Abstract
Background
The urine dipstick is widely used to screen for urinary tract infection (UTI) in daily practice. However, in older adults, especially women, its reliability is limited by the high prevalence of asymptomatic bacteriuria and leukocyturia. Recently, a consensus-based reference standard was introduced to classify patients based on the likelihood of UTI. Using this standard, we assessed the diagnostic accuracy of the urine dipstick in older adults.
Methods
We performed an observational study using prospectively collected data from two ongoing studies. We included adults aged ≥ 60 years with clinically suspected UTI who underwent dipstick testing, urine flow cytometry, and urine culture during the same episode. The Roche COMBUR-7 urine dipstick served as the index test, and episodes were classified using the consensus-based reference standard. We evaluated all possible dipstick thresholds and their combinations, calculating sensitivity, specificity, predictive values, and likelihood ratios using Generalized Estimating Equations to account for multiple episodes per participant. Receiver operating characteristic analysis was performed to determine the optimal balance between sensitivity and specificity.
Results
We included 368 episodes from 199 participants (mean age 73.8 years; 77.9% women). Dipstick sensitivity was 93.6% (95% CI 90.1%–95.9%) and specificity was 58.6% (95% CI 45.4%–70.6%) for the conventional threshold of nitrite + and/or leukocyte esterase ≥ 1 + . A higher threshold, “total of ≥ 2 + ” (leukocyte esterase ≥ 2 + or nitrite positive with leukocyte esterase ≥ 1 +) improved performance, yielding a sensitivity of 82.2% (95% CI 77.0%-86.4%) and a specificity of 75.7% (95% CI 61.6%-85.8%). With this threshold, the positive likelihood ratio improved from 2.3 (95% CI 1.76–3.06) to 3.4 (95% CI 2.0–6.1).
Conclusions
In older adults with suspected UTI, urine dipstick testing demonstrates limited diagnostic performance. Current thresholds yield high sensitivity but poor specificity, which may contribute to unnecessary antimicrobial treatment. Given the high prevalence of asymptomatic bacteriuria in older adults, a higher threshold of “total of ≥ 2 + ” (either LE ≥ 2 + , or nitrite + combined with LE ≥ 1 +) may improve diagnostic accuracy and be considered when assessing patients ≥ 60 years old for antibiotic therapy. Prospective studies are needed to confirm this recommendation.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12877-026-07741-y.
Keywords: Urinary Tract Infection, Older adults, Dipstick, Diagnostic stewardship, Diagnostic accuracy
Background
Urinary tract infection (UTI) is a major cause of antibiotic prescriptions in the older population [1, 2]. Antibiotic treatment for UTI reduces symptoms and helps prevent serious complications. However, frequent antibiotic treatment leads to the selection of resistant strains, increases healthcare costs, and exposes people to potential adverse effects. These risks are particularly relevant in older adults, who are more vulnerable to adverse drug reactions, drug–drug interactions, and complications such as Clostridioides difficile infections [3]. Therefore, it is crucial to accurately diagnose UTI in older adults and to avoid unnecessary antibiotic treatment. This is particularly challenging in older adults, in whom a higher prevalence of cognitive disorders may limit the reliability of the medical history. In addition, other age-associated urological conditions, such as genitourinary syndrome of menopause, cystocele, or prostate hypertrophy, can mimic the symptoms of UTI [4].
In clinical practice, the presence of pyuria or bacteriuria is often used as an indicator of UTI. However, asymptomatic bacteriuria (ASB) and leukocyturia are extremely common in older adults, particularly women, with reported ASB in up to 59% of women. Among those with ASB, 90% have leukocyturia as well [5–10]. As a result, urine tests that solely rely on these markers, such as the urine dipstick, may perform poorly in distinguishing true infection from colonization, further raising questions about the dipstick’s clinical value in routine care for older adults.
The urine dipstick is a rapid and inexpensive screening tool that detects nitrite, produced by most Enterobacterales, and leukocyte esterase (LE), an enzyme released by neutrophils [11]. Although appealing for its speed and simplicity, its diagnostic accuracy in older adults is uncertain. Most previous studies have compared dipstick results with urine cultures as the reference standard; however, culture alone cannot reliably distinguish ASB from a true infection, leading to potential misclassification [12, 13]. Moreover, studies that have applied more comprehensive diagnostic criteria rarely included older populations specifically [14–16].
