Urine Good Hands: Diagnosing UTIs With Urine Cultures

Feb. 8, 2021

Urinary tract infections (UTIs) include infections restricted to the bladder (cystitis), which are extremely common in women and may cause pain with urination, as well as more serious infections that also involve the kidneys (pyelonephritis). If you’re a clinician, you’re probably familiar with the process of requesting urine samples in patients with UTI symptoms, and equally familiar with receiving and acting on the results. But what exactly happens to that urine, and the organisms that may grow from it, between the time it leaves the bladder and the time the report appears in the medical record?

From the Bladder To the Cup To the Bench

One of the most important variables in the process of culturing urine is the method of collection. The bladder itself is generally considered a sterile environment (although, as we will discuss later, that isn’t always the case), but the external genitalia are colonized by commensal bacteria that can contaminate urine samples and ultimately grow in culture. A suprapubic aspirate, in which a needle is inserted directly through thoroughly cleansed skin into the bladder, is the most effective way to avoid the risk of urogenital contamination, but this method is relatively invasive and rarely used. For infants, young children and others who are not able to urinate directly into a specimen container (for example, people who have a neurogenic bladder), urine can be collected using a Foley catheter, which is inserted through the urethra into the bladder; this method also limits contamination.

Foley catheter.
Foley catheter.
Source: Wikimedia.org

Older children and adults who are able to do so can simply provide a “voided” urine specimen: that is, they pee in a cup. The midstream clean-catch approach is recommended for voided urine specimens in order to decrease the likelihood of contamination. However, there is no way to entirely prevent the possibility of contamination, and recent evidence suggests that cleaning and using a mid-stream specimen may not actually reduce contamination at all. (By contrast, the much-reviled “bagged urine” collection method sometimes used in infants, in which urine is collected in a plastic bag taped to the perineal region, may not be as prone to clinically significant contamination as is generally assumed). As we will see, the relative likelihood of contamination with different specimen collection methods becomes important in the clinical interpretation of urine culture results.

Once a urine sample has been collected, it must be transported to the laboratory. Because bacterial quantity is an important factor in assessing the potential clinical significance of any organisms present in the sample, it is important to limit bacterial growth between the time of sample collection and plating for culture. If the urine sample is kept at room temperature, it should be plated within 2 hours of collection. The time between collection and plating can be extended to 24 hours if the sample is kept refrigerated or is transported in a container with boric acid as a preservative.

The Preview: Urinalysis

It usually takes about a day for bacteria from a urine sample to grow to a sufficient quantity that they can be detected and identified using standard clinical microbiology lab techniques, and consequently it also takes at least this long to determine that bacteria aren’t present in the culture. However, valuable information about the likelihood of a UTI can be obtained rapidly through urinalysis. White blood cells in the urine, which reflect the inflammation that is typical of infection, can be detected and quantified by urinalysis. The presence of ≥10 white blood cells per μL (or >5 per high-power field) is almost always seen in people with a UTI. A urinalysis can also test for the presence of nitrites, which are produced by gram-negative bacterial species that are able to reduce nitrates to nitrites; these species include Escherichia coli, the most common cause of UTI. A point-of-care urine dipstick can provide preliminary information on these tests within minutes, while a microscopic urinalysis provides more quantitative and sensitive results. Urinalysis results also provide information on other parameters in the urine, including pH and the presence of red blood cells, protein and other materials that may be indications of a variety of  kidney diseases unrelated to infection.

If the urinalysis from a person with UTI symptoms confirms a likely UTI, a doctor can start empiric antibacterial treatment based on the most likely causative organisms while waiting for the culture results to tailor therapy. On the other hand, a normal urinalysis suggests that a UTI is less likely to be the cause of symptoms. A number of diagnostic stewardship programs have evaluated the implementation of “reflex” urine culture protocols, in which a culture is performed only if the urinalysis is suggestive of UTI. Studies of such approaches indicate that they may be effective at safely reducing unnecessary antibiotic consumption. It’s also important to note that asymptomatic bacteriuria, or the presence of bacteria in the urine of a person who is not having UTI symptoms, does not require treatment in most cases (pregnant women are an exception), so urine cultures should not generally be obtained in people in the absence of UTI symptoms.

