Urine has certain defense mechanisms against infection. The most important inhibitory factors include a very high osmolality (i.e., high urea concentration) and a high organic acid concentration (i.e., low pH). Both of these reduce bacterial growth by inhibiting phagocytosis and decreasing the reactivity of complement. In general, anaerobic bacteria and other fastidious organisms that make up most of the urethral flora do not multiply in urine. However, urine usually supports growth of nonfastidious bacteria (21,22).

There is accumulating evidence that the antibacterial defense mechanisms of the vaginal walls and periurethral area are important in preventing the progression of microorganisms from the rectum to the bladder. Normally, this area is colonized by gram-positive bacteria, lactobacillus, and diphtheroids
(organisms that grow very poorly in urine and do not cause UTIs). A number of studies have shown that females with recurrent cystitis first colonize their vaginal introitus and periurethral area with enterobacteria before the onset of the symptoms of cystitis and then are at risk for infection until this colonization reverses to a normal situation (4,21,23). Acidity of vaginal secretions may contribute to vaginal resistance to coliform bacteria. In premenopausal females, the vaginal pH is usually near 4.0. This low acidic pH prohibits the growth of organisms such as E. coli but promotes the growth of the normally present organisms (e.g., lactobacillus) that will interfere with the growth of uropathogens (24,25). High vaginal pH appears to be associated with the growth of enterobacteria (26). Treatment of menopausal patients with intravaginal estrogen leads to reappearance of lactobacilli, decline in vaginal pH, decrease in growth of uropathogens, and reduction in the incidence of UTI (27,28).

Periodic voiding is one of the most important known bladder defense mechanisms. One study noted the introduction of 10 million bacteria into normal male bladders failed to establish infection because the organisms were rapidly cleared by voiding, diluting with fresh urine, and voiding again (29). Voiding displaces infected urine with sterile urine and flushes out bacteria attached to desquamated uroepithelial cells. In addition, a thin film of urine remains in the bladder after emptying and any bacteria present are removed by the mucosal cell production of organic acids.

Unfortunately, episodic voiding is not enough to prevent infection after sexual intercourse. In case-controlled studies, voiding patterns before and after sexual activity were not associated with recurrent UTIs (30).

The ability of an organism to bind to the epithelial cell has been shown to correlate with its ability to infect the urinary tract. The ascending loop of Henle secretes Tamm-Horsfall protein, which is a uromucoid, rich in mannose. This protein may inhibit bacterial adherence and trap bacteria in the urine, allowing them to be flushed from the urinary tract (31). Also, the presence of urinary immunoglobulin and the lining of the bladder with a glycosaminoglycan may be important factors in the blocking of bacterial adherence. The reduction of glycosaminoglycan probably plays a role in recurrent cystitis (32,33).

Adherence of microorganisms to mucosal cells is considered to be a prerequisite for colonization and infection (34). As previously mentioned, when these organisms enter the urethra and bladder in most women, they do not adhere and are easily washed away. In patients who are susceptible to UTIs, the organisms will quickly lock into the defective epithelial cells. The fecal flora is almost invariably the source of the infecting organisms. E. coli is the major pathogen, although S. epidermidis and Enterococcus, Klebsiella, and Proteus species can sometimes be identified (Fig. 10.3). The interaction of the mucosal and bacterial cells is probably dependent on both receptors on the mucosa and some type of attachment mechanism used by the bacteria. E. coli has been shown to possess surface organelles that mediate attachment to specific host receptors. These structures are called pili and can be present in large numbers on the microbial cell. Two types that appear to be important in urinary infections have been identified. Type I pili seek mannose as a receptor and are isolated from individuals with cystitis. They tend to bind with a low affinity, and their presence is not correlated highly with pathogenicity. Type II pili are man-nose-negative or “p pili” and adhere to the P blood group antigens. E. coli strains possessing p fimbriae are more virulent and more likely to cause pyelonephritis than strains without them (35, 36, 37).

