General Practice

Fever can cause pyuria in children

BMJ 1995; 311 doi: http://dx.doi.org/10.1136/bmj.311.7010.924 (Published 07 October 1995) Cite this as: BMJ 1995;311:924
  1. G M Turner, paediatric senior registrara,
  2. M G Coulthard, consultanta
  1. aPaediatric Nephrology Unit, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP
  1. Correspondence to: Dr Coulthard.
  • Accepted 7 July 1995

Rapid, accurate diagnosis of urinary tract infection in childhood is important, and pyuria is often considered critical in diagnosis in addition to the presence of large numbers of bacteria. Clinically important pyuria has been defined as more than 10x106 leucocytes/l of urine. Only 1.5% of healthy schoolchildren exceed this.1 The view that pyuria might occur often in febrile children without a urinary tract infection, perhaps as a non-specific response to fever2 has recently been described as untenable.3 We investigated further our clinical impression that pyuria does indeed occur in febrile children without a urinary tract infection.

Subjects, methods, and results

We studied 157 children (73 girls), mean age 2.7 years (range 12 days to 15 years) in a paediatric day unit. Seventy febrile children (temperature >/=38°C) attended for clinical assessment, and 87 afebrile children (temperature </=37°C) attended for various investigations. The febrile and afebrile groups were similar in distribution of age and sex. No child had haematuria, proteinuria, or a urinary tract infection at the time of study, and none had renal tract disease. One of us (GMT) examined a urine sample within 30 minutes of collection by means of a Neubauer counting chamber and phase contrast microscopy at a magnification of 400. Leucocytes were counted in 0.9 μl of uncentrifuged midstream urine samples or bag collections of urine in which there were no bacteria present.4 Creatinine concentration and osmolality were measured in most specimens. When venesection was indicated clinically (32 febrile and 22 afebrile children) the blood neutrophil count was measured.

Moderate pyuria (10-100x106 leucocytes/l) occurred in 30 (43%) febrile children but in only five (6%) afebrile children (P<0.001; odds ratio 12.3 (95% confidence interval 4.9 to 30.8)). Obvious pyuria (100-700x106 leucocytes/l) occurred in six (9%) febrile children but in none of the afebrile children (P<0.025). Median values of urinary osmolality were similar in febrile and afebrile groups (660 and 670 mmol/l/kg). The higher urinary leucocyte counts in febrile children were not due to increased concentration of urine but to a higher leucocyte excretion rate, estimated from the ratio of urinary leucocyte count to creatinine concentration (P<0.001). Among the febrile children there was no significant relation between the degree of fever and age, sex, blood neutrophil count, or whether they had presented with a convulsion (29 cases).

Comment

This study shows that pyuria is common in feverish children without a urinary tract infection and that it is due to an increase in the urinary leucocyte excretion rate and is not a direct reflection of a raised blood neutrophil count. This suggests that pyuria may be a non-specific feature of fever in acute childhood illness and may reflect a generalised increase in white cell migration, perhaps mediated by changes in membrane permeability or white cell motility.

The fact that almost 9% of febrile children without a urinary tract infection have obvious pyuria is important, particularly now that pyuria may be readily identified by dipstick testing and is usually assumed to be diagnostic of a urinary tract infection. However, pyuria in a febrile child does not always indicate urinary tract infection; the false assumption that it does may lead to inappropriate antibiotic treatment and unnecessary investigations of the renal tract. Possibly more importantly, this assumption may also result in failure to consider other diagnoses in a feverish unwell child.

Footnotes

  • Funding None.

  • Conflict of interest None.

References

View Abstract