To address this gap, a Delphi consensus-based reference standard for UTI diagnosis was recently developed to minimize misclassification bias [17] (Additional file 1). Using this standard, the present study re-evaluates the diagnostic accuracy of the urine dipstick in older adults and explores whether adjusting the dipstick threshold could improve its clinical performance.
Methods
This single-center observational cohort study was conducted at Leiden University Medical Center, the Netherlands. Data were collected between April 2024 and December 2025.
Participants
For the current analysis, we included consecutive adults aged ≥ 60 years from the two prospective studies described below with a clinical suspicion of UTI who had, during the same episode, a urine dipstick, urine flow cytometry, and urine culture performed, along with documentation of urinary and systemic symptoms. All participants provided written informed consent for secondary use of their study data. Participants with an indwelling urinary catheter or those unable to provide informed consent were excluded. A single participant could contribute multiple episodes, provided these represented distinct clinical events, defined by complete resolution of symptoms after the previous episode.
Recruitment and data sources
Participants were selected from two clinical studies conducted at the Leiden University Medical Center: the E.mbrace study (NCT04899336) and the UTI-GOLD study (NCT06610721). If participants presented with symptoms and underwent urine dipstick testing, urine flow cytometry, and urine culture (standard diagnostic work-up), they were consecutively included in the study.
The E.mbrace study is a phase III randomized controlled trial evaluating a vaccine against invasive Escherichia coli disease. Participants aged ≥ 60 years were recruited through local advertisement and general practitioners. Inclusion required at least one UTI episode in the previous two years and a stable medical condition. Individuals with malignancies or severe immunodeficiencies were excluded.
The UTI-GOLD study [18] is a diagnostic accuracy study investigating urinary biomarkers for UTI. Participants aged ≥ 65 years were recruited from emergency departments, long-term care facilities, primary care offices, and outpatient clinics.
Procedures and data collection
Participants (or their caregivers) were interviewed by a research nurse or physician using a structured case report form, which systematically captured lower urinary tract symptoms (dysuria, frequency, urgency, suprapubic pain), systemic symptoms (fever, rigors, malaise), and other abdominal or non-specific symptoms (e.g., nausea, vomiting, confusion, falls). A physical examination was performed by a physician when systemic signs or symptoms were present. Participants were instructed to provide a midstream urine sample in a sterile container. If a voided sample could not be obtained, a sample was collected by single catheterization. Each urine specimen was divided into three aliquots: one for dipstick testing, one for flow cytometry analysis, and one for culture. Dipstick testing was performed immediately by a research nurse or study physician in accordance with the manufacturer’s instructions. After dipstick testing, the urine was immediately transported to the laboratory and processed on arrival. Laboratory staff performing urine flow cytometry and culture were blinded to the dipstick results. Research staff performing the dipstick were not informed of flow cytometry or culture results until the dipstick results had been recorded. As the reference standard required culture results, its application was performed after all data were available. Consequently, researchers applying the reference standard were not blinded to the index test results; however, as the reference standard followed a predefined algorithm, the risk of interpretation bias was minimized. All data were extracted from electronic patient records and stored in a CASTOR electronic database. Patient information was kept in a secure datasafe, accessible only to members of the research team.
Index test and laboratory analysis
The Roche Combur-7 Test (Roche Diagnostics, Almere, the Netherlands) was used as the dipstick, which provides categorical results for nitrite (positive or negative) and LE (negative, 1 +, 2 +, or 3 +). A positive dipstick was defined as a positive nitrite and/or at least 1 + LE, which is consistent with previous studies on this topic [4, 8, 12, 19, 20]. Leukocyturia was quantified by automated urine flow cytometry using the UF-4000 analyzer (Sysmex, Kobe, Japan), with a lower limit of detection of 1 leukocyte/µL and an upper limit of detection of 10.000 leukocytes/µL. Urine samples were cultured using agar plates, and bacterial growth was quantified as colony-forming units per milliliter (CFU/mL).
