Contamination of urine specimens by vaginal or labial bacteria is a serious problem. The most reliable way to obtain the specimen, especially in young children, is by suprapubic puncture and aspiration of the bladder with a needle and syringe. However, this technique has never become popular. Catheterization is another way to solve the difficulty, but approximately 2% of patients are reported to develop urinary tract infection following a single catheterization. A midstream “clean-catch” voided specimen is the third best procedure. The urinary opening is cleansed. The labia are held apart in the female and the foreskin pulled back in the male. The first portion of urine is allowed to pass uncollected, then the sterile container is quickly introduced into the urine stream to catch the specimen.

Quantitative urine cultures have now replaced simple cultures, since it has been generally accepted that a titer of 100,000 (105) or more organisms/ml has excellent correlation with clinical infection, whereas a titer of less than 1,000 (103) organisms/ ml is usually not significant. However, there are certain important reservations and exceptions to this commonly accepted statement. The original studies that established this concept were performed on nonsymptomatic women using clean-catch urine collection. In addition, it was found that a single urine specimen collected this way that contained more than 100,000 organisms/ml had only an 80% chance of representing true infection (i.e., there was a 20% chance that the specimen was contaminated). Two consecutive urine specimens containing more than 100,000 organisms/ml has a 95% chance of indicating true infection. If the patient is receiving antibiotics or if the specimen is obtained by catheterization, many authorities accept a colony count of 1,000 or more organisms/ml as significant. If the specimen is obtained by suprapubic puncture or by cystoscopy, any growth at all is considered significant. If the patient has symptoms of urinary tract infection (fever, chills, flank pain, and pyuria), many authorities are willing to accept a clean-catch specimen colony count of 1,000 or more organisms/ml as significant even if the patient is not taking antibiotics. Not all patients with symptoms of urinary tract infection have colony counts over 1,000 organisms, because some females who have dysuria and frequency have urethritis (acute urethral syndrome, discussed earlier) rather than cystitis. About 20%-30% of these cases of urethritis are due to Chlamydia trachomatis, which does not grow on standard urine culture media. There is also the possibility that some cases could be chlamydial in origin but that the urine culture becomes contaminated with other bacteria, which leads to a false impression of cystitis. Finally, in some cases of pyelonephritis there are few or intermittent symptoms, and the urine cultures may be consistently negative, intermittently positive, or contain relatively low numbers of organisms. If there is a strong suspicion of pyelonephritis, repeat cultures may be necessary if the initial culture is negative or equivocal.

Use of the quantitative (titer) technique with a cutoff level of 100,000 organisms/ml is supposed to compensate for small degrees of unavoidable contamination, but contamination can easily be severe enough to give a positive result. Various studies using midstream specimens have shown a surprisingly high rate of contamination in the female. In my experience, a microscopic examination of centrifuged specimen sediment in which more than 10 squamous epithelial cells per low-power field (10 x ocular, 10 x objective) are demonstrated suggests possible contamination. Three or more species of bacteria isolated from the same urine specimen also suggests contamination if the patient is not on long-term indwelling catheter drainage. All available information emphasizes that good technique in collecting the specimen is essential. Also essential is getting the specimen to the laboratory as soon as possible after collection. Urine is an excellent culture medium, and specimens that stand for more than 1 hour at room temperature permit bacterial incubation and proliferation to the point that quantitative counts are not reliable. If delivery to the laboratory must be delayed, the specimen should be refrigerated. The specimen can be preserved up to 12 hours in a refrigerator (4°C).

Quantitative urine culture has two other drawbacks. Culture involves trained technical personnel and relatively expensive media, thus curtailing use for mass screening to detect urinary tract infection. Culture takes 24 hours to determine bacterial quantity and another 24-48 hours to identify the organism, thus delaying treatment in suspected infection.

