Leukocyte counts in urine reflect the risk of concomitant sepsis in bacteriuric infants: a retrospective cohort study

Abstract

Background: When urine infections are missed in febrile young infants with normal urinalysis, clinicians may worry about the risk--hitherto unverified--of concomitant invasion of blood and cerebrospinal fluid by uropathogens. In this study, we determine the extent of this risk.

Methods: In a retrospective cohort study of febrile 0-89 day old infants evaluated for sepsis in an urban academic pediatric emergency department (1993-1999), we estimated rates of bacteriuric sepsis (urinary tract infections complicated by sepsis) after stratifying infants by urine leukocyte counts higher, or lower than 10 cells/hpf. We compared the global accuracy of leukocytes in urine, leukocytes in peripheral blood, body temperature, and age for predicting bacteruric sepsis. The global accuracy of each test was estimated by calculating the area under its receiver operating characteristic curve (AUC). Chi-square and Fisher exact tests compared count data. Medians for data not normally distributed were compared by the Kruskal-Wallis test.

Results: Two thousand two hundred forty-nine young infants had a normal screening dipstick. None of these developed bacteremia or meningitis despite positive urine culture in 41 (1.8%). Of 1516 additional urine specimens sent for formal urinalysis, 1279 had 0-9 leukocytes/hpf. Urine pathogens were isolated less commonly (6% vs. 76%) and at lower concentrations in infants with few, compared to many urine leukocytes. Urine leukocytes (AUC: 0.94) were the most accurate predictors of bacteruric sepsis. Infants with urinary leukocytes < 10 cells/hpf were significantly less likely (0%; CI:0-0.3%) than those with higher leukocyte counts (5%; CI:2.6-8.7%) to have urinary tract infections complicated by bacteremia (N = 11) or bacterial meningitis (N = 1)--relative risk, 0 (CI:0-0.06) [RR, 0 (CI: 0-0.02), when including infants with negative dipstick]. Bands in peripheral blood had modest value for detecting bacteriuric sepsis (AUC: 0.78). Cases of sepsis without concomitant bacteriuria were comparatively rare (0.8%) and equally common in febrile young infants with low and high concentrations of urine leukocytes.

Conclusion: In young infants evaluated for fever, leukocytes in urine reflect the likelihood of bacteriuric sepsis. Infants with urinary tract infections missed because of few leukocytes in urine are at relatively low risk of invasive bacterial sepsis by pathogens isolated from urine.