Schaeffer et al (38) studied the adherence of E. coli to vaginal epithelial cells in control subjects and in women who had experienced at least three UTIs in the past year. They found adherence to be
greater in the study patients than in the controls. The vaginal cells of those receiving a sustained course of antimicrobial showed less adherence than the vaginal cells of patients who were not taking antibiotics. If the antibiotics were discontinued, adherence returned, and reinfection usually occurred (39). In another study, Schaeffer et al (40) noted that adherence tended to be higher during the early estrogen-dependent phase of the menstrual cycle.

Furthermore, women at high risk for recurrent UTIs may be more genetically prone to recurrent infection. Although the mechanism is not understood, women who bear human leukocyte antigen A3 subtype (HLA-A3) are more likely to have had recurrent UTIs than those who lack this antigen (41). Other work also suggests that women of blood group B or AB who are nonsecretors of blood group substances are at significantly higher risk for developing infections than are women of other blood groups (42). In addition, patients with Lewis blood group types who are considered secretors have a lower incidence of UTIs. The Lewis blood groups exists at two genetic loci: Le_a and Le_b. Secretors possess the Lewis “b” genetic locus (i.e., Le_(a+,b+) and Le_(ā,b+), whereas nonsecretors are ). Evidence suggests that bacteria are unable to adhere to the urothelial cell because of alterations to the uromucoid, which inhibits binding (42, 43, 44, 45, 46).

Thus, these genetic differences at the cellular level appear to influence bacterial adherence and make certain women more prone to UTIs.

In women, sexual intercourse appears to be a major determinant for bacterial entry into the bladder. Prospective studies have shown that many UTIs develop the day after sexual intercourse (47). Both the frequency and recency of sexual intercourse increase the risk for UTI. It has been shown that women who have engaged in sexual intercourse within the prior 48 hours have a risk for infection 60 times greater than women who have not (47). Sexual intercourse with a new partner within the past year is also another independent risk factor.

Infection appears to occur through inoculation of periurethral bacteria into the bladder during active intercourse. Women who have not colonized their vaginal and periurethral areas with coliform bacteria will have introduction of normal vaginal flora (e.g., lactobacillus, diphtheroid, or S. epidermidis), which will not produce infection and are rapidly cleared with voiding. However, in the colonized women, the pathogenic organisms, such as E. coli, will infect the bladder. There is evidence that women who have recurrent UTI have shorter distances from the urethral meatus and posterior fourchette to the anus (48). Another commonly overlooked factor is the use of diaphragms. A number of studies have confirmed that diaphragm users are at increased risk for UTI even after statistically controlling for sexual activity and history of previous UTI (49,50). The mechanism is unknown; it is believed that it may be related to urethral obstruction caused by the diaphragm (51,52). Also, diaphragm users have reduced vaginal colonization with lactobacillus, but coliforms are isolated three times more often than in women using other contraceptive methods (50). Additionally, the spermicidal agent nonoxynol-9 is an independent risk factor for UTIs. Spermicides reduce vaginal lactobacilli, allowing growth of uropathogens.

Diabetic patients are prone to develop neurogenic bladder dysfunction and severe vascular disease, both of which can predispose to UTIs. Other genetic problems that are commonly associated with UTIs are gouty nephropathy, sickle cell trait, and cystic renal disease.

It must be understood that the explanations mentioned for the pathogenesis of UTIs apply only to those females who have normal urinary tracts. Bacteria in the presence of obstructions, stones, or a neurogenic bladder do not need to have special invasive properties other than the ability to grow in urine.

