellulitis (1), endocarditis (1) and infected facial cystic hygroma (1). Two children died, 1 following their 2nd episode (asplenia) and 1 following their 3rd episode (HIV). Children with no known underlying conditions accounted for 19 of the 90 patients (21%), and all but 1 of these had only 1 recurrence of infection. Underlying conditions associated with recurrent infection are shown in Table 1.

Cases of recurrent systemic pneumococcal infection over the study period.
Underlying Disease in Children With Recurrent Pneumococcal Infection

Fourteen of the 15 children with >2 episodes of infection had underlying conditions: HIV (6), leukemia (1), neurodegenerative disease (1), Mondini cochlear malformation (1), Down syndrome (1), end-stage renal disease (1), heart transplant with asplenia (1), end stage liver disease (1) and failure to thrive (1) (Table 2). The “normal” child with 3 episodes of infection and no underlying condition was given penicillin prophylaxis after the 1st episode of bacteremia (caused by serotype 14) but had a 2nd episode of bacteremia 4 weeks later caused by a different serotype (6B). Immunologic assessment was normal and he was given intravenous immunoglobulin for 1 year. He was reinfected 7 weeks after his 2nd infection with a serotype 14 pneumococcus genotypically identical by PFGE to the 1st isolate. He was immunized with polyvalent pneumococcal vaccine (PNEUMOVAX23) at 3 years of age and subsequently lost to follow-up.

Children With 2 or More Recurrent Invasive Pneumococcal Infections

At the time of their first illness, there was no statistical difference in age (3.45 years; SE = 0.40 vs. 2.99 years; SE = 0.06) or gender (64% male versus 58% male) between children with recurrent systemic pneumococcal infections and children with only 1 pneumococcal infection during the same time period (1993–2005). The mean duration between 1st and 2nd infections of 22.9 weeks for children with no known underlying condition and 43.0 weeks for children with an underlying condition was significantly different (P < 0.001). At first recurrence, children with no known underlying condition were younger (median = 1.7 years.) than children with an underlying condition (median = 3.3 year) (P < 0.006). When analyzed, stratifying for age, there was no difference between children less than or equal to 2 years or greater than 2 years either with or without underlying conditions with regard to the duration of time between infections.

Overall, 70 episodes of recurrent disease among the 90 patients were caused by a different serotype or a different genotype as demonstrated by the PFGE pattern (Table 3). As the duration between infection increased there was a significant shift at about 8 weeks to infections being caused by a different strain (χ2 for trend = 0.0004). The time between infections ranged from 12 days to 5.8 years. Twenty-four recurrences were caused by the same serotype and PFGE pattern. The time between these infections ranged from 11 days to 1.4 years. Four children whose recurring infections were caused by the same serotype and pulse field types had no underlying illness. Two children had recurrent infections caused by the same serotype, but the pulse field type indicated that they were infected by different strains. The recurrent strains of 14 episodes of infection could not be evaluated because 1 of the pairs was lost or nonviable. The longer the interval between recurrences, the more likely the infections were caused by different strains (serotype or genotype) P = 0.0004 (Table 3). Fifteen of 20 (75%) children who had recurrent infections caused by the same strain (serotype and genotype) of pneumococcus and 47 of 59 (79%) of those infected by a different strain had an underlying condition (P=0.79).

Time Between Recurrences and Strain Similarity Based on Serotype and Genotype

Sixteen children had intervals <30 days between infections, many of which might be considered by some a relapse of the original infection. However, 7 of these were caused by different strains (6 with different serotypes and different PFGE fingerprints and 1 with the same serotype but a different genetic fingerprint by PFGE); 7 recurrences were caused by identical strains (serotype and PFGE), and 2 episodes could not be classified because a second isolate was not available for typing or PFGE (Table 4).

Recurrent Invasive Pneumococcal Infections With Recurrence Intervals <30 Days

Overall, 36.5% of isolates were nonsusceptible to penicillin (22.3% intermediate, 14.2% resistant). Analysis of the 1st and 2nd infections revealed that in 42 cases (47.2%) the original and subsequent infections were caused by penicillin susceptible strains. In 16 cases (17.9%) both episodes were caused by nonsusceptible strains. In 18 cases (20.2) the 1st infection was caused by a susceptible strain and the 2nd infection was caused by a nonsusceptible strain. In 13 cases (14.6%) the 1st infection was caused by a nonsusceptible strain and the 2nd infection was caused by a susceptible strain.

