Suprapubic aspiration under ultrasound guidance in children with fever of undiagnosed cause

BMJ 1994; 308 doi: (Published 12 March 1994) Cite this as: BMJ 1994;308:690
  1. H Buys,
  2. L Pead,
  3. R Hallett,
  4. R Maskell
  1. Department of Paediatrics, St Mary's Hospital, Portsmouth PO3 6AD Public Health Laboratory, St Mary's Hospital, Portsmouth PO3 6AQ
  1. Correspondence to: Dr Hallett.
  • Accepted 12 November 1993


Objectives: To assess the ease of use of suprapubic aspiration of urine under ultrasound guidance in babies with fever of uncertain cause and to assess the importance of bacterial counts and pyuria in relation to abnormalities of the urinary tract and the importance of pyuria in the absence of bacteriuria.

Design: Analysis of urine samples obtained by suprapubic aspiration in babies and children from July 1991 to June 1992. The clinical records of the children with bacteriuria and sterile pyuria were examined retrospectively.

Setting: Neonatal and paediatric wards of a district general hospital. Subjects - 508 babies and children who had fever of uncertain cause or were seriously ill.

Results: No difficulties arose in the collection of 545 specimens. Bacteria were isolated from the specimens of 44 children, 24 of whom had abnormalities of the urinary tract. The bacterial count was <108/1 in 18 of the children with bacteriuria, 10 of whom had abnormalities. No white cells were seen in 22 of the 46 bacteriuric specimens; nine of the children with no pyuria had vesicoureteric reflux. 439 of the 499 non -bacteriuric specimens showed no white cells. 60 children had pyuria without bacteriuria.

Conclusions: The use of ultrasound guidance simplifies suprapubic aspiration of urine in babies. Low bacterial counts may be associated with abnormalities of the urinary tract. Laboratory techniques capable of detecting such counts reliably should be used. Pyuria is absent in half of babies and very young children with bacteriuria. It rarely occurs without bacteriuria, and if it does an explanation should be sought.

Clinical implications

  • Clinical implications

  • Detecting urinary tract infection in early childhood is important for preventing renal damage

  • Suprapubic aspiration of urine is considered in Britain to be a difficult technique

  • This study showed that use of ultrasound guidance makes suprapubic aspiration in children and babies easy and safe

  • Low bacterial counts were present in the urine of two fifths of the children with abnormalities of the urinary tract

  • Clinicians should regard low bacterial counts as important


The association of urinary tract infection in early childhood with renal scarring, hypertension, and renal failure and the importance of early diagnosis to prevent renal damage are recognised.1 The diagnosis may be difficult and unreliable owing to the problem of obtaining satisfactory urine specimens from infants and babies. Suprapubic aspiration is accepted as the definitive sampling technique,2,3 but it is not widely used in Britain because inexperienced medical staff often find it difficult. No large series have been reported from Britain. When used under ultrasound guidance, however, suprapubic aspiration is safe and easy,4,5 and is now part of the standard investigation of babies with fever of uncertain cause in our hospital. We studied the findings in specimens collected in this way over one year and recorded the results of imaging in children with bacteriuria and sterile pyuria.


From July 1991 to June 1992 senior house officers used suprapubic aspiration under ultrasound guidance to collect urine specimens from children in the neonatal and paediatric wards. Most specimens came from infants and babies with non-specific febrile illnesses but a few came from older children who were seriously ill or whose previous urine specimens, collected by other methods, had not yielded a definite diagnosis. Suprapubic aspiration was uniformly successful, and no ill effects were seen.

The specimens were transported to the laboratory immediately and processed on arrival. Uncentrifuged urine was examined by the inverted microscope technique (x20 objective).6 Specimens were inoculated with a 0.005 ml standard loop on to cystine lactose electrolyte deficient agar and incubated overnight in an atmosphere containing 5% carbon dioxide. We defined pyuria, for the purpose of this study, as the presence of any white cells on microscopy. We classified any growth on culture, whatever the count, as bacteriuria; pyuria and no growth as sterile pyuria; and no growth and no pyuria as a negative result. We obtained the clinical diagnosis and results of imaging of the children with bacteriuria retrospectively from the case notes. We examined the records of the children with sterile pyuria to find out whether they had been receiving antibacterial treatment when the specimen was collected. We also recorded the results of any imaging of children in this group.


In all, 545 suprapubic aspirates were collected from 508 children during the year. Table I shows the age and sex of the children when the urine specimens were collected and correlates these data with the results of microscopy and culture. Table II shows the bacterial counts of the organisms isolated from 46 specimens from the 44 children with bacteriuria.