Microscopic analysis of urine is an easy and valuable method of evaluating women with symptoms of UTI. A thorough microscopic examination of an uncentrifuged sample of urine showing minimal epithelial cells (thus not contaminated), bacteria moving in the field, and 2 to 6 leukocytes per highpower field correlates with above 10⁶ cfu/mL on culture. If infection with greater than 10⁴ cfu/mL is present, the finding of one or more bacteria on a Gram-stained specimen of urine correlates highly with the presence of UTI, having a sensitivity of 80% and a specificity of 90% with a positive predictive value of about 85% (60). Gram stain of the urine is useful in detecting abundant bacteriuria but is of little help in infection with colony counts of less than 10⁴ cfu/mL.

Fresh, unspun urine should also be quantitatively assessed with a hemocytometer for the number of white blood cells. The hemocytometer is positioned on the microscope stage. The number of leukocytes is counted in each of nine large squares, divided by 9 and multiplied by 10 to yield the number of white blood cells per milliliter. Pyuria is defined as greater than 10 leukocytes/mL. Pyuria is present in nearly all women with acute UTI. Studies note the presence of pyuria to be 80% to 95% sensitive (even when bacteria counts are less than 10⁴) and 50% to 75% specific for the presence of UTI. However, a study of pregnant patients presenting acutely to a labor ward showed that only 17% of patients with significant pyuria had a significant urine culture (61). It is also of value to ascertain whether red blood cells are present or to perform a urine dipstick for blood. Microscopic hematuria can be found in about 50% of women with acute UTI and is rarely present in patients who have dysuria from other causes (62,63).

If expertise for office microscopy is not available or feasible, it is reasonable to substitute a rapid diagnostic test for bacteriuria, pyuria, and hematuria, although in general these lead to less accurate results than microscopy.

The most common rapid detection test is the nitrite test. Certain bacteria such as Proteus species and occasionally E. coli have the enzyme nitrate reductase, which converts dietary nitrates into nitrite; this, in turn, causes the amine-impregnated dipstick to turn pink within 60 seconds of reaction. Numerous commercial urine dip tests are available, and one should check the sensitivity and specificity of the particular test kit used. Generally, a positive nitrite test is highly specific (92% to 100%) for UTI and deserves treatment. However, the test is not sensitive and is not a good screening tool (i.e., only 25% of patients with UTI test positive for nitrites). Lack of dietary nitrate, organisms that lack nitrate reductase, and diuretics can lead to false-negative results.

The nitrite test is often integrated with a test for leukocyte esterase (LE), which is an enzyme found in primary neutrophil granules. When LE reacts with reagents impregnated in the dipstick, a blue color is produced within 1 to 2 minutes, indicating a positive test. The LE test has a specificity of 94% to 98% (64). However, the sensitivity of the LE test is directly related to the bacterial load. Wu et al (65) showed a sensitivity of only 22% in infections with 10⁴ to 10⁵ cfu/mL versus 60% for those with greater than 10⁵ cfu/mL. In other words, low-level pyuria (5 to 20 white blood cells per highpower field microscopy) may be associated with a false-negative LE test. These tests are also best performed on concentrated first-morning voided specimens. It has been suggested that false-negative results are more likely if the test is used as a sampling technique at other times during the day (66). Furthermore, certain dyes such as bilirubin, methylene blue, or phenazopyridine may interfere with interpretation of the test.

Other rapid detection tests, such as filter methods (e.g., Back-T-Screen, Marion Laboratories, Inc., Kansas City, MO), concentrate a specific quantity of urinary sediment on a filter of controlled pore size. One milliliter of urine is mixed with 3 mL of a diluent containing glacial acetic acid and other ingredients that dissolve crystals and increase adherence of bacteria and leukocytes. The diluted mixture is then passed through the filter and rinsed with a diluent. A safranin dye is then used to stain the bacteria and leukocytes, and a decolorizer is added to remove excess dye. Resulting colors are compared with a reference to quantitate the presence of bacteria and leukocytes. The sensitivity of these tests for urine infected with 10⁴ to 10⁵ cfu/mL is 34% to 65%. As the number of organisms increases to greater than 10⁵, the sensitivity also increases to 79% to 85%. The specificity of this test at lower bacterial counts is about 75% (65,67). The main advantage of these tests is a more reliable detection of smaller numbers of bacteria at the expense of lower specificity (68). The test is believed by some to be a good screening method because it detects both bacteria and pyuria.