Twenty-five children received the 23-valent pneumococcal vaccine before an initial or subsequent pneumococcal infection (mean age at vaccination = 4.7 years). All children immunized before their pneumococcal infection had underlying conditions. Seven children, all with underlying conditions, were immunized with the 7-valent conjugate pneumococcal vaccine following its licensure in February of 2000. Five received only 1 dose prior to their infection; 2 of these were infected by serotype 9V, 2 by serotype 19A (not included in the vaccine) and 1 by 18C. One child received vaccine at 2, 6 and 8 months of age and had infection caused by a serotype 13, 2 years after the last vaccination. Another child with HIV received 3 doses following the initial bacteremia and was infected after the 2nd and 3rd doses by serotype 6B pneumococci identical to the strain causing the first infection. Prior to the introduction of the 7-valent conjugate pneumococcal vaccine (before July 2000), there were 172 episodes of systemic pneumococcal infection in 77 children, and after July 2000, there were 37 cases in 20 children. There was no difference between age at reinfection (2.7 vs. 4.8 years) or proportion with underlying conditions (15% vs. 21%) in children during the pre- and post-7-valent conjugate vaccine periods.


Recurrent invasive pneumococcal disease usually occurs in immunocompromised children and those children with a variety of underlying conditions such as HIV, asplenia, sickle-cell disease, renal disease or cranial malformations.3,11–13 Recurrent invasive disease in children without the above underlying problems is uncommon and may be an indication for evaluation for an immunologic defect. King et al14 studied 15,782 episodes of invasive pneumococcal disease between 1994 and 1998 and found 318 (including 65 children) or 2.3% survivors of their first infection who had ≥1 recurrent infection. While the majority of infections in that study occurred in adults, the rate of recurrence, 2.6%, is almost identical to the rate found in our study. Also, King and colleagues excluded infections that recurred within 29 days of each other (number not detailed), not considering them separate cases. Of our 16 cases that recurred within 30 days, 7 were new infections caused by isolates with different serotypes (6) or genotypes (1).

In the present study, during a 12.3 year period, we found that 19 of 90 (21%) children, who were apparently immunologically normal, had recurrent invasive pneumococcal disease at intervals ranging from 2 to 73 weeks between infections. With 1 exception, specific rigorous immunologic assessment of this group of children was not attempted; accordingly, subtle abnormalities might not have been detected. In the 1 case of the only apparently normal child to have more than 1 recurrence, evaluation did not reveal any defect. Over a 3-year interval, Lee et al5 found 5 infants with recurring invasive pneumococcal disease and 2 of these had no immunologic abnormality. All the recurrences were caused by different serotypes, including 1 infant with a recurrence after 11 days. Ten apparently normal children with repeated bacteremic pneumococcal infections were reported by Orlicek et al.2 In that study, during a 3-year period, 16 repeated infections occurred among 394 invasive episodes (4.1%). Serotype information was available for only 3 episodes, making it difficult to distinguish recurrence from relapse in those infections separated by short intervals.

Repeated episodes of invasive pneumococcal disease could represent relapses resulting from inadequate or inappropriate therapy, in which case the isolates are identical, or recurrences resulting from acquisition of different strains as defined by serotype or genotype of the strains. It is also possible that a reinfection with an identical strain could be a reinfection via strains that circulate among day care contacts or within the families of these children.14 Although relapses typically have intervals between infections of shorter durations, the present study demonstrates that almost half of these infections recurring within 30 days were caused by different strains. Further, recurrences caused by the same strain were documented to occur with times between infections of up to 72 weeks. Occurrence of pneumococcal disease in children who have been immunized with at least 2 doses of the conjugate vaccine should prompt further study to determine whether the causative strain is a serotype contained in the pneumococcal conjugate vaccine. If so, an assessment of the patient's immune system should be considered.15,16

The ability of pneumococci to switch capsule type was first described by Avery et al16 in 1955.17 Capsule switching is detected by molecular typing techniques and is documented when different serotypes display a similar or identical genotype.18 This phenomenon occurring in vivo was described by Nesin et al19 and is postulated to be a mechanism by which the pneumococcus evades host defenses. In a study of nasopharyngeal carriage of S. pneumoniae, Meats et al20 found that isolates from the same children that differed by serotype were also genotypically different and thus ruled out capsule switching. We also failed to detect any evidence of capsule switching in recurrent invasive disease in children and found 2 instances of reinfection by same serotypes belonging to different genotypes indicating acquisition of a different strain.

In summary, during more than 12 years of study of pneumococcal disease at 8 children's hospitals in the United States, we found a 2.6% rate of recurrent infection. The incidences of recurrence declined markedly after the introduction of the conjugate pneumococcal vaccine in 2000. While most of the children with recurrent disease had underlying conditions that are known to predispose them to pneumococcal infection, one-fifth of children had no known underlying disease. Recurrences of infection within 30 days are often considered relapses of the same infection; however, 7 of 16 such cases in our study were caused by acquisition of a new strain with 1 of these 7 cases caused by different strains of the same serotype. Thus, relapse of infection requires documentation that the pneumococcal isolates are not only the same serotype, but also have the same genotype patterns.