Reference standard for UTI
A predefined Delphi consensus-based reference standard for UTI was applied in this study [17]. Urine culture alone is insufficient in this population due to the high prevalence of ASB, which may lead to misclassification of colonization as infection. Therefore, we used a consensus-based reference standard that incorporates clinical features, urine flow cytometry, and culture results. This approach was specifically developed for research purposes to reduce misclassification and provide a more clinically meaningful classification of UTI episodes. The reference standard consists of four domains: symptoms and signs, systemic criteria, urine flow cytometry results, and urine culture results, yielding a total score ranging from 0 to 11. After all data were available, each episode was classified using this predefined algorithm. Episodes with a total score ≥ 5 were classified as “probable” or “definite” UTI and were considered true infections for the purpose of this analysis.
Statistical analysis
Statistical analysis was performed using SPSS IBM (version 30). A generalized estimating equation (GEE) model was applied to account for clustered data arising from multiple samples obtained from the same patient (with patient ID as the clustering variable). Estimated coefficients were used to calculate predictive probabilities, which were then used to derive sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive and negative likelihood ratios (LR + and LR-), and their associated 95% confidence intervals (95% CI). All analyses were repeated in prespecified subgroups comprising female participants and first UTI episodes. Diagnostic properties were calculated for all possible dipstick thresholds using GEE. Receiver operating characteristic (ROC) analysis was performed to determine the optimal trade-off between sensitivity and specificity. The optimal cut-off value was defined as the ROC-curve coordinate with the shortest Euclidean distance to the point (0.1). To evaluate the relationship between absolute pyuria and LE dipstick results, the median and interquartile range of urinary leukocyte counts were calculated for each LE dipstick category.
Missing data were not imputed. Analyses were performed using available data (complete case analysis). No formal sample size calculation was performed. The sample size was determined by the number of eligible participants included in the two parent studies during the study period. Participants were included if complete diagnostic data (dipstick, urine flow cytometry, and urine culture) and symptom documentation were available for a given episode.
Results
Cohort characteristics
A total of 199 participants met the inclusion criteria. Of these participants, 129 (65%) were enrolled from the EMBRACE study and 70 (35%) from the UTI-GOLD study. The mean age was 73.8 years, and 77.9% were women. More than half of the participants (53.8%) reported a history of recurrent UTI (≥ 2 UTI episodes in the last 6 months or ≥ 3 UTI episodes per year), and 25.1% had undergone prior urological surgery, most commonly transurethral resection of the prostate. Most participants were recruited through their general practitioner (64.3%) (Table 1).
Table 1.
Baseline characteristics (n = 199)
| Mean (SD) | Missing (%) | ||
|---|---|---|---|
| Age, years | 73.8 (7.3) | 0 | |
| UTI episodes per patient | 2.0 (1.6) | 0 | |
| N (%) | Missing (%) | ||
| Sex | 0 | ||
| Female | 155 (77.9) | ||
| Male | 44 (22.1) | ||
| Setting | 0 | ||
| General practice | 128 (64.3) | ||
| Outpatient clinic | 34 (17.1) | ||
| Emergency department | 24 (12.1) | ||
| Long term care facility | 13 (6.5) | ||
| Yes (%) | No (%) | Missing (%) | |
| Urological comorbidities | |||
| History of urological surgery | 50 (25.1) | 148 (74.4) | 1 (0.5) |
| Urological malignancy | 9 (4.5) | 190 (95.5) | 0 |
| Recurrent UTI* | 107 (53.8) | 88 (44.2) | 4 (2.0) |
| Urinary incontinence | 30 (15.1) | 169 (84.9) | 0 |
| Intermittent catheterization | 15 (7.5) | 184 (92.5) | 0 |
| Pelvic organ prolapse | 7 (3.5) | 192 (96.5) | 0 |
| BPH | 2 (1.0) | 197 (99) | 0 |
| Other comorbidities | |||
| Diabetes mellitus | 30 (15.1) | 168 (84.4) | 1 (0,5) |
| Hypertension | 75 (37.7) | 124 (62.3) | 0 |
| History of chronic kidney disease | 21 (10.6) | 178 (89.4) | 0 |
| Immunosuppressive therapy ** | 14 (7.0) | 185 (93) | 0 |
| Mean (SD) | Missing (%) | ||
| Charlson Comorbidity Index Score | 3.6 (1.5) | 0 | |
BPH Benign Prostate Hypertrophy, SD standard deviation
Legend:
* ≥ 2 UTI episodes in the last 6 months or ≥ 3 UTI episodes per year
** Within one week before urine sample collection
Episode characteristics
A total of 368 episodes with urine dipsticks were included. According to the consensus-based reference standard, 298 episodes (81.0%) were classified as “true UTI” (“probable” or “definite”). The most common urinary tract symptoms were dysuria (56.0%), frequency (53.5%), and urgency (44.3%).. A total of 299 episodes (81.3%) had a positive urine culture. Of these, Escherichia coli was the most frequently cultured pathogen (51.9%) (Table 2).
