Abstract
OBJECTIVE
To evaluate urinalysis testing patterns within the Veterans Health Administration (VHA), estimate the proportion and likelihood of patients who completed a urinalysis to have microscopic hematuria (MH), and explore how urinalysis testing patterns may influence MH detection.
METHODS
This was a retrospective cross-sectional study using VHA data. We identified adult patients without a known urologic cancer history who had at least 1 outpatient visit at any VHA site and at least 1 interpretable urinalysis performed in 2015. The factors associated with the number of urinalyses performed on each patient and associations with the presence of MH were investigated.
RESULTS
Among 5,719,966 adults, 39% completed a urinalysis. Variation in the proportion of patients who completed urinalyses was highest by age, among patients with hypertension and diabetes, and by region. Of patients who underwent urinalysis and had no prior genitourinary cancer history, 54% did not have an interpretable urinalysis result. Among patients with at least one interpretable microscopic urinalysis, 37% had MH. This was more common among older patients, females, current smokers, and patients with more comorbidities. Variation in the likelihood of patients having MH remained after adjusting for multiple factors and when contextualized by urinalysis completion and interpretability patterns.
CONCLUSION
The number of urinalyses performed in the VHA system is remarkably high. Detection of MH is influenced by the frequency of urinalysis testing and interpretability of results. The presence and detection of MH varies by factors which should be considered when adjudicating the need for further evaluation of MH.
The examination of urine has been used for thousands of years to understand human health. In contemporary medicine, urinalysis remains a commonly used laboratory study that can provide important information about myriad organ systems and disease processes. Urinalysis is an especially valuable test for urinary tract disease but is not recommended for disease screening outside of pregnancy. However, many patients still complete urinalyses for nonurologic or unclear indications. Therefore, when nonvisible blood (microscopic hematuria [MH]) is present on routine urinalysis, it is often an incidental finding. MH is a mostly benign1-3 or idiopathic clinical sign,4 but in approximately 1% of patients it may be the only indication of an occult genitourinary (GU) cancer. Accordingly, patients with MH are recommended to undergo a urologic evaluation to best rule out the presence of kidney or urothelial cancer as the source of bleeding.5,6
The prevalence of MH in the general population remains poorly defined. The most often cited estimates vary widely, from 2.6% to 54%,7-13 and they are from studies conducted over 30 years ago among high-risk populations or were conducted without microscopic examination (dipstick only) and/or a uniform definition of MH. In the absence of a population-level screening study that can provide an optimal estimate of MH prevalence and the related incidence of urologic disease, it is important to understand how the frequency and intensity of urinalysis testing may influence the detection of MH in clinical practice. Overdetection or underdetection of MH due to disparate testing patterns can influence the subsequent need for further diagnostic evaluation and can lead to either missed urologic cancers or unnecessary invasive testing.
We sought to assess contemporary practice patterns for outpatient urinalysis testing and estimate the patient-level variation and proportion of patients with MH on at least one urinalysis obtained at the Veterans Health Administration (VHA), the largest integrated healthcare system in the United States. Through a stepwise process beginning with identifying the likelihood of a patient completing a urinalysis, we assess the interpretability of urinalysis results for the purposes of detecting MH, and variation in the presence of MH by relevant sociodemographic and clinical factors.
METHODS
Study Design and Data Source
We conducted a retrospective cross-sectional study using VHA administrative and clinical data according to Strengthening Reporting of Observational Studies in Epidemiology (STROBE) guidelines.14 All data were sourced from clinical, cancer, and mortality registries within the VA Corporate Data Warehouse (CDW) and Centers for Medicare & Medicaid Services administrative claims data linked at the patient level via Social Security number.15
Cohort Identification
As part of a larger long-term study which aimed to identify factors associated with the presence of bladder cancer in veterans with MH, we identified all adults (≥18 years) who had at least one outpatient visit at any VHA site in calendar year 2015. Among these, we identified 2,250,654 patients with at least one completed urinalysis using VHA’s National Laboratory Test (NLT) list (Supplementary Methods). We then excluded patients with a known prior history of bladder cancer, kidney cancer, or upper tract urothelial cancer as these are known causes of hematuria. The resulting cohort of 2,196,989 patients without a known urologic cancer history who had at least 1 urinalysis completed during 2015 (Supplementary Fig. 1, Point A) represents a group of patients for whom a urologic evaluation would be indicated should they have MH. Patients with a known history of nephrolithiasis, medical renal disease, or chronic kidney disease (CKD) were not excluded, as these patients would still meet criteria for a urologic evaluation to rule out cancer if found to have MH. To rule out a temporal influence on our results, we explored urinalysis testing patterns of years prior to and subsequent to 2015 (2012-2021) and performed a sensitivity analysis for our outcomes of interest for the years 2017 and 2019, individually.