The authors thank the following for their invaluable help making this study possible: Linda Lamberth BS, Andrea Forbes RN, Wendy Hammerman RN, Nancy Tucker RN, Bev Petrites RN, BSN, Vanessa Guerrero, RN, and the staff of the Microbiology Laboratory at Texas Children's Hospital.


1. Gransden WR, Eykyn SJ. Recurrent pneumococcal bacteremia [comment]. Clin Infect Dis. 1992;15:1072–1073.
2. Orlicek SL, Herrod HG, Leggiadro RJ, Luedtke G, English BK. Repeated invasive pneumococcal infections in young children without apparent underlying immunodeficiency. J Pediatr. 1997;130:284–288.
3. Coccia MR, Facklam RR, Saravolatz LD, Manzor O. Recurrent pneumococcal bacteremia: 34 episodes in 15 patients. Clin Infect Dis. 1998;26:982–985.
4. McEllistrem MC, Mendelsohn AB, Pass MA, et al. Recurrent invasive pneumococcal disease in individuals with human immunodeficiency virus infection. J Infect Dis. 2002;185:1364–1368.
5. Lee W, Austrian R, Weiser JN. Recurrent pneumococcal bacteremia in normal children. Pediatr Infect Dis J. 1994;13:231–233.
6. Kaplan SL, Mason EO Jr, Barson WJ, et al. Three-year multicenter surveillance of systemic pneumococcal infections in children. Pediatrics. 1998;102:538–545.
7. Kaplan SL, Mason EO Jr, Wald E, et al. Six year multicenter surveillance of invasive pneumococcal infections in children. Pediatr Infect Dis J. 2002;21:141–147.
8. Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing. CLSI Document M100-S16. Sixteenth Informational Supplement. Pennsylvania: Clinical and Laboratory Standards Institute; 2006.
9. Murray BE, Singh KV, Heath JD, Sharma BR, Weinstock GM. Comparison of genomic DNAs of different enterococcal isolates using restriction endonucleases with infrequent recognition sites. J Clin Microbiol. 1990;28:2059–2063.
10. Tenover FC, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol. 1995;33:2233–2239.
11. Kuhls TL, Viering TP, Leach CT, Steele MI, Haglund LA, Fine DP. Relapsing pneumococcal bacteremia in immunocompromised patients. Clin Infect Dis. 1992;14:1050–1054.
12. Gransden WR, Eykyn SJ. Relapsing pneumococcal bacteremia in immunocompromised patients. Clin Infect Dis. 1992;15:558–559.
13. Reinert RR, Büssing A, Kierdorf H, Kühnemund O, Kaufhold A. Recurrent systemic pneumococcal infection in an immunocompromised patient. Eur J Clin Microbiol Infect Dis. 1994;13:304–307.
14. King MD, Whitney CG, Parekh F, Farley MM. Recurrent invasive pneumococcal disease: a population-based assessment. Clin Infect Dis. 2003;37:1029–1036.
15. American Academy of Pediatrics Committee on Infectious Diseases. Recommendations for the prevention of pneumococcal infections, including the use of pneumococcal conjugate vaccine (Prevnar), pneumococcal polysaccharide vaccine, and antibiotic prophylaxis. Pediatrics. 2000;106:362–366.
16. American Academy of Pediatrics. Pneumococcal infections. In: Pickering LK, Baker CJ, Long SS, McMillan JA, eds. Red book: 2006 Report of the Committee on Infectious Diseases. 27th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2006.
17. Avery OT, MacLeod CM, McCarty M. Studies on the chemical nature of the substance inducing transformation of pneumococcal types: induction of transformation by a desxoyribonucleic acid fraction isolated from a pneumococcus type III. Mol Med. 1955;1:344–365.
18. McEllistrem MC, Adams J, Mason EO, Wald ER. Epidemiology of acute otitis media caused by Streptococcus pneumoniae before and after licensure of the 7-valent pneumococcal protein conjugate vaccine. J Infect Dis. 2003;188:1679–1684.
19. Nesin M, Ramirez M, Tomasz A. Capsular transformation of a multidrug-resistant Streptococcus pneumoniae in vivo. J Infect Dis. 1998;177:707–713.
20. Meats E, Brueggemann AB, Enright MC, et al. Stability of serotypes during nasopharyngeal carriage of Streptococcus pneumoniae. J Clin Microbiol. 2003;41:386–392.

Streptococcus pneumoniae; invasive pneumococcal disease; recurrent infection

© 2007 Lippincott Williams & Wilkins, Inc.