Further evaluation of the Griess test to detect significant bacteriuria. II

Utility of a routine urinalysis in children who require clean intermittent catheterization

BACKGROUND

Children who require clean intermittent catheterization (CIC) frequently have positive urine cultures. However, diagnosing a urinary tract infection (UTI) can be difficult, as there are no standardized criteria. Routine urinalysis (UA) has good predictive accuracy for UTI in the general pediatric population, but data are limited on the utility of routine UA in the population of children who require CIC.

OBJECTIVE

To determine the utility of UA parameters (e.g. leukocyte esterase, nitrites, and pyuria) to predict UTI in children who require CIC, and identify a composite UA that has maximal predictive accuracy for UTI.

METHODS

A cross-sectional study of 133 children who required CIC, and had a UA and urine culture sent as part of standard of care. Patients in the no-UTI group all had UA and urine cultures sent as part of routine urodynamics, and were asymptomatic. Patients included in the UTI group had growth of ≥50,000 colony-forming units/ml of a known uropathogen on urine culture, in addition to two or more of the following symptoms: fever, abdominal pain, back pain, foul-smelling urine, new or worse incontinence, and pain with catheterization. Categorical data were compared using Chi-squared test, and continuous data were compared with Student's t-test. Sensitivity, specificity, and positive and negative predictive values were calculated for individual UA parameters, as well as the composite UA. Logistic regression was performed on potential composite UA models to identify the model that best fit the data.

RESULTS

There was a higher proportion of patients in the no-UTI group with negative leukocyte esterase compared with the UTI group. There was a higher proportion of patients with UTI who had large leukocyte esterase and positive nitrites compared with the no-UTI group (Summary Figure). There was no between-group difference in urinary white blood cells. Positive nitrites were the most specific (84.4%) for UTI. None of the parameters had a high positive predictive value, while all had high negative predictive values. The composite model with the best Akaike information criterion was >10 urinary white blood cells and either moderate or large leukocyte esterase, which had a positive predictive value of 33.3 and a negative predictive value of 90.4.

CONCLUSION

Routine UA had limited sensitivity, but moderate specificity, in predicting UTI in children who required CIC. The composite UA and moderate or large leukocyte esterase both had good negative predictive values for the outcome of UTI.

Best practice in primary care pathology: review 2

This second best practice review examines five series of common primary care questions in laboratory medicine: (1) laboratory testing for allergy, (2) diagnosis and monitoring of menopause, (3) the use of urine cytology, (4) the usefulness of the erythrocyte sedimentation rate, and (5) the investigation of possible urinary tract infection. The review is presented in a question-answer format. The recommendations represent a précis of guidance found using a standardised literature search of national and international guidance notes, consensus statements, health policy documents, and evidence based medicine reviews, supplemented by MEDLINE EMBASE searches to identify relevant primary research documents. They are standards but form a guide to be set in the clinical context. Most are consensus rather than evidence based. They will be updated periodically to take account of new information.

The Griess test: an inexpensive screening test for asymptomatic bacteriuria in pregnancy

This study was designed to compare urine microscopy and the Griess test as screening tests for asymptomatic bacteriuria (AB) in pregnancy, using urine culture as a gold standard. Urine microscopy, Griess test and urine culture were carried out on urine samples from 483 antenatal women. AB was detected by urine culture in 30 (6.8%) of the patients. The Griess test and urine microscopy were found to have a sensitivity of 63.3% and 60%, specificity of 99.5% and 83.6%; and positive predictive value of 90.5% and 21.1% respectively. Most of the infections missed by the Griess test were either <100,000 colony count on culture or caused by gram positive cocci. The cost of the Griess test was only one tenth of that of urine microscopy. The Griess test is a simple and inexpensive test that when compared to urine microscopy has equal sensitivity, and better specificity and positive predictive value for the detection of AB in pregnancy than urine microscopy.

Detection of urinary tract infections by reduction of nitroblue tetrazolium

BACKGROUND

Nitroblue tetrazolium (NBT) reduction to formazan has been used as a marker for nitric oxide synthase (NOS). Since inducible NOS activity is elevated in urine from patients with urinary tract infections (UTIs), we investigated the accuracy of NBT reduction as an early predictor of UTIs and quantified the relationship between inducible NOS and NBT.