Urinalysis Identification and Eligibility
The current definition of clinically significant MH is 3 or more red blood cells per high-power field (≥3 RBC/hpf). The interpretability of urinalysis requires a microscopic examination that quantifies the number of RBCs rather than one which provides a qualitative result like that from a dipstick assay.16 Therefore, dipstick-only urinalyses were not considered interpretable for the purposes of detecting MH. Urinalyses with non-numeric but still categorizable results such as “none” or “too numerous to count” were included as 0 RBC/hpf and > 100 RBC/hpf, respectively. We also excluded urinalyses without necessary adjunctive test results (leukocyte esterase, nitrite) required to rule out urinary tract infection (UTI) and/or gross hematuria (via urine color). There were 1,289,011 urinalyses with all supporting tests present among 862,906 unique patients, thus comprising the cohort of patients with an “interpretable” microscopic urinalysis (Supplementary Fig. 1, Point B). Patients with multiple urinalyses in the calendar year were considered to have microhematuria if any interpretable urinalysis had ≥3 RBC/hpf in the absence of other exclusion criteria.
Sociodemographic and Clinical Characteristics
For each patient, we include sociodemographic variables for age, sex, and self-reported race/ethnicity from the CDW. We identified select chronic medical conditions (eg, diabetes, hypertension, CKD, coronary artery disease [CAD], and medical renal disease) by querying diagnoses from the 2 years prior to a patient’s index urinalysis date using International Classification of Diseases, Ninth and Tenth Revisions (ICD-9 and ICD-10) diagnosis codes. We used an algorithm developed and validated within VHA data17 to categorize smoking history as never, former, or current. Estimated glomerular filtration rate (eGFR) was calculated from serum creatinine (± 12 months from a person’s index date) using the CKD-EPI method; from the calculated eGFR, individuals were categorized into CKD stages I-V according to standard thresholds. Proteinuria was categorized according to standard estimated albumin/creatinine ratio estimates.18
Statistical Analyses
We first assessed differences among those who did or did not complete a urinalysis. The proportion of patients with MH was then calculated using the number of patients with at least one interpretable microscopic urinalysis as the denominator and those with 1 or more interpretable microscopic urinalyses that had ≥3 RBC/hpf as the numerator (Supplementary Fig. 1, Point B).
To determine factors associated with the number of interpretable microscopic urinalyses performed per patient, we used multivariable negative binomial regression. Included in the model were age (plus restricted cubic splines as nonlinear terms), sex, race and ethnicity, smoking history, diabetes, hypertension, CAD, CKD (stages 3-5 vs 1-2), and medical renal disease status; and location within the VA’s system of 18 regional Veterans Integrated Services Networks (VISNs). This approach contextualizes the frequency of urinalysis testing as it relates to the presence of microhematuria among those with more than 1 urinalysis during the study period. Among those with an interpretable microscopic urinalysis in 2015, we also sought to identify factors associated with the presence of MH at the patient-level using multivariable mixed effects logistic regression, adjusting for the same variables listed above and using a random intercept for patient.
We conducted all analyses using SAS Enterprise Guide, version 8.3 (SAS Institute), and R, version 4.3.1 (R Project for Statistical Computing). All statistical tests were 2 sided with P < .05 indicating statistical significance. Confidence intervals for proportions are not given, as they are very narrow due to the large sample size.
RESULTS
Variation in Urinalysis Ordering
A total of 5,719,966 adult patients had an outpatient visit at a VHA facility during calendar year 2015 and 38% without a prior GU cancer history completed at least one urinalysis, regardless of interpretability (n = 2,196,989). The median number of ordered and interpretable urinalyses at the patient level was 1 (IQR 1-2). Age group (urinalyses were conducted more frequently in older patients), hypertension (43% vs 34% without), diabetes (45% vs 36% without), renal disease (53% vs 38% without), and geographic region (range 25%-52%) were the factors associated with the largest differences in urinalyses completed. The absolute difference in the proportion of male and female patients with a completed urinalysis of any kind was 2% (Table 1). No differences were seen among patients based on CKD stage, and a slightly higher proportion of current smokers (41%) completed a urinalysis than never (38%) or former (37%) smokers (Table 2). During the years 2012-2019, the number of completed urinalyses among eligible VA patients ranged from 3,697,426-3,886,503 indicating minimal temporal change to practice patterns at the system level. The years 2020 and 2021 had significantly fewer patients complete a urinalysis (2,587,468 and 2,752,058, respectively), likely related to the COVID-19 pandemic.
Table 1.
Cohort demographics for patients who completed at least 1 urinalysis (UA), and patients with an interpretable UA.