Distribution of bacteriuria and sterile pyuria in 545 urine specimens collected by suprapubic aspiration by sex and age of children from whom specimens were taken. Figures are numbers of specimens

View this table:

Results of culture of 46 urine specimens from 44 children with bacteriuria. Figures are numbers of specimens (numbers of specimens from children with pyuria in parentheses)

View this table:

Thirty eight of the 44 children with bacteriuria underwent imaging of the urinary tract; of the rest, three had surgery for acute abdomen due to non-renal causes, one died of an infected meningomyelocele, one recovered and was discharged with a diagnosis of neonatal jaundice, and the parents of one cancelled the investigations. Ultrasound scanning and micturating cystourethrography were performed in 35 children; table III shows the findings. Vesicoureteric reflux was classified according to Smellie's criteria.7 Micturating cystourethrography alone was performed in three children: none showed reflux but one had a bladder residue. Intravenous urography or radionuclide scanning was undertaken in six children to elucidate abnormalities detected with ultrasound scanning or micturating cystourethrography. In all, 24 children (16 boys), with bacteriuria had radiological abnormalities (table IV).


Findings of ultrasound scanning performed in 35 children with bacteriuria related to findings on micturating cystourethrography

View this table:

Age of 24 children with bacteriuria in whom radiological abnormalities were found

View this table:

Table V shows the bacterial counts in the 24 children with radiological abnormalities and how many of the children had pyuria; 10 (42%) had counts of <108/l and 9 (38%) had no pyuria. Of the 499 non- bacteriuric specimens, 439 (88%) showed no pyuria and the specimens from the remaining 60 children (42 boys) showed sterile pyuria. Forty one of the 60 children with sterile pyuria were neonates, of whom 17 were known to have been receiving antibiotics when the urine was collected. No definitive diagnosis was made in most of the neonates; many of them were premature and in incubators. Imaging of the urinary tract was undertaken in only 13 of the children with sterile pyuria and showed abnormalities in four: hydronephrosis (one), hydronephrosis and calculi (one), calculi (one), and bladder residue (one). Six of the 60 children with sterile pyuria either had evidence of previous bacteriuria or developed bacteriuria before the end of the study.


Results of culture of urine specimens from 24 children with bacteriuria in whom radiological abnormalities were found. Number of children with pyuria in parentheses

View this table:


The use of ultrasound scanning makes suprapubic aspiration of urine simple and has enabled us to incorporate the procedure into the routine investigation of babies with fever of undiagnosed cause. In one year bacteriuria was identified in this way in 44 children and sterile pyuria in 60. Urinary tract infection was excluded in the rest of the 508 children; this averted the need for unnecessary treatment with antibiotics, prolonged stays in hospital, and expensive and unnecessary investigations that may be prompted by false positive results in specimens collected by other methods.8 A study of the investigation of children in hospital with febrile convulsions found that only two urine specimens from 228 children were collected by suprapubic aspiration.9

About two thirds of the bacteriuric children who were investigated had abnormalities of the urinary tract; most of these were treated and managed according to accepted practice. Vesicoureteric reflux was detected in 14 children with normal findings on ultrasound scanning; as diagnosis at this early stage is accepted as the best way to prevent renal damage micturating cystourethrography should always be performed.

Low bacterial counts were found in the urine of two fifths of the children with abnormalities. Although other authors have accepted low counts in urine collected by suprapubic aspiration as evidence of bacteriuria, they did not state in how many children such counts were associated with abnormalities.2 Clearly, not only should clinicians accept low counts as important but laboratories should use methods that can detect such counts and laboratory staff who read and report results should understand the importance of low counts. Many laboratory methods,10,11 often promoted to save time and money, do not detect low bacterial counts reliably. When they are substituted for culture of urine from sick children12 important diagnoses may be missed.

Bacteriuria was detected in a slightly higher percentage of specimens from boys than girls (9% versus 8%). This agrees with Airede's findings in Nigerian children of twice as many abnormalities in boys as girls.3 Airede's finding is clearly related to the age group studied. Chiu et al,4 who studied children from birth to 14 years, found no significant difference between the sexes in the percentage with abnormalities but reported a male to female ratio of 3:1 for infections in children aged <2 years.

Our findings challenge two widely held beliefs about pyuria. Firstly, we found that children with bacteriuria that is clinically important do not necessarily have pyuria: about half of the bacteriuric specimens showed no pyuria even when a low criterion for the detection of white cells was used, and nine children with no pyuria had vesicoureteric reflux. Secondly, few of the non-bacteriuric specimens showed pyuria. All the children from whom these specimens came were being investigated for febrile illnesses. Thus the view that pyuria often occurs as a non-specific response to fever seems untenable. Two thirds of the children with sterile pyuria were neonates, of whom many were premature and many were taking antibiotics; the possibility of bloodborne urinary tract infection, known to occur in neonates,13 is strong. Radiological abnormalities were found in about a third of the children who were investigated.

We suggest that clinicians should always seek an explanation for sterile pyuria and should consider the possibility of urinary tract infection if sterile pyuria persists.

We thank Drs G M Lewis, M J Hardman, and E R Wozniak for permission to include their patients and Veronica Symes for typing the manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
View Abstract