Table 2.
Characteristics of individual UTI episodes (n = 368)
| Mean (SD) | Missing (%) | ||
|---|---|---|---|
| Symptom duration, days | 3.9 (5.9) | 62 (16.9) | |
| N (%) | Missing (%) | ||
| Sex | 0 | ||
| Female | 310 (84.2%) | ||
| Male | 58 (15.8) | ||
| Yes (%) | No (%) | Missing (%) | |
| Pretreated with antibiotics | 51 (13.9) | 317 (86.1) | 0 |
| Symptoms | |||
| Urgency | 163 (44.4) | 204 (55.3) | 1 (0.3) |
| Frequency | 197 (53.5) | 170 (46.2) | 1 (0.3) |
| Dysuria | 206 (56.0) | 161 (43.7) | 1 (0.3) |
| Cloudy urine | 125 (34.0) | 232 (63) | 11 (3.0) |
| Suprapubic pain | 137 (37.2) | 230 (62.5) | 1 (0.3) |
| Flank pain | 35 (9.5) | 331 (90) | 2 (0.5) |
| Fever | 18 (4.9) | 346 (94) | 4 (1.1) |
| N (%) | Missing (%) | ||
| Pathogens | 0 | ||
| Escherichia coli | 191 (51.9) | ||
| Klebsiella species | 36 (9.8) | ||
| Enterococcus species | 17 (4.6) | ||
| Proteus species | 5 (1.4) | ||
| Streptococcus species | 7 (1.9) | ||
| Mixed flora | 19 (5.2) | ||
| Other species | 24 (6.5) | ||
| No bacterial growth | 69 (18.8) | ||
| Reference standard score | 0 | ||
| No UTI (0–2 points) | 40 (10.9) | ||
| Possible UTI (3–4 points) | 30 (8.2) | ||
| Probable UTI (5–7 points) | 148 (40.2) | ||
| Definite UTI (8–11 points) | 150 (40.8) | ||
Legend:
Other species = all cultured pathogens not listed above
SD standard deviation
Diagnostic accuracy of the dipstick
Using the standard definition of a positive dipstick (nitrite positive and/or LE ≥ 1 +), dipsticks were positive in 308 of 368 episodes (83.7%). Of these, 279 represented a true UTI according to the reference standard. The GEE analysis yielded a dipstick sensitivity of 93.6% (95% CI 90.1%–95.9%) and a specificity of 58.6% (95% CI 45.4%–70.6%). This corresponded to a positive likelihood ratio of 2.26 (Table 3).
Table 3.