| Completed UA N = 2,196,989 |
Interpretable UA N = 862,906 |
Proportion Interpretable | |
|---|---|---|---|
| Sex | |||
| Male | 2,025,045 (39%) | 772,723 (15%) | 38% |
| Female | 171,944 (37%) | 90,183 (19%) | 52% |
| Age categories | |||
| Under 30 years of age | 80,722 (29%) | 26,713 (9.7%) | 33% |
| 30-39 years of age | 151,231 (32%) | 54,347 (12%) | 36% |
| 40-49 years of age | 201,736 (37%) | 73,036 (13%) | 36% |
| 50-59 years of age | 366,838 (41%) | 141,977 (16%) | 39% |
| 60-69 years of age | 756,033 (42%) | 295,839 (17%) | 39% |
| 70-79 years of age | 376,062 (40%) | 154,668 (16%) | 41% |
| 80-89 years of age | 222,783 (34%) | 96,710 (15%) | 43% |
| 90+ years of age | 41,584 (29%) | 19,616 (13%) | 47% |
| Race and ethnicity | |||
| Non-Hispanic White | 1,487,120 (38%) | 540,520 (14%) | 36% |
| Non-Hispanic Black or African | 388,308 (42%) | 166,833 (18%) | 43% |
| American | |||
| Hispanic | 150,836 (44%) | 89,169 (26%) | 59% |
| American Indian or Alaska Native | 16,520 (38%) | 6546 (15%) | 40% |
| Asian | 18,661 (34%) | 7556 (14%) | 41% |
| Pacific Islander | 17,237 (36%) | 6890 (15%) | 40% |
| Missing | 118,307 (32%) | 45,392 (12%) | 38% |
| Veterans Integrated Service Network | |||
| New York/New Jersey | 117,491 (43%) | 56,579 (20%) | 48% |
| New England | 74,599 (33%) | 21,151 (9.3%) | 28% |
| Pennsylvania | 97,259 (37%) | 27,306 (10%) | 28% |
| Capitol | 75,872 (41%) | 23,394 (13%) | 31% |
| Mid-Atlantic | 141,636 (43%) | 50,696 (15%) | 36% |
| Southeast | 202,142 (52%) | 83,683 (21%) | 41% |
| Florida | 239,925 (45%) | 124,283 (23%) | 52% |
| Mid-South | 100,895 (40%) | 35,571 (14%) | 35% |
| Ohio, IN, MI | 124,331 (28%) | 40,035 (8.9%) | 32% |
| Great Lakes | 92,813 (35%) | 21,904 (8.2%) | 24% |
| Heartland | 107,451 (47%) | 47,411 (21%) | 44% |
| South Central | 189,727 (49%) | 57,552 (15%) | 30% |
| Heart of Texas | 109,450 (31%) | 72,307 (20%) | 66% |
| Rocky Mtn | 67,426 (25%) | 20,336 (7.6%) | 30% |
| Northwest | 76,222 (29%) | 44,420 (17%) | 58% |
| Sierra Pacific | 89,201 (30%) | 28,446 (9.5%) | 32% |
| Desert Pacific | 182,049 (41%) | 73,589 (17%) | 40% |
| Midwest | 108,500 (36%) | 34,243 (11%) | 32% |
Percentage listed is the row proportion who completed an UA or completed an interpretable UA (respectively for each column) among all patients seen at Veterans Health Administration (VHA) during 2015. Proportion interpretable is the proportion of those who had an interpretable UA (numerator) among those with at least one completed UA (denominator).
Table 2.
Cohort comorbidities for patients who completed at least 1 urinalysis (UA), and patients with an interpretable UA.
| Completed UA N = 2,196,989 | Interpretable UA N = 862,906 | Proportion Interpretable | |
|---|---|---|---|
| Smoking history | |||
| Never | 564,843 (38%) | 238,465 (16%) | 42% |
| Former | 981,192 (37%) | 362,208 (14%) | 37% |
| Current | 650,954 (41%) | 262,233 (16%) | 40% |
| Diabetes | |||
| No | 1,573,071 (36%) | 577,373 (13%) | 37% |
| Yes | 623,918 (45%) | 285,533 (20%) | 46% |
| Hypertension | |||
| No | 898,783 (34%) | 315,021 (12%) | 35% |
| Yes | 1,298,206 (43%) | 547,885 (18%) | 42% |
| Chronic kidney disease stage | |||
| 1-2 | 1,787,503 (43%) | 670,582 (16%) | 38% |
| 3-5 | 400,611 (44%) | 189,318 (21%) | 47% |
| Coronary artery disease | |||
| No | 1,913,184 (38%) | 737,209 (15%) | 39% |
| Yes | 283,805 (43%) | 125,697 (19%) | 44% |
| Medical renal disease | |||
| No | 2,177,392 (38%) | 850,566 (15%) | 39% |
| Yes | 19,597 (53%) | 12,340 (33%) | 63% |
| Proteinuria at index UA | |||
| Negative | 1,677,251 (98%) | 572,424 (33%) | 34% |
| Trace | 232,739 (96%) | 156,791 (65%) | 67% |
| Small | 101,838 (94%) | 76,700 (71%) | 75% |
| Moderate-large | 23,256 (95%) | 17,256 (70%) | 74% |
Percentage listed is the row proportion who completed UA or completed an interpretable UA (respectively for each column) among all patients seen at VHA during 2015. Proportion interpretable is the proportion of those who had an interpretable UA (numerator) among those with at least one completed UA (denominator).
Variation in Urinalysis Interpretability
Urinalyses were considered uninterpretable for the purposes of detecting clinically significant MH if they lacked a numeric (n = 1,175,363) or interpretable (n = 87,987) RBC/hpf result, or if they had missing adjunctive urinalysis results (n = 49,432). Urinalyses that had a concomitant UTI (n = 44,077) or obvious gross hematuria (n = 12,731) were also excluded as these would not meet the criteria for MH evaluation. A higher proportion of women compared to men had an interpretable microscopic urinalysis despite lower overall testing among women (52% vs 38%). Similar differences were seen in the proportion of completed urinalysis of any kind to interpretable microscopic urinalyses by region and by race/ethnicity (Table 1). Similar patterns and effect sizes were seen as part of sensitivity analyses for the years 2017 and 2019 (data not presented).