A symptomatic patient should be treated with antibiotics even if an office urinary kit is negative for both nitrites and leukocytes. Empiric antibiotic use is guided by symptoms because treatment leads to faster resolution of lower urinary tract symptoms and will reduce the median duration of constitutional symptoms (fever and shivers) by 4 days (69). Additionally, treatment can reduce time of restricted activity and leave from work.

In the patient who has clinical signs of acute lower UTI and is noted to have pyuria, bacteriuria, or hematuria on one of the previously mentioned office tests, it is reasonable to initiate antibiotic therapy without obtaining a urine culture. However, if one of the screening techniques is deemed inappropriate or inconclusive, if the patient has recurrent infection that has not been subjectively relieved with previous antibiotics, or if signs and symptoms are consistent with upper UTI, a bacterial culture and sensitivity should be performed.

The traditional approach to the interpretation of a urinary culture has been that there must be growth of at least 10⁵ cfu/mL to consider it positive. This criterion is based on studies demonstrating that the finding of at least 10⁵ cfu/mL on two consecutive urine cultures distinguishes women with asymptomatic bacteriuria or pyelonephritis from those with contaminated specimens (55,56,60).

The use of this cutoff, however, has two limitations for the clinician who treats these patients. First, 20% to 24% of women with symptomatic urinary infections present with less than 10⁵ bacteria/mL of urine (57,70, 71, 72). This is probably secondary to a slow doubling time of bacteria in urine combined with frequent bladder emptying from persistent irritation. Stamm et al (73) proposed that the best diagnostic criterion for culture detection in young symptomatic women is 10² cfu/mL, not 10⁵ cfu/mL.

The second limitation of the 10⁵ cutoff is one of overdiagnosis. In the original studies by Kass (55,56,74), a single culture of at least 10⁵ cfu/mL had a 20% chance of representing contamination. Because patients who are susceptible to infection often carry large numbers of pathogenic bacteria on the perineum, contamination of an otherwise sterile urine can occur. For this reason, care in the collection of the urine specimen must again be emphasized. Most health care workers do not spend much time and effort to explain adequately how a patient should collect a midstream urine clean-catch specimen. However, one study showed contamination rates to be similar among patients whose urine samples were collected with traditional instructions (midstream urine sample, perineal cleansing, and spreading of the labia) compared with urine samples of patients told to urinate into a clean container without cleansing (75).

Although methods of obtaining cultures in the office are available, most clinicians use commercial
laboratories. One should be familiar with the individual laboratory policy of reporting culture results. Some laboratories report any culture of less than 10⁵ cfu/mL as negative and often report only the predominant organism in mixed cultures.

Sensitivity testing is also usually obtained using a commercial laboratory, even though office tests have been described. The disadvantages of sensitivity testing include the time involved, which is typically 24 to 48 hours; the absence of control of processing by the referring physician; and the relatively high cost.

Cystitis may appear as diffuse inflammation (nonraised, red areas distributed throughout) on cystoscopy. Occasionally, one may note one or multiple small clear cysts throughout the bladder from a resolved UTI. This is termed “cystitis cystica.”

Cystoscopy is not routinely indicated for the evaluation of lower UTI, and routine endoscopic evaluation in females with UTI is a controversial issue. Fowler and Pulaski (76) reported on 74 cystoscopies performed in women with two or more previous infections and noted the only abnormality that altered treatment was the presence of a urethral diverticulum in three cases. Engel et al (77) reviewed 153 women who had undergone cystoscopy for UTI. Although abnormalities were noted in 62% of the cases, 84% of these abnormalities were inflammatory in nature and presumably secondary to prior infection. Only one abnormality, a colovesical fistula, had an effect on treatment (77).