Screening for Asymptomatic Bacteriuria in Adults: Evidence for the U.S. Preventive Services Task Force Reaffirmation Recommendation Statement

Methods

The USPSTF requested that this update address 2 primary key questions:

1. What are the benefits of screening and treatment for asymptomatic bacteriuria in pregnant women, nonpregnant women, and men?

2. What are the harms of screening for asymptomatic bacteriuria in pregnant women, nonpregnant women, and men?

The Task Force determined that this update need not review new evidence for the harms of antibiotic treatment for asymptomatic bacteriuria because the adverse effects of commonly used antibiotic medications are well established.

Data Sources and Searches

We performed literature searches for the benefits of screening for asymptomatic bacteriuria and the harms of screening, limited 1 January 2002 through 30 April 2007, using the search terms asymptomatic bacteriuria, screening, and urine culture. Initial searches were limited to English-language articles indexed in the Cochrane Database of Systematic Reviews and PubMed core clinical journals. Core journals are a subset of 120 English-language journals defined by the National Library of Medicine, previously known as the Abridged Index Medicus. When initial searches revealed few articles, we expanded searches to include noncore journals. We supplemented these searches by reviewing reference lists of recent systematic and narrative reviews and clinical guidelines.

Study Selection

We searched for studies on the benefits and harms of screening and the benefits of treatment for asymptomatic bacteriuria. We included studies of adults 18 years of age or older from the United States and from other countries with patient populations generalizable to the United States. We excluded studies of very high-risk or special patient populations, including patients with a history of recurrent urinary tract infections, immunocompromised patients, and patients with sickle cell disease.

For benefits of screening or treatment of screened populations, we included randomized, controlled trials (RCTs); meta-analyses; and systematic reviews. For harms of screening, we included systematic reviews, meta-analyses, RCTs, cohort studies, case–control studies, and case series of large multisite databases. We excluded editorials, case reports, narrative reviews, and guideline reports.

We evaluated all articles for predetermined exclusion criteria at each stage of review (title, abstract, and full article). Articles selected by at least 1 team member advanced to the next stage of review. At the full article stage, we resolved differences of opinion by consensus.

Data Extraction

We abstracted information on sample size, entry criteria, demographic characteristics, comorbid conditions, study design, treatment group allocation, and clinical outcomes of interest.

Data Synthesis and Analysis

Data from included studies were synthesized qualitatively in a narrative format.

Role of the Funding Source

The work of the USPSTF is supported by the Agency for Healthcare Research and Quality. This review did not receive separate funding.

Results

We identified 93 potentially eligible articles and entered them into a reference database. After sequential application of exclusion criteria (Figure), 1 systematic review of treatment for asymptomatic bacteriuria in pregnant women and 1 RCT of treatment for asymptomatic bacteriuria in nonpregnant women with diabetes met inclusion criteria for this update. An additional prospective cohort study of outcomes of asymptomatic bacteriuria in diabetic women that did not meet the inclusion criteria is also reviewed in detail. These 3 new studies are discussed in the next sections (Table).

1. What Are the Benefits of Screening and Treatment for Asymptomatic Bacteriuria in Pregnant Women, Nonpregnant Women, and Men?

Pregnant Women

We identified no new RCTs of screening for asymptomatic bacteriuria in pregnant women. A 2007 Cochrane systematic review and meta-analysis of randomized trials comparing antibiotic treatment with placebo or no treatment included 14 studies involving 2302 pregnant women. The review found statistically significant reductions in the incidence of pyelonephritis (relative risk [RR], 0.23 [95% CI, 0.13 to 0.41]) and low-birthweight babies (RR, 0.66 [CI, 0.49 to 0.89]) (3).

Although all of the studies included in the Cochrane review were published in 1987 or earlier, we considered the meta-analysis to represent new evidence because it came to a different conclusion than a previous Cochrane review (4) about the effect of antibiotic treatment on the incidence of preterm delivery. The previous review relied on the standard definition of preterm delivery from the 1960s, birthweight less than 2500 g, rather than the currently accepted definition of birth before a certain gestational age. When the Cochrane review investigators included only the 3 studies that used a gestational age–based definition of preterm delivery, antibiotic treatment for asymptomatic bacteriuria had no effect on rates of preterm delivery (RR, 0.37 [CI, 0.10 to 1.36]) (3).