Prevalence of MH
The overall proportion of patients who had at least one interpretable microscopic urinalysis with ≥3 RBC/hpf was 37% (n = 319,772). MH was found in 46% of women compared to 36% of men. MH was more prevalent in older patients than younger patients (24% ages 30-39 and 42% ages 80-89). The percent probability of MH increased with advancing age, with a notable acceleration of this increase around 62-65 years (Supplementary Fig. 2B). The proportion of current smokers with MH (39%) was higher than both never (36%) and former (36%) smokers. Patients with CAD, worse CKD, hypertension, or known renal disease also more often had MH than those without (Table 3). Among interpretable microscopic urinalyses that had MH present, 68% (n = 289,770) had RBC/hpf values in the range 3-10 and 17% (n = 72,795) had > 25 RBC/hpf (Supplementary Fig. 3). Some degree of proteinuria was found in 18.2% of all patients with a moderate-to-large degree seen in 1.2%; MH was more often seen in patients with a greater degree of proteinuria.
Table 3.
Proportion of patients with at least 1 interpretable urinalysis (UA) positive for microscopic hematuria (MH; 3 or more RBC/hpf).
| Proportion With MH | |
|---|---|
| Sex | |
| Male | 278,564 (36%) |
| Female | 41,208 (46%) |
| Age categories | |
| Under 30 years of age | 8482 (32%) |
| 30-39 years of age | 18,428 (34%) |
| 40-49 years of age | 25,507 (35%) |
| 50-59 years of age | 51,711 (36%) |
| 60-69 years of age | 108,256 (37%) |
| 70-79 years of age | 57,810 (37%) |
| 80-89 years of age | 40,211 (42%) |
| 90+ years of age | 9367 (48%) |
| Race and ethnicity | |
| Non-Hispanic White | 200,915 (37%) |
| Non-Hispanic Black or African | 72,710 (44%) |
| American | |
| Hispanic | 23,836 (27%) |
| American Indian or Alaska | 2043 (31%) |
| Native | |
| Asian | 2008 (27%) |
| Pacific Islander | 2337 (34%) |
| Missing | 15,923 (35%) |
| Smoking history | |
| Never | 85,518 (36%) |
| Former | 131,605 (36%) |
| Current | 102,649 (39%) |
| Diabetes | |
| No | 214,755 (37%) |
| Yes | 105,017 (37%) |
| Hypertension | |
| No | 111,253 (35%) |
| Yes | 208,519 (38%) |
| Chronic kidney disease stage | |
| 1-2 | 240,902 (36%) |
| 3-5 | 77,727 (41%) |
| Coronary artery disease | |
| No | 267,407 (36%) |
| Yes | 52,365 (42%) |
| Medical renal disease | |
| No | 314,159 (37%) |
| Yes | 5613 (45%) |
| Proteinuria at index UA | |
| Negative | 190,237 (33%) |
| Trace | 70,748 (45%) |
| Small | 39,399 (51%) |
| Moderate-large | 9166 (53%) |
Factors Associated With Interpretable Urinalysis Testing and the Presence of MH
After adjustment, several patient factors were associated with a higher likelihood of having interpretable microscopic urinalysis testing completed (Table 4), most strongly female sex (incidence rate ratio [IRR] 1.66, 95% CI 1.65, 1.68), Hispanic ethnicity (IRR 2.00, 95% CI 1.98, 2.01), and renal disease (IRR 2.18, 95% CI 2.14, 2.23). Regionally, there was significant variation in interpretable microscopic urinalyses. The predicted count by age followed an inverted U-shaped curve; specifically, the predicted count of interpretable urinalyses steadily increased with advancing age, peaking around age 60, and steadily decreasing thereafter (Supplementary Fig. 2A).
Table 4.
Side by side comparison of the multivariable negative binomial regression estimating the number of eligible/in-terpretable urinalyses (UAs) and a multivariable mixed effects logistic regression with patient as a random intercept estimating the odds of having a UA with microscopic hematuria (MH; ≥3 RBC/hpf).