METHODS

Urine samples from 434 patients were screened for the presence of UTIs with leukocyte-esterase and nitrite dipsticks and with NBT reduction. The rapid screening results from each test were compared to urine culture results. In addition, NBT reduction parameters were measured in urine pellet at 595 nm after incubation with one of four factors: NOS cofactors, NOS inhibitors, NADH, or superoxide dismutase/catalase.

RESULTS

As a urine screening test for UTIs, NBT reduction was more sensitive with a higher negative predictive accuracy than the nitrite dipstick. NBT reduction also was more specific with a higher positive predictive accuracy and negative predictive accuracy than the leukocyte-esterase dipstick. In infected urine pellet, both NADPH, a NOS cofactor, and NADH increased NBT reduction. Superoxide dismutase/catalase decreased NBT reduction.

CONCLUSIONS

Although NOS may not be the only NBT reducing enzyme, rapid, visible reduction of NBT is induced in urine from patients with UTIs.

Nitric oxide synthase: an endogenous source of elevated nitrite in infected urine

To further define the endogenous sources of urine nitrite in urinary tract infections, we measured urinary nitrite levels by the Griess method and assayed urinary nitric oxide (NO) synthase activity by the conversion of 14C-arginine to 14C-citrulline. Endogenous production of 14C-citrulline was confirmed by thin layer chromatography. Exogenous L-arginine increased nitrite production in whole infected urine, but not in bacteria isolated from infected urine. Urinary tract infections significantly increased NO synthase activity in soluble urine fractions, although soluble activity was less than 10% of particulate activity. Urine particulate fractions from women with non-infected urine had greater NO synthase activity than particulate fractions from men with non-infected urine, 11 +/- 2 and 0.2 +/- 0.1 picomol/min/mg protein, respectively. Urinary tract infections increased NO synthase activity in urine particulate fractions from women and men, 99 +/- 20 and 48 +/- 9 picomol/min/mg protein, respectively. The conversion of 14C-arginine to 14C-citrulline required NADPH, was calcium independent, and was inhibited to a greater extent by L-canavanine than by NG-monomethyl-L-arginine or NG-nitro-L-arginine. Human infected urine contains an isoform of NO synthase which is an endogenous source of urine nitrite.

Periurethral colonization and urinary leukocytes as markers for bacteriuria in children with neurogenic bladder

Bacteriuria and associated renal damage is common in children with a neurogenic bladder, but the pathogenesis of urinary tract infection (UTI) is undefined. We examined the association between periurethral bacterial colonization and the presence of urinary leukocytes in 76 catheter urine specimens from children with neurogenic bladders. Although all the children were asymptomatic, 38/76 (50%) of the urine cultures were positive. Periurethral colonization was significantly more common with positive than with negative urine cultures, suggesting a pathogenetic role for periurethral bacteria in infection of the neurogenic bladder. Urinary leukocytes were present in 24/38 (63%) with positive cultures, as against none (0/38) of those with negative urine cultures, and their presence represents a host response to bladder bacteriuria.

Accuracy of Reagent Strip Testing for Urinary Tract Infection in the Elderly

The accuracy of reagent strip testing for urinary tract infection (UTI) was assessed in 100 elderly patients (50 acute patients admitted to hospital and 50 attending the day hospital). Reagent strip sensitivities were: Acute patients – urinary nitrite 83%, blood 67%, protein 72% and leucocytes 72%, and day hospital patients — urinary nitrite 90%, blood 65%, protein 30% and leucocytes 60%. Urinary nitrite specificities were 100% for both groups of patients. Only 28% of patients with a UTI had specific symptoms of the infection; pyrexia and a raised WBC also proved poor indicators. Urinary nitrite was thus the most accurate immediate indicator of UTI.

Performance characteristics of urine dipsticks stored in open containers

Dip and read urinalysis is a laboratory test commonly performed by emergency physicians. Although the manufacturer states that the capped vials containing the dipsticks must be closed immediately after removal of a strip, this recommendation may not be followed in a busy emergency department. In a simple, two-part, blinded, and controlled trial the authors found that the reagents for determining leukocyte esterase, pH, protein, glucose, ketones, urobilinogen, bilirubin, and blood showed good reproducibility when fresh dipsticks were compared with dipsticks exposed to room temperature and humidity for 1 to 15 days. In contrast to this, the nitrite portion of the exposed dipsticks showed a rapid and cumulative loss of specificity over time. By the end of a week of exposure, one third of the nitrite tests gave false-positive readings. At the end of a second week, nearly three quarters gave false-positive readings for a specificity of only 28%. It is concluded that the nitrite reagent, in contrast to the other eight reagents on the Chemstrip-9 dipstick (Biodynamics, Indianapolis, IN), rapidly loses accuracy when stored in uncapped vials.