Nonpregnant Women

We found 1 good-quality RCT of treatment that compared antibiotics with placebo in diabetic women with asymptomatic bacteriuria (5). Endocrinology offices in 2 tertiary care hospitals and community practices in Canada recruited 105 participants. Eligible patients were nonpregnant women older than 16 years of age with diabetes, normal renal function, no symptoms consistent with a urinary tract infection, and 2 consecutive positive urine cultures (≥10⁵ colony-forming units of an organism per mL of urine). After randomization, patients received either trimethoprim–sulfamethoxazole (or ciprofloxacin if they could not use sulfa drugs) or placebo. Four weeks after treatment, and every 3 months for up to 36 months, investigators rescreened patients for bacteriuria with urine cultures. Patients in the experimental group with positive results received additional antibiotics for progressively longer periods. Patients in either group who developed symptomatic urinary infections were also treated with the study drugs.

After the initial round of therapy, significantly (P < 0.001) more women in the placebo group continued to have bacteriuria. However, during follow-up, the time to symptoms consistent with a urinary tract infection was similar in both groups. The placebo group was no more likely than the treatment group to have a symptomatic urinary tract infection (RR, 1.19 [CI, 0.28 to 1.81]), pyelonephritis (RR, 2.13 [CI, 0.81 to 5.62]), or hospitalization for a urinary tract infection (RR, 1.93 [CI, 0.47 to 7.89]). Patients in the placebo group averaged 34 days of antibiotic use per 1000 days of follow-up, compared with 158 days of antibiotic use per 1000 days of follow-up in the treatment group.

Two additional clinically important outcomes thought to be related to asymptomatic bacteriuria are renal function decline and development of hypertension. We identified a multicenter prospective cohort study of interest that did not meet inclusion criteria for this review because it was not a randomized trial (6). We present it here as a good-quality study that contributed to the USPSTF's understanding of screening for asymptomatic bacteriuria in women with diabetes.

A total of 644 women with type 1 and type 2 diabetes, who were from hospital outpatient diabetic clinics in the Netherlands, were evaluated for asymptomatic bacteriuria and monitored for a mean of 6.1 years for changes in creatinine clearance and blood pressure. A multivariate analysis, adjusted for patient age, length of follow-up, duration of diabetes, and presence of microalbuminuria at study entry, found no association between the presence of asymptomatic bacteriuria and subsequent changes in creatinine clearance. Similarly, investigators observed no relationship between asymptomatic bacteriuria and the development of hypertension. These findings suggest that screening and treatment for asymptomatic bacteriuria in women with diabetes would have no effect on either outcome.

Men

We identified no studies of screening or treatment for asymptomatic bacteriuria in men.

2. What Are the Harms of Screening for Asymptomatic Bacteriuria in Pregnant Women, Nonpregnant Women, and Men?

We identified no new studies of the harms of screening for asymptomatic bacteriuria in pregnant women, nonpregnant women, or men.

Discussion

Although urine culture remains the gold standard for screening for and diagnosis of asymptomatic bacteriuria in pregnant women, it is time- and labor-intensive and patients may have difficulty providing uncontaminated samples for testing. It would be useful to identify a rapid test with a high negative predictive value for asymptomatic bacteriuria that could replace urine culture as a screening test. Urine cultures would then be performed only on patients with positive screening results.

The 2004 USPSTF evidence update on screening for asymptomatic bacteriuria (7) highlighted the Uriscreen (Savyon Diagnostics, Ashdod, Israel) enzymatic screening test as having good potential because of its reported 100% sensitivity and negative predictive value in 1 study of 313 consecutive pregnant patients in Israel. However, in a 2005 study of 150 asymptomatic pregnant women in Venezuela, the test detected only 17 of 28 patients whose catheterized urine samples showed bacteriuria, yielding a sensitivity of 60.7% and negative predictive value of 90.8% compared with urine culture (8).

Whether pregnant women benefit from additional screening for asymptomatic bacteriuria after the first trimester is unknown. A 2005 Canadian study conducted in outpatient clinics at an urban teaching hospital compared the diagnostic yield of a single urine culture, 2 urine cultures, and 3 urine cultures at fewer than 20 weeks', 28 weeks', and 36 weeks' gestation (9). Additional cultures were performed at routine prenatal visits if a leukocyte esterase–nitrite test result was positive; the gold standard for asymptomatic bacteriuria was any single positive urine culture.

Forty-nine of the 1050 patients in the Canadian study had asymptomatic bacteriuria. A single urine culture before 20 weeks' gestation (consistent with the USPSTF recommendation) detected only 40.8% of cases, whereas the 3-urine-culture strategy detected 87.8% of cases. However, the study did not assess the effect of increased detection on clinical outcomes.

In summary, we found some new evidence that continues to support screening for asymptomatic bacteriuria in pregnant women, as well as evidence that suggests no benefit from screening other groups of adults. No currently available screening tests have a high enough sensitivity and negative predictive value for asymptomatic bacteriuria in pregnant women to replace the urine culture. Future research is needed to clarify the optimal timing and periodicity of screening in pregnant women.