Saturday 28 September

ORIGINAL

No 8770

1991

ARTICLES

Diagnosis of urinary tract infection in children: fresh urine microscopy or culture?

Fresh unspun and unstained urine specimens from 342 children with previous urinary tract infection (UTI) or symptoms compatible with a UTI were examined by microscopy at a magnification of × 400 in a mirrored counting chamber by a clinician, and sent for culture in a microbiology laboratory; 200 samples were also plated onto dip-slides. When microscopy and culture results were discrepant, further urine samples were collected until a diagnosis of UTI (24) or sterile urine (318) could be confirmed. Initial microscopy correctly identified 23 of 24 UTIs and 286 of 318 sterile urines; 1 false-positive result was caused by vaginal contamination with lactobacilli. 32 specimens (9%) gave an equivocal result on microscopy; the 1 other true-positive result was identified correctly on microscopy of the next urine specimen obtained. Culture of the initial urines correctly identified all 24 UTIs, but only 82% of the negative samples. Of the samples from uninfected children, 35 (11%) showed a mixed growth which was sterile on repeat sampling, and 21 (6·6%) initially grew a false-positive pure growth of more than 105 colony-forming units/ml of one organism. True UTIs were associated with bacterial counts above 107/ml. Microscopy by a clinician represents a cheaper, quicker, and more reliable screening test for UTI in children than does routine culture in a

microbiology laboratory. Introduction

Urinary tract infection (UTI) in a child requires prompt because it may rapidly lead to a serious,

treatment

debilitating illness, especially in an infant, and because even fairly brief infection may cause permanent renal parenchymal scars in susceptible children.’ Prompt treatment depends on rapid diagnosis. However, UTI is usually confirmed by culture of a pure growth of a single bacterial pathogen at a viable count of more than 105 colony forming units (CFU) per ml urine,2-7 a process with inherent delays. Apart from the day taken to incubate the bacteria, many microbiology laboratories do not routinely issue an interim report but wait a further day for the results of subcultures for sensitivity testing. In practice, clinicians often do not receive confirmation that a child has a UTI until a written report is delivered 3 or more days after the urine was collected. If the result is ambiguous, such as a

heavy

mixed

growth, diagnosis

and

treatment may

be

further delayed while a repeat culture is awaited, or a confident diagnosis may be impossible if the patient was treated with an antibiotic after the specimen was taken. In such circumstances, a clinician may feel that further investigation and follow-up are necessary, despite the lack of clear initial evidence of infection. Collection of several samples for culture before antibiotic treatment is started would reduce this risk, but would be expensive, delay treatment, and seldom be practicable outside hospital. For several years, urine specimens from children seen as inpatients and outpatients in our paediatric renal unit have been examined by light microscopy as well as sent for

laboratory culture. Although various workers have suggested that microscopy of unstained urine may miss infection, 4-7 our experience indicated that all children who had a UTI were readily identified. Occasionally we saw no Department of Child Health, Royal Victoria Newcastle upon Tyne NE1 4LP, UK (D. Vickers, MRCP, T. Ahmad, MRCP, M. G. Coulthard, FRCP). Correspondence to Dr M G. Coulthard. ADDRESS.

Infirmary,

768

TABLE I-INITIAL RESULTS BY MICROSCOPY AND URINE CULTURE

bacteria on urine microscopy, but received laboratory reports of "significant" bacterial infection in samples from children in whom subsequent investigations or clinical course made the diagnosis of a UTI appear unlikely. We therefore set up a prospective study to test the hypothesis that microscopy of fresh unspun and unstained urine can distinguish between sterile and infected urine as well as, or better than, routine culture.

I

*Inltlal culture sterile, lactobacilli

a

dip-slide.