Detecting bacteriuria in a primary maternal and child health care programme

Urinary tract infection in pregnancy has not been adequately dealt with in developing countries, though its consequences are well recognised. This is primarily because of constraints on resources coupled with a lack of technological infrastructure. An evaluation of the Griess test for the mass screening of urinary tract infection among antenatal women was carried out prospectively using a case-control method. The Griess test was found to be a valid, reliable, and economical screening test for urinary tract infection which can be integrated into a primary maternal and child health care programme.

The urinalysis: a critical appraisal

The urinalysis, an inexpensive office test, is often performed unnecessarily. Improved chemical testing using reagent strips obviates the need for routine microscopy in many cases. More information is needed to make specific recommendations on the use of the routine urinalysis as a screening procedure.

Urinary nitrite in symptomatic and asymptomatic urinary infection

The dipstrip test for urinary nitrite is fairly unreliable in symptomatic urinary infections and only 104 (52%) of 200 symptomatic children with urinary infection attending an emergency department had a positive result. The test yielded positive results, however, in 83 of 100 outpatients with largely asymptomatic urinary infection attending a follow up clinic because of known predisposition to urinary infection. This difference was highly significant. The finding of urinary nitrite is highly specific for urinary infection and only 1% of 300 uninfected urine specimens gave a positive result. After addition of a broth culture of Escherichia coli to sterile urine incubation at 37 degrees C for four to six hours was required before the nitrite test yielded positive results. This suggests that frequency of micturition in urinary infection reduces the reliability of the nitrite test. On the other hand, the use of overnight, first morning urine specimens may further improve the sensitivity. If nitrite testing is used for screening for urinary infection at home, however, patients should be warned not to rely on a negative result in the presence of symptoms of urinary infection.

Urinary tract infection during pregnancy. Asymptomatic bacteriuria, acute cystitis, and acute pyelonephritis

Urinary tract infection is one of the most common medical complications of pregnancy, occurring in roughly 10% of all pregnancies. The clinical entities most commonly seen are bacteriuria, acute cystitis, and acute pyelonephritis. Relative stasis of urine due to pregnancy-induced changes in the urinary tract is a significant causative factor. Escherichia coli is the most common responsible organism. Patients with untreated or inadequately treated asymptomatic bacteriuria are at high risk for development of acute pyelonephritis, a serious febrile illness that can lead to major maternal and fetal complications. Patients with acute pyelonephritis should be hospitalized and treated aggressively.

The efficacy of the nitrite test and microscopic urinalysis in predicting urine culture results

The efficacy of the nitrite dipstick and microscopic urinalysis in predicting culture results were evaluated for 146 urine cultures from 56 women at risk for recurrent pyelonephritis. There were 111 negative cultures, 18 positive cultures, and 17 contaminated cultures. Nitrites were not detectable in either negative or contaminated cultures. Bacilluria was present in 10% of negative cultures and in 18% of contaminated cultures. Pyuria was present in 14% of negative and in 24% of contaminated cultures. Among the 18 positive cultures, nitrite was present in only 22%, bacilluria in only 61%, and pyuria in only 67%. These results suggest that positive cultures cannot be accurately predicted by either microscopic urinalysis or the nitrite dipstick, and that routine culturing should be performed as part of the outpatient management of women at risk for recurrent pyelonephritis.

Evaluation of new urinary tract infection screening devices

Several new methods for detection of bacteriuria were studied to evaluate their usefulness as screening procedures. A new filter paper device incorporating dehydrated media and tetrazolium was found to be reliable when compared with the standard pour plate method in the laboratory and with the dip-slide method in a field test. It failed to detect yeasts and slowly growing streptococci. Antibiotics blocked the test when susceptible organisms were present. An agar-cup method was found to be quite reliable, but could be improved by use of differential media. The Griess test was confirmed in a small trial to be highly specific when used in conjunction with a first morning specimen, but of little value with random specimens. Phenzopyridine was found to give false positive reactions. The subnormal glucose test, although highly sensitive and specific, gave too many false positive tests to be useful other than as a screening method.