| Number of UAs |
Odds of MH |
|||
|---|---|---|---|---|
| Characteristic | IRR | 95% CI | OR | 95% CI |
| Age, y (nonlinear terms)† | ||||
| Female sex (ref: male) | 1.66 | 1.65, 1.68 | 1.71 | 1.68, 1.75 |
| Race and ethnicity | ||||
| Non-Hispanic White | — | — | — | — |
| Non-Hispanic Black or African American | 1.20 | 1.20, 1.21 | 1.23 | 1.22, 1.25 |
| Hispanic | 2.00 | 1.98, 2.01 | 0.45 | 0.44, 0.46 |
| American Indian or Alaska Native | 1.14 | 1.11, 1.17 | 0.81 | 0.76, 0.87 |
| Asian | 1.23 | 1.20, 1.26 | 0.76 | 0.72, 0.82 |
| Pacific Islander | 1.11 | 1.08, 1.14 | 0.85 | 0.80, 0.90 |
| Missing | 0.94 | 0.93, 0.95 | 0.95 | 0.92, 0.97 |
| Smoking status | ||||
| Never | — | — | — | — |
| Former | 0.96 | 0.95, 0.96 | 1.04 | 1.03, 1.06 |
| Current | 1.16 | 1.16, 1.17 | 1.29 | 1.27, 1.31 |
| Comorbidities | ||||
| Diabetes | 1.32 | 1.31, 1.33 | 0.78 | 0.77, 0.79 |
| Hypertension | 1.26 | 1.25, 1.26 | 0.95 | 0.94, 0.97 |
| Chronic kidney disease stage | ||||
| 1-2 | — | — | — | — |
| 3-5 | 1.37 | 1.37, 1.38 | 1.01 | 0.99, 1.02* |
| Coronary artery disease | 1.19 | 1.18, 1.20 | 1.16 | 1.14, 1.18 |
| Medical renal disease | 2.18 | 2.14, 2.23 | 1.28 | 1.23, 1.34 |
| Veterans Integrated Service Network | ||||
| New York/New Jersey | — | — | — | — |
| New England | 0.43 | 0.43, 0.44 | 2.33 | 2.23, 2.43 |
| Pennsylvania | 0.45 | 0.45, 0.46 | 2.43 | 2.33, 2.53 |
| Capitol | 0.65 | 0.64, 0.66 | 2.88 | 2.77, 3.00 |
| Mid-Atlantic | 0.64 | 0.64, 0.65 | 2.22 | 2.15, 2.30 |
| Southeast | 0.87 | 0.86, 0.88 | 6.00 | 5.80, 6.21 |
| Florida | 1.03 | 1.02, 1.04 | 3.39 | 3.29, 3.50 |
| Mid-South | 0.62 | 0.61, 0.63 | 3.15 | 3.03, 3.27 |
| Ohio, IN, MI | 0.39 | 0.39, 0.40 | 4.26 | 4.10, 4.42 |
| Great Lakes | 0.35 | 0.35, 0.36 | 3.75 | 3.59, 3.91 |
| Heartland | 0.91 | 0.90, 0.93 | 4.06 | 3.92, 4.21 |
| South Central | 0.60 | 0.59, 0.60 | 3.23 | 3.13, 3.34 |
| Heart of Texas | 0.87 | 0.86, 0.88 | 0.54 | 0.52, 0.55 |
| Rocky Mtn | 0.32 | 0.32, 0.33 | 1.87 | 1.79, 1.96 |
| Northwest | 0.77 | 0.76, 0.78 | 0.82 | 0.79, 0.85 |
| Sierra Pacific | 0.45 | 0.44, 0.45 | 1.91 | 1.84, 1.99 |
| Desert Pacific | 0.70 | 0.70, 0.71 | 2.59 | 2.51, 2.67 |
| Midwest | 0.53 | 0.52, 0.54 | 2.14 | 2.07, 2.22 |
CI, confidence interval; IRR, incidence rate ratio; OR, odds ratio; RBC/hpf, red blood cells per high-power field.
All variables were statistically significant at the < .001 level except when indicated with an asterisk (*). The P value for the relationship between chronic kidney disease stage 3-5 and the odds of having MH was 0.2.
For MH+ OR, age term is per 10 y; see Supplementary Figure 2.
By comparing the likelihood (IRR) of having an interpretable microscopic urinalysis with odds of having MH (OR) on a urinalysis, we were able to assess whether the presence of MH is related to testing patterns. For example, factors with concordant IRRs and ORs both > 1.0 would indicate more frequent testing and higher odds of having MH. Conversely, factors that have IRRs and ORs < 1.0 would suggest less frequent testing and lower odds of finding MH. A discordant IRR and OR is more difficult to interpret because it is unclear whether a higher or lower probability of finding MH could be related to more or less selective testing, respectively.
Among factors associated with more frequent testing, older age (Supplementary Fig. 2), female sex (OR 1.71, 95% CI 1.68, 1.75), currently smoking (OR 1.29, 95% CI 1.27, 1.31) relative to never smoking, having CAD (OR 1.16, 95% CI 1.14, 1.18), and having renal disease (OR 1.28, 95% CI 1.23, 1.34) also had a higher odds of MH (Table 4). Several factors had a higher likelihood of urinalysis testing but a lower likelihood of MH, most notably the presence of diabetes, hypertension, and Hispanic ethnicity. Patients in many VISN regions were less likely to have interpretable urinalyses than the referent region (New York/New Jersey) but had significantly increased odds of having MH.
