Uninterpretable urine samples from children: could molecular analytic methods be more accurate than culture?

Problem

The prompt and accurate diagnosis of urinary tract infection (UTI) in young children is important both to reduce morbidity and reduce the possibility of renal scarring. The diagnosis usually relies on microbiological confirmation as presenting symptoms are non-specific. Urine sampling is difficult in young children and contamination is a big problem. Culture results are often inconclusive with 48% urine culture results from children less than five reported as ‘mixed growth’, requiring repeat urine samples. False positive and negative results are also suspected. We aimed to explore whether molecular methods have the potential to more accurately diagnose UTI in children compared with standard culture.

Approach

Urine samples were obtained from children aged less than five from healthy and acutely ill children, using clean catch or nappy pad sampling techniques. Urine samples from the ill children were analysed by National Health Service (NHS) microbiology laboratories according to their standard operating procedures and frozen. Molecular methods (qPCR and pyrosequencing) were used to determine the type and load of bacteria in the samples from acutely ill and healthy children. The results for the healthy children were used as a baseline. Results from molecular methods were compared with standard NHS culture results for the acutely ill children. We considered molecular methods possibly indicative of true infection if nucleic acid from the dominant bacteria represented more than 107 reads/ml and more than 50% of the overall bacterial DNA.

Findings

128 urine samples from acutely ill children, were analysed using standard NHS urine culture. These, plus a further 57 urine samples from healthy children totalling 185 samples, were analysed using PCR and pyrosequencing techniques.The urine from healthy children had a wide variety of bacteria, with the predominant bacteria representing less than 50% of overall bacterial growth in all cases. For the ill children, findings from molecular methods largely reflected NHS culture results. However, 23.9% of NHS negative results had >107 reads/ml with >50% of total bacterial DNA from the dominant organism, suggesting false negative results. Similarly, 26.7% of positive results were suggestive of false positive results. In those with mixed growth on culture, 48.8% had molecular results suggestive of UTI.

Consequences

False positive and negative results are likely to arise when using culture to diagnose UTI, particularly in young children where sampling is difficult and there is a high risk of contamination. Our findings support this view with an estimation of the extent the problem. Given the importance of accurate diagnosis in young children, alternatives to traditional culture need to be considered. Diagnostic thresholds would need to be validated, but we have shown that molecular methods may provide a diagnostic alternative to culture in children, which is less affected by sampling method and contamination.