Each urine specimen was examined 3 times at x 400 magnification by a microscopist who was unaware of the patient’s clinical details. The first examination was with a simple and inexpensive monocular microscope with an external light source (Open University, Milton Keynes, UK); the second examination was made on a more expensive binocular microscope with a teaching attachment (’Labophot’, Nikon, Tokyo, Japan), used as a standard light microscope, and the third on the same instrument under phase contrast. The result of each examination stage was recorded before the next was started. A Neubaur counting chamber with a mirrored surface (’Crystalite’, Hawksley, Lancing, West Sussex, UK) was used in all examinations. A clean cover slip was firmly applied to the slide surface to leave a space 0-1mm deep above the counting chamber surface, into which urine was drawn by capillary action from a reagent test strip. Each sample was examined for red and white blood cells, epithelial cells, crystals, debris, and bacteria; morphology was noted and numbers counted per ml. With the simple microscope only one high-power field (about 0.008 nl) was examined; with the second microscope 0 1 lil urine was rapidly scanned by use of a stage mover. White cells settled quickly onto the grid surface, and those at other depths in the field could be readily identified and counted with adjustments of focus. However, for specimens that contained many bacteria only about 10% of organisms could be seen at any one time; 90% were out of the plane of focus of the microscope. To allow for this observation the slide was scanned whilst the grid was kept in clear focus, and the number of bacteria seen was multiplied by a factor of 10 to give the number of organisms present in the sample. Microscopy results were defmed as: positive if the sample contained at least 107 bacteria of the same morphology per ml (equivalent to 8 organisms per high-power field or 100 organisms/grid inspected [with the grid in the plane of focus]); negative if no bacteria were seen; and uncertain for any other result (eg, bacteria of different morphology, bacteria of the same morphology but less than 107/rnl, or presence of large numbers of epithelial cells or debris to indicate contamination). In the microbiology laboratory urine was plated onto MacConkey and blood agar, and the interval between the collection and plating noted. Dip-slides (’Uricult’, Orion Diagnostica, Espoo, Finland) were also used in the last 200 urines collected, inoculated by the clinician at the time of urine microscopy and then incubated in the laboratory. Culture results were defined as: positive if there were more than 105 CFU/ml of one species of organism; negative if there was no growth or below 104 CFU/ml; and uncertain if growth of more than one species of organism (mixed growth) was observed. A child was considered to have a UTI when urine microscopy and culture results were both positive, and to have uninfected urine when the culture result was negative irrespective of the microscopic

I

anaerobic culture

TABLE II-COMPARISON OF INITIAL MICROSCOPY AND FINAL DIAGNOSIS

Patients and methods Urine samples were obtained consecutively from 357 children, aged 6 weeks to 18 years, who attended an outpatient clinic (many referred because of a previous diagnosis of UTI) or who were admitted to hospital with a wide range of diagnoses and symptoms and signs that included unexplained pyrexia, abdominal pain, haematuria, malaise, and weight loss; a few had known chronic renal failure. Samples were collected as midstream specimens or by use of sterile urine bags in younger children. In girls the vulva was washed with benzyl alcohol (09% in water) on cottonwool balls; urine was collected from boys without washing. Each specimen was divided immediately after collection: part was reserved for microscopy within 1 h, part sent to the laboratory for routine culture, and, for 200 specimens, a third aliquot was used to inoculate

I

on

*Inrtial culture sterile, lactobacilli

on

anaerobic culture

findings. If the initial urine specimen showed uncertain results on microscopy or culture a repeat specimen was taken as soon as possible; for uncertain results on microscopy the second specimen was usually the next sample passed, but there were obviously longer delays when the culture result was uncertain. Urine samples were obtained repeatedly until a specimen fulfilled criteria for a UTI or no

infection.

Results The findings of the three microscopic examinations of each urine specimen were identical regardless of the type of instrument, mode of use, or whether 1 high-power field or the whole grid was examined. Bacteria could readily be seen and their morphology defined by standard light microscopy and a mirrored slide, but phase contrast made the task more easy. In all but 2 of the last 200 specimens the culture results from samples plated in the laboratory and from the dip-slides were the same; 22 positive, 154 negative, and 22 mixed growths; the 2 exceptions both showed mixed growths on dip-slides and were sterile on laboratory

plates. 15 of the 357 children were excluded from analysis. Urine samples never reached the microbiology laboratory in 9; another 5 were from children whose initial urine microscopy and culture results were discrepant but who either received antibiotic treatment before they could be repeated (3) or lived too far from the hospital for delivery of a repeat sample to our unit to be practicable (2); and in 1 patient a candida UTI was diagnosed by both microscopy and culture, but no bacterial infection was present. For the other 342 specimens, initial microscopy findings are compared to initial culture findings in table I and to final diagnosis of a UTI or uninfected urine in table n, initial culture findings are similarly compared to final diagnosis in table ill. Only 24 (7%) of the urine specimens were obtained from children with a UTI; of these 23 were positive both on initial culture and microscopy, with 107-109 bacteria/ml seen on microscopy. The other urine from a child with a UTI also had bacteria of similar morphology on TABLE III-COMPARISON OF INITIAL URINE CULTURE AND D FINAL DIAGNOSIS

769

Fig 1-Summary of results by microscopy and culture.

considered doubtful because only 0-8 x107 bacteria/ml were seen; the next urine voided was positive on microscopy with > 107 bacteria/ml. Of 318 urine specimens fmally considered uninfected, 244 (77%) were negative on initial microscopy and culture. 42 (13%) were negative on initial microscopy but grew either a mixed growth (29) or a pure growth ( > 105 CFU/ml) of one organism (13) on initial culture; none of these children received antibiotics, and repeat urine samples obtained after

microscopy, but

was

scrupulous cleaning were negative both on microscopy and culture. (Where culture was positive, yet no organisms were seen on microscopy, it is possible that the cultures contained just over 105 organisms/ml; at such a count only 1 organism would be likely to be seen over the entire counting chamber when scanned with the grid kept in clear focus.) Similarly, 14 (4%) specimens doubtful on initial microscopy grew either a mixed growth (6) or a pure growth (8), but the next

samples passed were sterile. 17 samples (5%) were doubtful on initial microscopy and sterile on culture. In 1 sample from a teenage girl many bacilli were seen but the culture was negative; repeat specimens all showed the same pattern and lactobacilli were found on subsequent anaerobic culture, presumably from vaginal contamination.

