Are resistance patterns in uropathogens published by microbiological laboratories valid for general practice?

During 7 months from August 1994, 171 urine samples were collected consecutively in general practice in Western Norway from female patients with suspected lower urinary tract infection. For each of the 171 samples, 2 samples from adult females received from general practice at the microbiological laboratory on the same day were selected using a predetermined system. Samples noted as treatment controls and samples from pregnant patients were discarded. Bacteriuria was found in 101/171 (59.1%) vs 220/342 (64.3%) of the samples. The general practice material contained more bacteriuric samples with Escherichia coli (83.2% vs 71.8%, p<0.05) and Staphylococcus saprophyticus (11.9 vs 6.4%), and fewer with other Gram-negative rods (4.0% vs 15.9%, p<0.01) and enterococci (1.0% vs 5.9%, p<0.01). The frequency of resistant isolates was substantially lower in the samples from general practice for all antibacterial agents tested: amoxycillin 18.9% vs 23.9%, mecillinam 1.1% vs 4.7%, trimethoprim 12.9% vs 18.5%, cotrimoxazole 12.0% vs 15.4%, sulphonamide 20.0% vs 28.4%, nitrofurantoin 3.0% vs 9.7% (p<0.05). Data from local laboratories exaggerate the resistance problems among uropathogens found in urine samples in general practice.