Home screening for bacteriuria in children with spina bifida and clean intermittent catheterization

Clean intermittent catheterization (CIC) and antibiotic prophylaxis have reduced the incidence of parenchymal kidney damage in children with spina bifida [1‐3]. In these patients, the main objective of any urinary diagnostic test is to detect or exclude urinary tract infections (UTIs) to prevent under- and over-treatment. The diagnosis of UTI is made on clinical symptoms, leukocyturia and significant bacteriuria (SBU). Several simple tests to detect a UTI, such as dip slides and leukocyte esterase tests (LET) were studied extensively. Two recent meta-analyses showed that a LET had a negative predictive value (NPV) of 90%, with a positive predictive value (PPV) of 60% [4, 5]. A UTI therefore has to be confirmed with a urine culture [4], which takes at least three days, and treatment is postponed. Dip slides with two or three culture media were tested to diagnose SBU in primary care [6‐9]. As PPV was poor, it was concluded that the use of dip slide urine cultures should only be used to exclude SBU.

The aim of the present study was to evaluate the reliability of the LET and dip slide in children with spina bifida. Children with clinical symptoms of UTI participated in a parallel study, and were not included in this study [10]. Only children with asymptomatic SBU were included. We also assessed whether general patient characteristics, such as sex, age and use of prophylactic antibiotics can predict asymptomatic SBU. The hypothesis was that LET and Uricult® Duo dip slide are as accurate as laboratory cultures in determining significant bacteriuria.

One hundred and twelve patients with spina bifida on CIC known at the Gasthuisberg University Hospital Leuven, Belgium participated in the study. Patients catheterized themselves or were catheterized by their parents or primary care takers for a fresh urine sample at the quarterly control visit. Patients who had completed treatment for UTI less than 4 weeks before the visit to the clinic, or who had febrile episodes immediately preceding the visit, or a clinical suspicion for a UTI at the visit were excluded. A leukocyte esterase test (LET, Combur-2® test strip, Roche, Switzerland) was performed on the urine sample, regarded “positive” in every range of discoloration. The sample was also inoculated onto a dip slide (Uricult® Duo, Orion Diagnostics, Finland), which contains an aselective cystine-lactose-electrolyte deficient (CLED) agar for Gram-positive bacteria and enterobacteriaceae, and MacConkey agar for non-glucose fermenting Gram-negative rods. The dip slide was incubated in a bottle warmer (Philips® SBC 215/00, Philips SA, Belgium) at 36.3 ± 2.5 °C, as measured over 48 hours with a calibrated Dickson® SK 180 temperature logger (Dickson Corporation, Addison, USA). After 24 hours in the bottle warmer, the dip slide was evaluated for colony forming units by a trained research nurse, and the result was reported as ‘no growth’ or ‘growth’ (visible colonies). The same urine sample was also sent for ‘gold standard’ agar plated culture. SBU was defined as a colony count of ≥10⁴ per milliliter of one single species in a catheterized sample. Of patients with multiple samples, only the first was used for analysis. To establish whether general patient characteristics could discriminate SBU from no SBU, logistic regression was used with gold standard outcome (positive / negative) as dependent variable and age and sex as independent variables. This model was then extended with prophylaxis (model 2), LET testing (model 3), and dip slide testing (model 4). Model results are expressed as odds ratios (95% confidence intervals, and p-values). Discriminative capacity for these four models was evaluated using areas under the Receiver Operator Characteristic (ROC) curves (AUC). This study is approved by the ethics committee from the Leuven University Hospital, and performed after parental or guardian consent.

Of the 112 asymptomatic patients, 45 had a positive agar plated culture, hence the prior probability for SBU was 40%, which is consistent with previous studies. The patients had an age range of 0 to 35 years (median 13.0, long-term spina bifida follow-up patients over 18 years of age are included). Fifty (45%) were boys and 61 (68%) were on antibiotic prophylaxis. Table 1 shows the results of four consecutive models predicting the gold standard culture outcome. The LET had the strongest discriminative power, while dip slide testing did not add significantly. In Figure 1, the discriminative capacity of the models is shown graphically as ROC curves for age and sex (AUC = 0.64, p = 0.01); age, sex and prophylaxis (AUC = 0.76, p = 0.003); age, sex, prophylaxis and LET (AUC = 0.91, p < 0.0001) and age, sex, prophylaxis, LET and dip slide (AUC = 0.91, p < 0.0001). As this study addressed the role of LET and dip slide to rule out SBU in spina bifida patients without complaints of UTI, we proceeded with these tests only, as shown in Table 2. Given an a-priori chance of SBU of 40%, a positive LET had a PPV of 72%, while a negative LET substantially decreased the chance of a SBU to 4% (NPV = 96%). Dip slide testing had a similar PPV (73% versus 72% for LET) but substantially lower NPV (78% versus 96% for LET). Combining LET with dip slide improved neither PPV (positive LET 72% versus both LET and dip slide positive 74%) nor NPV (negative LET 96% versus both LET and dip slide negative 98%). Pathogens found were Escherichae coli (N=26), Klebsiella pneumonia (4), Streptococcus species (4), Enterococcus species (3), Proteus mirabilis (3), Pseudomonas aeruginosa (2), Serratia marcescens (1), Staphylococcus aureus (1) and Providencia rettgeri (1). In three of the 45 positive cultures (one Streptococcus, one Staphylococcus and one Enterococcus species) the LET was negative, resulting in 93.3% sensitivity. There were 16 false positive LETs in 67 negative cultures, resulting in 76% specificity.

In this study of 112 spina bifida patients on clean intermittent catheterization, a negative LET excludes SBU in a home setting with a NPV of 96%. A negative dip slide alone was not effective to rule out SBU, and a negative LET together with a negative dip slide did not improve NPV. Both a positive LET and dip slide had a false positive rate of more than 20 percent compared to laboratory cultures, and cannot be used to diagnose SBU.

In home testing of spina bifida children on clean intermittent catheterization, leukocyte esterase testing can be used to exclude significant bacteriuria. Both leukocyte esterase test and dip slide are not sensitive enough to predict significant bacterial growth, and a agar plated culture should therefore be performed when either test is positive. Other than serving as transport medium, dip slide testing has no added to diagnose or exclude significant bacteriuria.