tVHL.tU&ULty

Fig 2-Comparison microscopy.

of

leucocyte

counts

on

fresh and

laboratory

Excludes samples for which both fresh and laboratory leucocyte samples for which the laboratory

counts were below 50 000/ml, and result was not quantified

Fig 1 shows the efficacy of microscopy or culture alone of a single urine sample for diagnosis of a UTI or no infection in the children studied. Microscopy for bacteria had a positive predictive value of 96% and a negative predictive value of 100% in the 91 % of specimens that gave a clear initial result. 9% of initial microscopic examinations gave an uncertain result, mostly indicative of contamination, and could be repeated almost at once; 97% were then negative. Culture had a positive predictive value of only 53%; 21 samples produced a pure growth of > 105 CFU/ml which was not associated with a positive microscopic result and was not confirmed in repeat samples taken without antibiotic treatment (for 8 of these patients the repeat sample was the next urine voided). The organisms cultured in these false-positive cases were similar species to those seen in true UTIs-

By definition, culture alone had a negative predictive value of 100%. 10% of initial cultures were mixed growths; repeat samples were all negative. All but 1 of the 24 children with

UTI had more than 50 000 leucocytes/ml (> 50/mm3) on fresh microscopy; several children without a UTI also had more than 50 000 leucocytes/ml, all of whom had pyrexia. Many samples with high leucocyte counts on fresh microscopy still had cells noted when they were examined in the microbiology laboratory, but there were less in almost all specimens and occasionally no white blood cells could be seen (fig 2). a

Discussion

Microscopy of urine widely practised, and

for unstained bacteria is no longer with standard light microscopy

bacteria appear as small bright shapes against a very bright background. A slide with a mirrored surface makes identification of bacteria much easier, because of the reduction of background glare, and a counting chamber helps to provide a clear plane (the grid) on which to focus. Even with a cheap microscope it was possible correctly to identify rods or chains of bacilli (and a yeast in 1 specimen), and phase contrast made identification straightforward and easy to teach. High counts of white blood cells in urine specimens may help to confirm a diagnosis of UTI, but we and others have observed that leucocytes may occur in large numbers in uninfected urine, and that leucocyte counts may fall with time.9 Absence of leucocytes cannot be used to exclude UTI, especially when there is a long delay before the sample reaches the laboratory or if the urine pH is high 10 (eg, in association with proteus infections). The definition of positive microscopy at a bacterial count of 107jml was made for practical reasons: this count is

equivalent to 8 organisms per high-power field, or 100 per grid, and can be readily identified even if only 1 high-power field is examined with the microscope focused on the counting chamber. A small number of organisms were identified on microscopy of some urine specimens from uninfected children who showed mixed growths or just over 105 CFU/ml of one species on culture, but not in others. There was no significant increase in the number of organisms cultured from the urine because of multiplication during transfer to the laboratory; culture results from dip-slides inoculated at the time of initial microscopy were

770

Fig 3-Suggested protocol

for

investigation of possible

UTI.

virtually identical to those processed in the laboratory. Although an arbitrary figure of 105 CFU/ml of a single organism is widely used to define a positive culture result, the actual count of bacteria in urine can only be measured accurately by predilution of the sample. Our data indicate that children with a true UTI usually have many more than 105 organisms/ml: we counted as many as 109/ml. Our findings indicate that microscopy of unstained, unspun fresh urine can be used to determine whether a child has a UTI or uninfected urine. In 310 (91%) of 342 specimens, microscopy of the initial specimen gave a clear diagnosis of no infection (with 100% predictive value) or of UTI (with 96% predictive value, because contamination with vaginal lactobacilli gave a false-positive result in 1 sample). In the other 9%, where initial microscopy was inconclusive, a clear answer could be obtained with minimal delay when the child next passed urine. Urine culture alone produced mixed growths in 11 % of children who had uninfected urine. Culture should be repeated for heavy mixed growths in case one of the organisms grown represents a true UTI, but in our series repeat samples were all negative; in such circumstances the delay between sample collection and the culture result often means that children have to be recalled from home to provide the repeat sample. More disturbingly, 6-6% of children with uninfected urine would have a false diagnosis of a UTI: despite the culture results, a true UTI seems highly unlikely in the absence of excess white blood cells on microscopy, negative culture results on repeat samples, and clinical course. If contamination of a urine sample with several organisms can occur it seems reasonable to assume that, on occasion, contamination can also occur with a single species. Other workers have described positive urine cultures when microscopy was negative, and have concluded that urine microscopy misses some genuine urinary tract infections;5 could some of these urine samples have been contaminated by small numbers of one species of organism? Although urine microscopy could quite easily miss a UTI if the count is only just over 105/ml, our experience indicates that bacterial counts in children with genuine infection are usually much higher. Although there has been recent support for the view that urine specimens from boys with low colony counts might indicate urinary infection," this observation is based on a report that boys might occasionally harbour small numbers of urethral organisms between frank UTIs.12 Urine microscopy is an inexpensive, straightforward, easily taught, reproducible, and rapid investigation which can readily be combined with routine clinical urine analysis.