The Main Feature: Culture

Growth of a lactose-fermenting organism, such as Escherichia coli, on a urine biplate with MacConkey agar (L) and sheep’s blood agar (R).
Growth of a lactose-fermenting organism, such as Escherichia coli, on a urine biplate with MacConkey agar (L) and sheep’s blood agar (R).
Source: Thea Brennan-Krohn

Once the urine sample reaches the clinical microbiology lab, it is typically plated onto 2 types of media: a MacConkey agar plate, which inhibits growth of gram-positive bacteria and also allows some early predictions about the identity of gram-negative bacteria, and a blood agar plate, which permits growth of nearly all bacteria that cause UTIs. The great majority of UTIs are caused by gram-negative bacteria, most commonly E. coli, which grows as pink colonies on MacConkey agar due to its ability to ferment lactose. Other Enterobacterales, such as Klebsiella and Proteus species, can also cause UTI, as can a few types of gram-positive bacteria, including Enterococcus species and Staphylococcus saprophyticus. Urine biplates, in which each of the 2 types of agar fill half the plate, allow for more efficient plating. Plates are incubated at 35-37°C and examined at 20 hours and, if there is no growth at this point, may be incubated for an additional day and re-examined.

Calibrated inoculating loops, 10 µL on the left, 1 µL on the right.
Calibrated inoculating loops, 10 µL on the left, 1 µL on the right.
Source: Wikimedia.org

Urine cultures are plated quantitatively, using a calibrated inoculating loop that picks up either 1 or 10 µL of urine; when colonies grow on the agar, the number of colony-forming units per milliliter (CFU/mL) can be calculated by multiplying by 1000 or 100, respectively. Quantifying bacteria in urine cultures is essential, particularly for voided specimens because, as we noted above, contamination of urine samples with urogenital flora is common. The identity of any organisms that grow, the quantity in which they grow and the specimen type are all taken into account when interpreting the results of the culture. Clinical microbiology labs use detailed algorithms to determine which bacteria are reported to the clinician and how they are described in the report. The full algorithms for reporting are complex and vary to some extent among labs, but certain principles are common to all:

  • When only 1 or 2 types of bacteria grow and are present in large quantities (i.e., ≥10,000 CFU/mL), they are almost always identified at the species level and reported as such.
  • When 3 or more types of bacteria grow and no single one predominates (i.e., none is present at >100,000 CFU/mL), the results may be reported as “mixed bacterial flora.”
  • When bacteria are present in lower quantities (i.e., <10,000 CFU/mL), they may be reported in more detail if they are from specimens that are more likely to be sterile (e.g., catheterized urine) than from specimens that are more likely to be contaminated (e.g., voided urine).
  • Antimicrobial susceptibility testing is not usually performed on organisms that are rarely uropathogens (e.g., Streptococcus species).

The goal of these algorithms is to ensure that bacteria that are causing disease are accurately reported so that patients can be treated, while avoiding unnecessary reporting of bacteria that are very unlikely to be causing a UTI in order to avoid excessive antibiotic use. Of course, there may be certain clinical situations in which it could be appropriate to evaluate in more detail a mixed culture or a culture with an organism that isn’t usually a uropathogen (for example, if a patient has indwelling urinary tract hardware), so, as always, clinicians can call the microbiology lab to ask for more detail about exactly what grew on a particular plate or to request that additional workup be performed in specific cases.

The relative ease of obtaining a urine specimen and the rapid growth of most uropathogens in culture mean that UTI is often a seemingly straightforward diagnosis. However, interpreting cultures from a specimen that has to pass through the dense microbiota of the urogenital region before reaching the specimen container requires a great deal of careful work in the clinical microbiology lab, where medical laboratory professionals, using their experience in colony recognition in concert with detailed algorithms, must balance the need for a diagnosis with the risk of Too Much (clinically irrelevant) Information.

Author: Thea Brennan-Krohn

Thea Brennan-Krohn
Thea Brennan-Krohn is a diplomate of the American Board of Medical Microbiology at Beth Israel Deaconess Medical Center (BIDMC). She is an attending in Pediatric Infectious Diseases at Boston Children's Hospital and a postdoctoral fellow at Beth Israel Deaconess Medical Center,