Diagnostic accuracy of urine dipstick and distribution of dipstick results
| UTI* | No UTI** | Total (%) | |||
|---|---|---|---|---|---|
|
Positive dipstick |
279 | 29 | 308 (83.7) | ||
| Negative dipstick | 19 | 41 | 60 (16.3) | ||
| Total (%) | 298 (81.0) | 70 (19.0) | 368 (100) |
| Sensitivity (%) (CI) |
Specificity (%) (CI) |
PPV (%) (CI) |
NPV (%) (CI) |
LR + (CI) | LR—(CI) |
|---|---|---|---|---|---|
|
93.6 (90.1—95.9) |
58.6 (45.4—70.6) |
90.6 (86.5—93.5) |
68.3 (54.4—79.6) |
2.26 (1.76—3.06) |
0.11 (0.06—0.22) |
Positive dipstick = nitrite positive and/or LE ≥ + 1
CI 95% Confidence Interval, NPV Negative Predictive Value, PPV Positive Predictive Value, LR + Positive Likelihood Ratio, LR Negative Likelihood Ratio
Legend: * Reference standard score ≥, ** Reference standard score < 5
Optimal dipstick threshold
All possible dipstick thresholds were evaluated to identify the cut-off that provided the optimal balance between sensitivity and specificity (Fig. 1 and Additional file 2). The ROC curves indicated that the LE subgroup (1 +, 2 +, and 3 +) demonstrated the highest overall diagnostic performance with an approximate AUC of 0.846 (95% CI 0.795–0.897; p < 0.001), with the “total” subgroup (total 2 + and total 3 +) showing slightly lower performance with an approximate AUC of 0.833 (95% CI 0.783–0.883; p < 0.001). The “Total of 2 + ” threshold had the smallest Euclidean distance to the upper left corner of the ROC space(0,1), suggesting the most optimal balance between sensitivity and specificity (d = 0.301) (Additional file 3). At this threshold, 262 dipsticks (71.2%) were positive, corresponding to a sensitivity of 82.2% (95% CI 77.0%—86.4%) and a specificity of 75.7% (95% CI 61.6%—85.8%). With this threshold the positive likelihood ratio improved to 3.38 (95% CI 2.0–6.1) with a negative likelihood ratio of 0.24 (95% CI 0.22–0.27).
Fig. 1.
Diagnostic properties of dipstick thresholds with 95% confidence intervals
Legend: Sensitivity and specificity for different urine dipstick cut-offs. LE = leukocyte esterase. * = conventional threshold. “Total ≥ 2 + ” and “Total ≥ 3 + ” represent combined dipstick scores calculated by summing the leukocyte esterase (LE) level (0–3 +) and the nitrite result (0 or 1 +). A threshold is met when the sum is equal to or exceeds the specified value, whether driven by LE alone or by any combination of LE and nitrite (e.g., LE 1 + + nitrite 1 + = 2 +; LE 2 + + nitrite 1 + = 3 +)
Urine leukocyte counts per dipstick category
Median urine leukocyte counts increased stepwise with higher LE dipstick categories, from 18 cells/µL for negative LE to 3,798.5 cells/µL for LE 3 +. However, leukocyte counts showed substantial variability within each dipstick category, with overlapping distributions across categories (Fig. 2 and Additional file 4).
Fig. 2.
Boxplot of leukocyte counts (cells/µL) by dipstick leukocyte esterase score
Legend: Median and interquartile ranges of leukocyte counts per dipstick leukocyte esterase result
Subgroup analysis
Subgroup analysis in the female population showed a sensitivity of 93.4% (95% CI 89.5%–95.9%) and a specificity of 52.8% (95% CI 37.1%–68.0%). When considering only the first UTI episode, sensitivity was 92.5% (95% CI 87.3%–95.7%) and specificity was 61.5% (95% CI 45.6%–75.3%). All data from these subgroup analyses are presented in Additional files 5 and 6.
Discussion
Summary of findings
In this study, we evaluated the diagnostic accuracy of the urine dipstick for UTI in older adults using a consensus-based reference standard. At the conventional threshold (nitrite positive and/or LE ≥ 1 +) the dipstick showed a high sensitivity of 93.6% (95% CI 90.1%−95.9%) but a poor specificity of 58.6% (95% CI 45.4%−70.6%). Raising the threshold to a “total of ≥ 2 + ” (defined as either LE ≥ 2 +, or nitrite positive combined with LE ≥ 1 +) improved specificity to 75.7% (95% CI 61.1%−85.8%), with a moderate reduction in sensitivity to 82.2% (95% CI 77.0%−86.4%). Overall, these findings indicate that the urine dipstick has limited discriminative ability for diagnosing UTI in older adults.