We propose more widespread use of microscopy in the examination of urine specimens from children with a suspected UTI according to the protocol in fig 3. Fresh urine should first be examined with a mirrored counting chamber and light microscope, or under phase contrast. All negative samples can be discarded; if microscopic findings are equivocal, microscopy should be repeated on the next specimen obtained; and if microscopy is positive, the rest of the urine specimen should be sent for culture and analysis of antibiotic sensitivities and antibiotic treatment started immediately (direct sensitivity plating can be requested at the same time as there is already evidence of infection, so the correct antibiotic treatment can be confirmed after 1 day). The extra laboratory workload from direct sensitivity testing of microscopy-positive urines would be far outweighed by disposal of the microscopy-negative samples. Such a scheme need not be restricted to hospital paediatrics; the practical advantages might be even more apparent in general practice, and positive urines could be inoculated into dip-slides for later laboratory inspection. With current laboratory opening hours, it is our experience that many children who present to general practitioners with a possible UTI at a weekend are treated with antibiotics without urine examination; many are referred for further investigations, perhaps unnecessarily. Incorrect diagnosis of a UTI after a false-positive urine culture is even more likely to lead to such treatment and investigation. Similarly, false-negative culture could have unfortunate results. Since this study we have seen a boy with 108 bacteria/ml (chains of streptococci) and 400 000 leucocytes/ml on urine microscopy, but a urine culture reported as sterile. A repeat sample the following day gave identical results, but a microaerophilic enterococcus was subsequently cultured under anaerobic conditions. A routine culture request would have produced a false-negative result in this boy; it is difficult to determine the frequency of such results. We suggest that urine culture should be replaced by microscopy as a screening test for urinary tract infection in children. REFERENCES 1.

Ransley PG, Risdon RA. Reflux nephropathy: effects of antimicrobial therapy on the evolution of the early pyelonephritic scar. Kidney Int 1981; 20: 733-42.

2. Kass EH. Bacteriuria and the diagnosis of infections of the urinary tract with observations on the use of methionine as a urinary antiseptic. Arch Intern Med 1957; 100: 709-14. 3. Kass EH. Asymptomatic infections of the urinary tract. Trans Assoc Am Physicians 1956; 69: 56-63. 4. Littlewood JM, Jacobs SI, Ramsden CH. Comparison between microscopical examination of unstained deposits of urine and quantitative culture. Arch Dis Child 1977; 52: 894-96. 5. Robins DG, White RHR, Rogers KB, Osman MS. Urine microscopy as an aid to detection of bacteriuria. Lancet 1975; i: 476-78. 6. Flanagan PG, Rooney PG, Davies EA, Stout RW. Evaluation of four screening tests for bacteriuria in elderly people. Lancet 1989; i: 1117-19. 7. Pylkkanen J, Vilska J, Koskimies O. Diagnostic value of symptoms and clean-voided urine specimens in childhood urinary tract infection. Acta Paediatr Scand 1979; 68: 341-44. 8. Stansfeld JM, Webb JKG. Observations on pyuria in children. Arch Dis Child 1953; 28: 386-91. 9. Kierkegaard H, Feldt-Rasmussen U, Horder M, Anderson HJ, Jorgensen JP. Falsely negative urinary leucocyte counts due to delayed examination. Scand J Clin Lab Invest 1980; 40: 259-61. 10. Stansfeld JM. The measurement and meaning of pyuria. Arch Dis Child 1962; 37: 257-62. 11. Report of a working group of the Research Unit, Royal College of Physicians. Guidelines for the management of acute urinary tract infection in childhood. J R Coil Physicians Lond 1991; 25: 36-42. 12. Hallett RJ, Pead L, Maskell R. Urinary infection in boys: a three-year prospective study. Lancet 1976; ii: 1107-10.

Diagnosis of urinary tract infection in children: fresh urine microscopy or culture?

Fresh unspun and unstained urine specimens from 342 children with previous urinary tract infection (UTI) or symptoms compatible with a UTI were examin...
541KB Sizes 0 Downloads 0 Views