DISCUSSION
This study presents contemporary estimates within the largest integrated healthcare system in the United States of the proportion of adult outpatients who completed a urinalysis, and of those, the proportion with MH when a microscopic urinalysis was performed. It provides important context to the detection of MH relative to variation in urinalysis testing and the use of microscopic urinalysis, specifically. We found that 37% of VHA outpatients who had an interpretable microscopic urinalysis and did not have a prior GU cancer history, gross hematuria, or concomitant UTI had clinically significant MH (≥3 RBC/hpf). The association between female sex, medical renal disease, and Hispanic ethnicity with urinalysis completion patterns is a novel and important finding of our study that may help explain previously reported differences in the completion of a MH evaluation and resultant disparate cancer detection among women and non-White patients.19-21
Without understanding which patients in an observational study are most likely to complete a urinalysis, it is difficult to interpret MH prevalence estimates. Prior studies have reported differences in who receives a MH evaluation, and the referral and detection biases inherent to these studies have skewed our estimates of the incidence of MH-related cancers and disease.19-21 Our results, particularly the factors associated with both an increased likelihood of testing and higher odds of having MH (female sex, currently smoking), raise the question of whether MH is: (1) truly associated with an underlying medical problem and worthy of evaluation in all patients, (2) detected more frequently in some due to testing patterns, resulting in over-evaluation, or (3) so ubiquitous that the process of identifying and evaluating MH is an entirely stochastic process based on factors unrelated to the patient’s likelihood of having an underlying GU cancer or nephrologic disease. In the absence of a population-level screening study that assesses the prevalence of MH and the resulting incidence of urologic cancers found if each patient was evaluated with cystoscopy and imaging, we believe these results help frame the reported incidence of MH-related cancers as an overestimation. The percentage of patients who complete a urologic evaluation is but a small, select, and likely enriched proportion of individuals with MH and may not even reflect those at highest risk of having a urologic cancer.22-24 Extending this further and using a conservative microhematuria prevalence estimate from our results (30% of the health adult population over age 35), census estimates of roughly 120 million adults over age 35 in the United States, and assuming a liberal estimate of bladder cancer incidence among those with MH (~2%) this would imply that there should be ~720,000 new cases of bladder cancer each year detected as a result of MH evaluation alone. The current best estimate of yearly incident bladder cancer in the US is roughly 80,000 and while one could argue that population-level screening and improved early detection of bladder cancer would increase this number, this estimate is an order of magnitude higher than what the actual bladder cancer incidence is currently. Accordingly, our estimates should be used in future comparative effectiveness research to assess the value of MH evaluations at various thresholds, adjusted for testing frequency.
In this cohort of roughly 5.7 million patients seen for at least one outpatient visit in the VHA during a calendar year, 39% had a urinalysis ordered. The indication for each urinalysis was not available, but patients with other non-GU medical problems and urinary symptoms or complaints are more likely to receive a urinalysis. This may, in part, explain the high proportion of patients who had MH in our cohort as we were unable to determine symptomatology at the time of specimen collection. Patients complete urinalysis for myriad reasons, as dysfunction in nearly all organ systems can show itself in the urine. The frequency of dipstick-only urinalysis testing in our study may be explained by the pre-test indication for urine testing. For example, patients with diabetes or CKD are frequently monitored for proteinuria or glucosuria, and patients with liver disease may develop bilirubinuria. Conversely, urinalyses can also be ordered reflexively without high suspicion for underlying urinary tract disease and without a specific clinical question or concern. This overuse of urinalysis testing, in general, may result in overdetection of MH in patients who are unlikely to have any underlying urinary tract pathology, thus increasing cost and use of unnecessary diagnostic testing like cystoscopy, renal biopsy, and imaging without meaningful benefit in early diagnosis. Although we are unable to determine the appropriateness of urinalysis testing in our cohort, the sheer number of urinalyses ordered (both dipstick and microscopic) leads us to question the necessity of many of these tests.
Among all adult patients, we found a high prevalence (37%) of MH even with conservative exclusion criteria. This proportion may be influenced by detection bias, as it is impossible to know whether those who did not have a microscopic urinalysis had MH. However, in our study, urinalysis was more likely to be conducted for patients with risk factors for urologic cancer (older, male, smokers). Therefore, it is possible that since patients with bladder cancer risk factors and other comorbidities potentially related to microhematuria were more likely to complete a urinalysis that the true proportion of patients with microhematuria, if all completed a urinalysis, would be lower. Prior studies have estimated MH prevalence at 2.6%-54%8-10,25,26 and many of these were done with dipstick chemical urinalysis assays, one of our exclusion criteria, and not a microscopic approach. Notably, current dipstick technology has been found to have similar diagnostic accuracy when compared with microscopic urinalyses, albeit with limitations as a confirmatory diagnostic test.27 Variations in patient populations may also explain differences in our results; however, many of the historical studies included enriched cohorts of older, male patients who smoked and are at high risk for bladder cancer.
Our final objective was to identify factors associated with increased odds of having MH relative to the likelihood of a urinalysis being conducted. It is possible that the comorbidities associated with MH prevalence are also associated with urinalysis ordering practices, making it harder to assess the relationship between possible predictors and MH prevalence. Roughly 18% of patients had proteinuria, with increasing severity being associated with a higher likelihood of MH being present. Accordingly, if physicians have a greater propensity to order urinalyses for a given group of patients, such as those with renal disease, this would tend to lower the incidence of MH in that group. As a toy example to illustrate this point, imagine that a doctor has four patients, two with and two without kidney disease, where one of each pair has had a recent negative urinalysis. Normally, the doctor would only order a urinalysis for patients without recent testing, but in this case the doctor decides to do a repeat urinalysis for the patient with kidney disease to determine a change in proteinuria. If the previously negative patient with kidney disease remains negative and the untested patients are both positive, that would give an incidence of MH of 50% for patients with kidney disease and 100% for patients without, making it appear that the prevalence of MH was lower in the high-risk group. This may explain many of the findings in Table 4, where none of the odds ratios for diabetes, hypertension, and CKD are greater than 1, despite all being known causes of MH.
