Use of plasmid analysis to determine the source of bacterial invasion of the urinary tract.

Gram negative colonisation and infection of the urinary tract is a well recognised complication of the neuropathic bladder caused by spinal cord injury (SCI). K. pneumoniae accounts for one third of all urinary tract infections in hospitalised SCI patients. Plasmid analysis has been shown to reliably fingerprint bacterial strains, particularly K. pneumoniae, so that growth from two separate locations in or on the body can be accurately analysed as to migration from a reservoir to a target location. Eighty seven hospitalised SCI patients on intermittent catheterisation for a total of 586 patient-weeks were studied. Twice weekly catheterised urine specimens and once weekly rectal swab cultures were taken from each patient. Thirty seven patients experienced at least one clinically significant (colony count greater than 10,000/mL) urinary tract colonisation caused by K. pneumoniae, representing 66 total colonisations. Further analysis of 31 of these 37 patients revealed: K. pneumoniae in all of their stool cultures (p less than 0.05) and the identical strain of K. pneumoniae in the urine as well as the stool in 72% of the 66 colonisations (p less than 0.05). Analysis of 14 patients without K. pneumoniae urinary colonisations showed absence of faecal K. pneumoniae in 3, and predominant growth in only 4. In 22 of the 37 patients, multiple K. pneumoniae urinary colonisations were noted, representing 27 pairs of colonisation. Fifteen of the pairs were found to be relapsing (caused by two identical bacterial strains), and 12 were recurrent (caused by two different bacterial strains). Thirteen of the 15 relapsing pairs also had identical urine and stool K. pneumonia strains (p less than 0.05). All colonisations were treated with appropriate antibiotics based on culture and sensitivity reports. Fourteen of the 15 relapsing colonisation pairs have identical antibiograms (p less than 0.05), while all 12 of the recurrent colonisation pairs had different antibiograms (p less than 0.05). The differences noted on sensitivity patterns (antibiograms) correlated with differences among strains of K. pneumoniae based upon plasmid analysis. Treatment of bacteriuria did not affect the nature of repeated colonisations regardless of the antibiotic chosen, the route of administration or the duration of treatment.