Interpretation and comparison with previous studies
This is the first study to evaluate the diagnostic performance of urine dipstick testing for UTI in older adults from a predominantly general practice setting. Our findings reinforce and extend existing concerns about the clinical usefulness of urine dipstick testing in this population.. Although “Total of 2 + ” demonstrated the most favorable balance of sensitivity and specificity, differences in ROC curve results across multiple thresholds were modest. The relatively wide confidence intervals, likely due to the limited sample size, indicate uncertainty in identifying an optimal cutoff. Therefore, this threshold should be considered approximate rather than definitive.
The high sensitivity but low specificity for the conventional threshold of nitrite + and/or LE ≥ 1 + observed in this study mirrors results from nursing home and hospital-based studies, suggesting that the generation of many false-positive dipstick results is not confined to institutionalized patients but occur broadly among older adults.
Previous studies using less comprehensive reference standards have focused on younger populations or exclusively on nursing home or hospital settings [14–16, 21–23]. Despite differences in study populations, the diagnostic properties of the urine dipstick observed in these studies are consistent with our findings. In studies where the dipstick was compared to a standardized reference standard, specificity did not exceed 65%, with Latour et al. [14] reporting a specificity as low as 20.7% (95% CI 14.3%−28.9%) for the conventional threshold (nitrite + and/or LE ≥ 1 +) in nursing home residents. In contrast, sensitivity was consistently high, exceeding 85% across studies. The differences between studies may be explained by the reference standards used. Latour et al. [14] confirmed only 11.7% of cases using a consensus-based reference standard, compared with 84.5% in our study. Additionally, our study population mainly consisted of individuals with a history of recurrent UTI, which may have resulted in a different clinical spectrum and thereby influenced diagnostic performance.
Moragas et al. [24] conducted a meta-analysis that showed a pooled specificity of less than 40% (95% CI 19%−62%) for the dipstick in predicting bacteriuria among institutionalized patients. Although the studies included in this meta-analysis used bacteriuria as the outcome rather than a clinical definition of UTI, this low specificity in symptomatic patients is consistent with the findings of the studies discussed above and with our current findings. Importantly, our study provides evidence that the remarkably low specificity observed is not limited to clinical or long-term care settings, but extends to the older population more broadly.
Hertz et al. [15] evaluated the diagnostic accuracy of similar dipstick thresholds in men and women (median age 76 years) presenting to the emergency department with suspected infection. The diagnosis of UTI was adjudicated retrospectively by an expert panel and served as the reference standard. They observed that thresholds of LE ≥ 3 or the combination of nitrite and LE ≥ 1 provided the highest proportion of correctly classified episodes (true positives/true negatives). Despite methodological differences, both we and Hertz et al. [14] identified substantially higher optimal thresholds than those currently used in clinical practice. Although LE ≥ 3 demonstrated high specificity in our cohort of predominantly community-dwelling older adults with suspected UTI, sensitivity at this threshold was limited (48.3%, 95% CI 41.9%−54.8%). The generalizability of the observed specificity warrants cautious interpretation. Our study included relatively few long-term care residents and likely reflects a population with lower background rates of asymptomatic bacteriuria compared with those of more frail or institutionalized older adults. In settings with a higher prevalence of asymptomatic bacteriuria, specificity may be reduced, increasing the risk of false-positive results and overtreatment. At present, we do not support the use of LE ≥ 3 as a standalone rule-in test for UTI in older adults.
Previous research has shown that the degree of pyuria can help distinguish asymptomatic bacteriuria from true UTI in older women, emphasizing the importance of accurate quantification [5]. In contrast, our findings demonstrate that the urine dipstick cannot provide such quantitative discrimination: even high LE categories corresponded to a wide range of leukocyte counts, limiting its utility for clinical decision-making. This was previously demonstrated by van den Broek et al. [25]. Although no absolute leukocyte counts were provided, the box and-whisker plot presented in that study shows patterns similar to those observed here. Median leukocyte concentrations increased stepwise across dipstick LE categories, but with substantial overlap, supporting the interpretation that dipstick LE results provide only semiquantitative reflections of pyuria rather than reliable indicators of infection.
Taken together, these findings suggest that the limited specificity of urine dipstick testing is a consistent finding across settings, while our study adds evidence from a predominantly community-based older population using a clinically oriented reference standard.