Female patients were more likely to complete an interpretable microscopic urinalysis and have MH. However, prior studies have shown that, despite this, female patients are less likely to complete a guideline-recommended evaluation for MH,19 suggesting a potentially significant disparity. Strict inclusion criteria in our study assessing ordered and positive urinalyses in the VA CDW may have led to an underestimation of MH prevalence by excluding patients with visible hematuria, prior urologic cancer diagnoses, or a higher prevalence of UTIs. However, accurate identification of true de novo MH, in the absence of a known benign or malignant cause and therefore requiring further evaluation, was required to strengthen the external validity of our results. Given the often-asymptomatic nature of MH, patients for whom a urinalysis was conducted may have initially sought care for other conditions. It is also possible that for patients with discordant IRR and ORs, more selective but less frequent testing was purposeful, thus selecting for a higher proportion of patients of MH based on clinical suspicion.
As expected, the demographic profile of patients receiving ambulatory care at VHA in 2015 matched the VA census population from 2015. However, the VA population tends to be older and more male than the general population, which should be considered when interpreting our findings. Nonetheless, the focus of this study, patients receiving ambulatory care within the VHA system, remains a significant strength. The VA is an equal-access system treating diverse patient populations, and previous studies have highlighted equal or improved treatment and screening within the VHA system compared to patients at other healthcare centers.28-30
CONCLUSION
The frequency of urinalysis testing in the VHA system is remarkably high. The detection of MH, which is relatively common, is influenced by the frequency of urinalysis testing, interpretability of microscopic RBC/hpf results, and varies among patients with different sociodemographic and clinical factors. Ordering practices for urinalyses may be at least partially responsible for whether MH is detected, but clear variation in the presence of MH existed even after adjustment for patient, comorbidity, and regional factors. Recommendations for the evaluation of MH should be revisited with consideration of how these newly identified factors may influence MH detection and the need for subsequent evaluation.
Supplementary Material
Acknowledgment
The authors thank Dr Yair Lotan for his input and feedback on the manuscript and study design.
Disclosures
This work was supported in part by the National Institutes of Health/National Cancer Institute (NIH/NCI) with a Cancer Center Support Grant to Memorial Sloan Kettering Cancer Center [P30 CA008748] and the 2021-2023 AUA/SUO Research Scholar Award (RSM).
Appendix A. Supporting Information
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.urology.2025.07.004.
Footnotes
CRediT Authorship Contribution Statement
Richard S. Matulewicz: Writing – original draft, Resources, Investigation, Formal analysis, Conceptualization, Writing – review & editing, Supervision, Methodology, Funding acquisition, Data curation. Samuel A. Gold: Writing – review & editing, Methodology, Formal analysis, Writing – original draft, Investigation. Fady Baky: Writing – review & editing, Investigation, Writing – original draft, Formal analysis. Andrew Nicholson: Writing – review & editing, Visualization, Software, Formal analysis, Writing – original draft, Validation, Methodology, Data curation. Eric Wahlstedt: Writing – review & editing, Investigation, Writing – original draft, Formal analysis. Patrick Alba: Investigation, Formal analysis. Bernard H. Bochner: Investigation, Supervision, Formal analysis. Harry W. Herr: Supervision, Formal analysis, Investigation. David S. Goldfarb: Investigation, Supervision, Formal analysis. Julie A. Lynch: Project administration, Formal analysis, Supervision, Investigation. Lamont J. Barlow: Project administration, Formal analysis, Supervision, Investigation. Melissa Assel: Validation, Methodology, Visualization, Software, Investigation, Conceptualization, Formal analysis. Andrew J. Vickers: Validation, Investigation, Conceptualization, Visualization, Methodology, Formal analysis. Scott E. Sherman: Project administration, Formal analysis, Supervision, Investigation. Danil V. Makarov: Supervision, Investigation, Formal analysis, Conceptualization, Resources, Funding acquisition, Data curation.
Ethical Declaration
The study was approved by the VA NY Harbor Healthcare System Institutional Review Board (NYHHS IRB # 1751).
Declaration of Competing Interest
The authors have no conflict of interest to declare.
Data Availability
Due to VA data privacy issues, we are unable to share the raw data.
