Accuracy of urinary dipstick for glucose and protein determination in canine cerebrospinal fluid

To evaluate the precision of urinary dipstick (UD) to assess protein and glucose concentrations in canine CSF samples compared to the standard methods.

Cerebrospinal fluid protein and glucose were measured in 22 samples from dogs with neurological diseases affecting the CNS using UD and biochemistry (pyrogallol red and glucose oxidase reaction, respectively). Results were converted into scores to allow comparison between methods. The proportion of divergence between methods and its CI were calculated. The sensitivity (Se), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV), and accuracy (Ac) of UD were determined for 2 cutoff levels of CSF protein (15 and 30 mg/dL) and glucose (40 and 100 mg/dL).

The proportion of divergence between methods was 64% (95% CI, 44% to 84%) for CSF protein (representing 14 of 22 erroneous samples), of which 92.9% (13 of 14) had a UD score lower than biochemistry. For CSF glucose, 73% (16 of 22 erroneous samples; 95% CI, 54% to 91%) had divergence between methods, of which 87.5% (14 of 16) had a UD score higher than biochemistry. Urinary dipstick had better results when the cutoff level was 15 mg/dL for protein (Se, 78.9%; Sp, 66.7%; PPV, 93.7%; NPV, 33.3%; Ac, 77.3%) and 40 mg/dL for glucose (Se, 89.5%; Sp, 33.3%; PPV, 89.5%; NPV, 33.3%; Ac, 81.8%) concentrations.

Urinary dipstick was unreliable in estimating canine CSF protein and glucose concentrations.

The UD underestimated protein and overestimated glucose levels in the canine CSF, which could have a significant diagnostic impact and should discourage its use as a bedside test.