Screening tests for bacteriuria

Because of the delay inherent in culture methods, several screening tests for rapid detection of significant degrees of bacteriuria have been introduced. If organisms are seen on Gram stain of uncentrifuged urine, reports indicate about 93% (range, 88%-95%) probability of a positive (100,000 organisms/ml) quantitative culture. The same is said to be true of direct examination of unstained centrifuged urine sediment. Gram-stained centrifuged sediment apparently gives too many false positive results. Gram-stained uncentrifuged urine and unstained urine sediment examination are said to give relatively few false positive results. Microscopic examination of centrifuged urine sediment to detect pyuria (increased number of WBCs) is widely used to screen for urinary tract infection. However, pyuria is found in only about 70%-80% (range, 13%-93%) of patients with positive urine cultures, and about 30% (range, 2%-40%) of patients with pyuria do not have a positive urine culture. For example, in one group of infants with culture-proven urinary tract infection, 42% had normal urine sediment and no bacteria seen on Gram-stained smears of noncentrifuged urine. Leukocyte esterase dipsticks have been used as a substitute for microscopic examination. They are positive in about 80%-90% (range, 72%-98%) of patients with a positive culture. Some urine test dipsticks have both the leukocyte esterase and a biochemical nitrite test; if results of both are positive, this detects 85%-90% (range, 78%-100%) of positive results from culture. Several chemical methods have been advocated. The most successful include the triphenyl tetrazolium chloride (TTC) test, the Greiss nitrate test (previously mentioned), and the catalase test. Best results have been reported with the TTC test. Nevertheless, reports vary as to its accuracy, with a range of 70%-90% correlation with quantitative culture, including both false positive and false negative results. The Greiss nitrate test detects about 55% (range, 35%-69%) of cases with a positive culture. The catalase test is even less reliable.

Finally, there are several other bacteriuria screening systems now available, and more keep appearing. The Marion Laboratories Bac-T-Screen filters the urine, trapping any bacteria on the filter, which is then stained to reveal presence of bacteria. Sensitivity (compared to culture) is reported to be about 93% (range, 88%-99%). However, false positive results are about 29% (range, 16%-55%). Two companies market tests using firefly luciferase (bioluminescence) that reacts with bacterial ATP. Three companies market automated screening instruments, two detecting turbidity produced by bacterial growth (Autobac and MS-2 systems) and one (AMS systems) scanning a multiwell card containing various test media. Results from all of these systems are similar to those of the Bac-T-Screen. The advantages of these systems are that negative test results suggest that culture would not be necessary. Results are available the same day. Drawbacks are that positive test results by any of these methods must be followed by quantitative culture (except possibly the AMS system). False positive test results are relatively frequent, and negative test results do not completely rule out urinary tract infection (especially since the sensitivity figures quoted earlier are for 100,000 organisms/ml, and detection rates become less if bacterial numbers are less).

Drawbacks of urine quantitative culture

In summary, although quantitative culture with a cutoff titer value of 100,000 organisms/ml is widely considered the most reliable index of urinary tract infection, there are several major limitations:

1. Using a titer of 100,000 organisms/ml as the cutoff point for a positive culture decision eliminates specimens with low-grade contamination but does not identify those with greater degrees of contamination.
2. The cutoff point of 100,000 organisms/ml does not apply to catheterized or suprapubic aspiration specimens and may be misleading if used to interpret them.
3. There are circumstances in which fewer than 100,000 organisms/ml may be significant in clean-catch specimens.
4. A positive culture does not reveal which area of the urinary tract is involved.
5. Tuberculosis of the urinary tract, anaerobic infections, and chlamydial infections of the urethra give negative culture results on ordinary culture media.
6. Bacteremia from many causes, even if transient, involves organisms that are filtered by the kidney and may produce a temporarily positive quantitative culture result.
7. Some cases of urinary tract infection may give negative cultures at various times. In several studies, a sizable minority of patients required repeated cultures before one became positive. This may be due to the location of the infection in the kidney and to its degree of activity.

Therefore, two cautions are required when one is interpreting a urine culture report. If the results suggest that urinary tract infection is present, the physician must be certain that the specimen was properly collected, especially in the female. If the result is negative, this does not rule out chronic pyelonephritis. The problem of pyelonephritis and bacteriuria is also discussed in Chapter 12.

Kits with agar-coated slides that not only yield quantitative urine culture results but also act as a culture medium for the organisms are now available. These are reported to have 90%-98% accuracy in various university hospital laboratories when compared with standard quantitative culture methods. The average is about 95% reliability. There is some question whether office laboratories can approach this figure and what provision would be made to identify the organisms.

It should be mentioned that a few studies have reported little or no difference in contamination rates between clean-catch cultures and no-precaution cultures.