Strengths and limitations
A key strength of this study is its focus on community-dwelling older adults, a population in which dipsticks are frequently used due to limited access to immediate laboratory testing. The use of a Delphi consensus-based reference standard minimized misclassification related to asymptomatic bacteriuria, enhancing diagnostic validity compared with studies relying solely on culture results [17]. Another important strength is the completeness of clinical data; because participants were drawn from two prospective studies, detailed information on UTI-related symptoms was systematically collected, resulting in minimal missing data and allowing the reliable application of the reference standard.
Several potential sources of bias should be considered. First, selection bias may be present, as inclusion required availability of complete diagnostic data, and not all eligible patients underwent dipstick testing due to logistical reasons. However, as this was unrelated to patient characteristics or disease severity, the impact on the results is likely limited. In addition, a substantial proportion of participants had a history of recurrent UTI and were recruited through ongoing studies, which may further limit generalizability. Second, the high prevalence of UTI in our study population (81%) suggests a potential spectrum effect, which may have influenced the observed diagnostic performance, particularly specificity and predictive values. Third, all included episodes underwent the same diagnostic work-up, including both the index test and the reference standard, making classical verification bias unlikely; however, the requirement for complete testing may have contributed to selection effects. Fourth, measurement bias may have occurred due to inter-observer variability in the interpretation of the dipstick results, as different researchers were responsible for reading the test. Although standardized procedures and interpretation guidelines were applied, some variability between observers cannot be completely excluded. Finally, participants in the E.mbrace study received either a vaccination against invasive Escherichia coli disease (IED) or a placebo, which could theoretically have influenced leukocyte responses, although a meaningful impact on local bladder leukocyte response is considered unlikely.
Conclusion and implications
In older adults with suspected UTI, urine dipstick components show markedly different diagnostic performance: current thresholds provide high sensitivity but poor specificity, which may contribute to inappropriate antimicrobial treatment. In the context of a relatively high prevalence of baseline asymptomatic bacteriuria in older adults, and a significant potential for deleterious effects of unnecessary antibiotic use, our findings suggest that a threshold of “total of ≥ 2 + ” (defined as either LE ≥ 2 +, or nitrite positive combined with LE ≥ 1 +) could improve overall diagnostic accuracy and may be considered when assessing patients ≥ 60 years old for antibiotic therapy. These results should be interpreted as preliminary and warrant confirmation in larger, prospective studies before being used to guide clinical decision-making.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- ASB
Asymptomatic bacteriuria
- GEE
Generalized Estimating Equations
- IED
Invasive E.coli disease
- LE
Leukocyte esterase
- ROC
Receiver Operating Characteristic
- UTI
Urinary tract infection
Authors’ contributions
AB: Investigation, Resources, Supervision, Writing – Original Draft, Writing – Review and Editing. CO: Investigation, Formal Analysis, Writing – Original Draft. RG: Investigation. NEM: Investigation. JG: Formal analysis. LV, MB, LC, AH: Writing-Review and Editing. ML: Conceptualization, Methodology, Resources, Supervision, Writing – Review and Editing.
Funding
No funding was provided for this study.
Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by the Research committee of the Leiden University Center for Infectious Disease (LUCID) on 28 February and by the non-WMO commission of the Leiden University Medical Center (reference number: 2025014). This study was conducted in accordance with the 1964 declaration of Helsinki and its subsequent amendments [26].
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Annelies M. Baart and Carmen I. Oosterkamp contributed equally to this work.