References
- 1.Mariani AJ, Mariani MC, Macchioni C, et al. The significance of adult hematuria: 1,000 hematuria evaluations including a riskbenefit and cost-effectiveness analysis. J Urol. 1989;141:350. [DOI] [PubMed] [Google Scholar]
- 2.Topham PS, Harper SJ, Furness PN, et al. Glomerular disease as a cause of isolated microscopic haematuria. Q J Med. 1994;87:329. [PubMed] [Google Scholar]
- 3.Stapleton FB. Idiopathic hypercalciuria: association with isolated hematuria and risk for urolithiasis in children. The Southwest Pediatric Nephrology Study Group. Kidney Int. 1990;37:807. [DOI] [PubMed] [Google Scholar]
- 4.Bolenz C, Schroppel B, Eisenhardt A, et al. The investigation of hematuria. Dtsch Arztebl Int. 2018;115:801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Barocas DA, Boorjian SA, Alvarez RD, et al. Microhematuria: AUA/SUFU guideline. J Urol. 2020;204:778. [DOI] [PubMed] [Google Scholar]
- 6.Barocas DA, Lotan Y, Matulewicz RS, et al. Updates to microhematuria: AUA/SUFU guideline (2025). J Urol. 2025;213:547–557. [DOI] [PubMed] [Google Scholar]
- 7.Thompson IM. The evaluation of microscopic hematuria: a population-based study. J Urol. 1987;138:1189. [DOI] [PubMed] [Google Scholar]
- 8.Messing EM, Young TB, Hunt VB, et al. Home screening for hematuria: results of a multiclinic study. J Urol. 1992;148:289. [DOI] [PubMed] [Google Scholar]
- 9.Messing EM, Young TB, Hunt VB, et al. Urinary tract cancers found by homescreening with hematuria dipsticks in healthy men over 50 years of age. Cancer. 1989;64:2361. [DOI] [PubMed] [Google Scholar]
- 10.Britton JP, Dowell AC, Whelan P, Harris CM. A community study of bladder cancer screening by the detection of occult urinary bleeding. J Urol. 1992;148:788. [DOI] [PubMed] [Google Scholar]
- 11.Bangma CH, Loeb S, Busstra M, et al. Outcomes of a bladder cancer screening program using home hematuria testing and molecular markers. Eur Urol. 2013;64:41. [DOI] [PubMed] [Google Scholar]
- 12.Loo RK, Lieberman SF, Slezak JM, et al. Stratifying risk of urinary tract malignant tumors in patients with asymptomatic microscopic hematuria. Mayo Clin Proc. 2013;88:129. [DOI] [PubMed] [Google Scholar]
- 13.Froom P, Shimoni Z, Dusseldorp N, Benbassat J. Asymptomatic microscopic hematuria in inpatient nonsurgical adults. Am J Clin Pathol. 2023;159:221. [DOI] [PubMed] [Google Scholar]
- 14.Ghaferi AA, Schwartz TA, Pawlik TM. STROBE reporting guidelines for observational studies. JAMA Surg. 2021;156:577. [DOI] [PubMed] [Google Scholar]
- 15.VA’s National Laboratory Test (NLT) List, Appendix B of the Laboratory Emerging Pathogens Initiative (EPI) Version 5.2 Roll Up Modifications Technical and User Manual: National Laboratory Test 2015. [Google Scholar]
- 16.Barocas DA, Boorjian SA, Alvarez RD, et al. Microhematuria: AUA/SUFU guideline. J Urol. 2020;204:778. [DOI] [PubMed] [Google Scholar]
- 17.Song RJ, Ho Y-L, Nguyen X-MT, et al. Development of an electronic health record-based algorithm for smoking status using the Million Veteran Program (MVP) Cohort survey response. Circulation. 2016;134:A18809. [Google Scholar]
- 18.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int. 2024;105:S117–S314. [DOI] [PubMed] [Google Scholar]
- 19.Matulewicz RS, Demzik AL, DeLancey JO, et al. Disparities in the diagnostic evaluation of microhematuria and implications for the detection of urologic malignancy. .e1 Urol Oncol. 2019;37:300. [DOI] [PubMed] [Google Scholar]
- 20.Han DS, Zhou W, Seigne JD, et al. Geographic variation in cystoscopy rates for suspected bladder cancer between female and male medicare beneficiaries. Urology. 2018;122:83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ark JT, Alvarez JR, Koyama T, et al. Variation in the diagnostic evaluation among persons with hematuria: influence of gender, race and risk factors for bladder cancer. J Urol. 2017;198:1033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Matulewicz RS, Rademaker A, Meeks JJ. A simplified nomogram to assess risk of bladder cancer in patients with a new diagnosis of microscopic hematuria. Urol Oncol. 2020;38:240–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Waisbrod S, Natsos A, Wettstein MS, et al. Assessment of diagnostic yield of cystoscopy and computed tomographic urography for urinary tract cancers in patients evaluated for microhematuria: a systematic review and meta-analysis. JAMA Netw Open. 2021;4:e218409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Richards KA, Ruiz VL, Murphy DR, et al. Diagnostic evaluation of patients presenting with hematuria: an electronic health record-based study. Urol Oncol. 2018;36:88.e19. [DOI] [PubMed] [Google Scholar]
- 25.Messing EM, Young TB, Hunt VB, et al. The significance of asymptomatic microhematuria in men 50 or more years old: findings of a home screening study using urinary dipsticks. J Urol. 1987;137:919. [DOI] [PubMed] [Google Scholar]
- 26.Messing EM, Young TB, Hunt VB, et al. Comparison of bladder cancer outcome in men undergoing hematuria home screening versus those with standard clinical presentations. Urology. 1995;45:387. [DOI] [PubMed] [Google Scholar]
- 27.Matulewicz RS, DeLancey JO, Pavey E, et al. Dipstick urinalysis as a test for microhematuria and occult bladder cancer. Bladder Cancer. 2017;3:45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Trivedi AN, Grebla RC. Quality and equity of care in the veterans affairs health-care system and in medicare advantage health plans. Med Care. 2011;49:560. [DOI] [PubMed] [Google Scholar]
- 29.Rasmussen KM, Patil V, Li C, et al. Survival outcomes by race and ethnicity in veterans with nonmetastatic castration-resistant prostate cancer. JAMA Netw Open. 2023;6:e2337272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Riviere P, Luterstein E, Kumar A, et al. Survival of African American and non-Hispanic white men with prostate cancer in an equal-access health care system. Cancer. 2020;126:1683. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Due to VA data privacy issues, we are unable to share the raw data.