Streptococcus pneumoniae in Urinary Tracts of Children with Chronic Kidney Disease

Streptococcus pneumoniae is not commonly considered an agent of urinary tract infections. We report 3 children with urinary tract abnormalities who had high numbers of S. pneumoniae in their urine (>104 CFU/mL) and varying clinical symptoms.

In September 2009, a 12-year-old boy (case-patient 2) sought care at the nephrology department of the University of Heidelberg Children's Hospital for his yearly control examination 7 years after kidney transplantation. He had no clinical signs of current infection. His medical history was remarkable for kidney insuffi ciency, congenital obstruction and refl ux in the urethral valve, kidney transplantation in 2002, ileocecal pouch, chronic transplant nephropathy, metabolic acidosis, hypertension, and renal anemia. A urine sample showed 10 5 CFU/mL S. pneumoniae and 10 3 CFU/mL Enterobacteriaceae. Urine leukocyte levels were slightly elevated (16 cells/μL). Further analysis showed that S. pneumoniae was serotype 34 and fully susceptible to all antibacterial drugs tested (Table 1).
In November 2009, a 7-year-old girl (case-patient 3) was sent to the emergency department of the University of Heidelberg Children's Hospital by her pediatrician because of abnormal results in a control urine sample 4 weeks after percutaneous nephrolithoapraxia and concrement removal. She was known to have cystinuria and had already undergone extracorporeal shock wave treatment with concrement removal in 2004. Her temperature was slightly elevated (37.5°C), but she had no dysuria or pain. Urinalysis showed elevated leukocyte levels (158 cells/μL), and 10 4 CFU/mL S. pneumoniae could be grown. Because of the mild symptoms, no antimicrobial drug treatment was started. Further analysis showed that the isolate was a 19F serotype (Table 1).
For each patient, urine was routinely cultured as follows: 2 samples of 1 μL each were placed on a 5% sheepblood agar plate and a MacConkey agar plate. chromID CPS medium (bioMérieux, Nürtingen, Germany) was injected with 10 μL of urine. All plates were incubated for 18-24 h at 36°C ± 1°C in ambient air (5). Susceptibility testing was performed by using the BD Phoenix Automated Microbiology System with SMIC/ID panels (Becton Dickinson, Heidelberg, Germany).

Discussion and Conclusions
The literature on urinary tract infections with S. pneumoniae is scarce. In 1980, Green  A UTI is defi ned as bacteriuria (>10 5 CFU/mL in adults, >10 4 CFU/mL in children) of 1 uropathogen and typical clinical signs, i.e., dysuria and urgency. Depending on the age of the patient, clinical signs might be less typical, especially in children <2 years of age. Generally pyuria is present (8,9). By contrast, asymptomatic bacte-riuria is defi ned as a uropathogen (>10 5 CFU/mL in adults, >10 4 CFU/mL in children) without pyuria (<10 leukocytes/ μL) (9).
Applying these criteria to the 3 cases in this report yields the following results. Assuming that S. pneumoniae is a uropathogen, case 1 is a UTI accompanied by septicemia. No other focus of infection with S. pneumoniae was apparent or could be identifi ed. We believe this is an ascending UTI in a boy with known oliguria from bilateral  cystic-dysplastic kidneys. Because case-patient 3 showed only mild symptoms, diagnosis of UTI is not obligatory. Nevertheless, it is not a mere pneumococcosuria because of the high numbers of leukocytes in the urine (158 cells/ μL).
Case 2 is more diffi cult to classify. The episode described might be pneumococcosuria because the leukocyte level is not high; nevertheless, it is above normal. Reassessment of all microbiological data of case-patient 2 indicated that since 2005 we have received 29 different urine samples. In March 2007, we had already identifi ed S. pneumoniae (10 4 CFU/mL) and pyuria (92 CFU/μL). Additionally, 10 4 CFU/mL Proteus mirabilis had been present and considered the cause of the pyuria; 11 of 27 samples had contained α-hemolytic streptococci >10 3 CFU/mL (Table  2). Unfortunately, no test to differentiate between pneumococci and the other α-hemolytic streptococci was originally performed on these bacteria. Therefore, we can only speculate whether at least some of these samples contained S. pneumoniae. Taking these facts into account, we believe that this might be a case of bacterial persistence according to the criteria described by Chang and Shortliffe (cultures ± same organism) (8).
Reassessment of all available laboratory data on casepatient 1 showed that we received 3 urine samples before the described episode and 8 afterward. In none of the samples did we fi nd S. pneumoniae or α-hemolytic streptococci. From case-patient 3 we did not receive any other material than described.
Further analysis of the S. pneumoniae isolates indicated that all 3 children were affected by different serotypes. Therefore, we do not have an indication that a single serotype has a predilection for the urinary tract. In theory, 2 cases could have been avoided by vaccination, i.e., case 3 (serotype 19F) and case 1 (serotype 15B). However, casepatient 3 was 4 years of age when the routine vaccination program aimed at children <2 years of age started in Germany in 2006. Because no catch-up program existed, this child was never vaccinated against pneumococci. Casepatient 1 was regularly vaccinated with heptavalent pneumococcal conjugate vaccine in 2006-2007 but was just 23 months old at disease onset disease, i.e., 1 month too young to be eligible for the 23-valent pneumococcal polysaccharide vaccine. Case-patient 2 is not yet vaccinated against pneumococci. Thus, we suggest that S. pneumoniae be added to the potential UTI-causing pathogens in children with urinary tract abnormalities.