References
- 1.Ahmed H, Farewell D, Jones HM, Francis NA, Paranjothy S, Butler CC. Incidence and antibiotic prescribing for clinically diagnosed urinary tract infection in older adults in UK primary care, 2004-2014. PLoS ONE. 2018;13(1):e0190521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Holm A, Cordoba G, Aabenhus R. Prescription of antibiotics for urinary tract infection in general practice in Denmark. Scand J Prim Health Care. 2019;37(1):83–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ge IY, Fevrier HB, Conell C, Kheraj MN, Flint AC, Smith DS, et al. Reducing risk of Clostridium difficile infection and overall use of antibiotic in the outpatient treatment of urinary tract infection. Ther Adv Urol. 2018;10(10):283–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gbinigie OA, Ordonez-Mena JM, Fanshawe TR, Pluddemann A, Heneghan C. Diagnostic value of symptoms and signs for identifying urinary tract infection in older adult outpatients: systematic review and meta-analysis. J Infect. 2018;77(5):379–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bilsen MP, Aantjes MJ, van Andel E, Stalenhoef JE, van Nieuwkoop C, Leyten EMS, et al. Current Pyuria Cutoffs Promote Inappropriate Urinary Tract Infection Diagnosis in Older Women. Clin Infect Dis. 2023;76(12):2070–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rodhe N, Lofgren S, Matussek A, Andre M, Englund L, Kuhn I, et al. Asymptomatic bacteriuria in the elderly: high prevalence and high turnover of strains. Scand J Infect Dis. 2008;40(10):804–10. [DOI] [PubMed] [Google Scholar]
- 7.Ouslander JG, Schapira M, Fingold S, Schnelle J. Accuracy of rapid urine screening tests among incontinent nursing home residents with asymptomatic bacteriuria. J Am Geriatr Soc. 1995;43(7):772–5. [DOI] [PubMed] [Google Scholar]
- 8.Biggel M, Heytens S, Latour K, Bruyndonckx R, Goossens H, Moons P. Asymptomatic bacteriuria in older adults: the most fragile women are prone to long-term colonization. BMC Geriatr. 2019;19(1):170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang RC, Schneeberger C, Halonen K, Smalbrugge M, Geerlings SE, de Greeff SC, et al. Bacteriuria in Nursing Home Residents in the Netherlands: Point Prevalence and Associated Factors. J Am Med Dir Assoc. 2026;27(3):106070. [DOI] [PubMed] [Google Scholar]
- 10.Boscia JA, Abrutyn E, Levison ME, Pitsakis PG, Kaye D. Pyuria and asymptomatic bacteriuria in elderly ambulatory women. Ann Intern Med. 1989;110(5):404–5. [DOI] [PubMed] [Google Scholar]
- 11.Mundt LA. Chapter 4: chemical analysis of urine. In: Graff’s textbook of routine urinalysis and body fluids, 2nd ed. Wolters-Kluwer / Lippincott Williams & Wilkins, Philadelphia, PA. 2010. p. 49–50.
- 12.Bafna P, Deepanjali S, Mandal J, Balamurugan N, Swaminathan RP, Kadhiravan T. Reevaluating the true diagnostic accuracy of dipstick tests to diagnose urinary tract infection using Bayesian latent class analysis. PLoS ONE. 2020;15(12):e0244870. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Werter DE, Schneeberger C, Geerlings SE, de Groot CJM, Pajkrt E, Kazemier BM. Diagnostic accuracy of urine dipsticks for urinary tract infection diagnosis during pregnancy: a retrospective cohort study. Antibiotics (Basel). 2024;13(6):567. [DOI] [PMC free article] [PubMed]
- 14.Latour K, De Lepeleire J, Catry B, Buntinx F. Nursing home residents with suspected urinary tract infections: a diagnostic accuracy study. BMC Geriatr. 2022;22(1):187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hertz MA, Johansen IS, Rosenvinge FS, Brasen CL, Andersen ES, Ostergaard C, et al. Urine flow cytometry and dipstick analysis in diagnosing bacteriuria and urinary tract infections among adults in the emergency department-a diagnostic accuracy trial. Diagnostics (Basel). 2024;14(4):447. [DOI] [PMC free article] [PubMed]
- 16.Kristensen LH, Winther R, Colding-Jorgensen JT, Pottegard A, Nielsen H, Bodilsen J. Diagnostic accuracy of dipsticks for urinary tract infections in acutely hospitalised patients: a prospective population-based observational cohort study. BMJ Evid Based Med. 2024;29(1):1–8. [DOI] [PMC free article] [PubMed]
- 17.Bilsen MP, Conroy SP, Schneeberger C, Platteel TN, van Nieuwkoop C, Mody L, et al. A reference standard for urinary tract infection research: a multidisciplinary Delphi consensus study. Lancet Infect Dis. 2024;24(8):e513–21. [DOI] [PubMed] [Google Scholar]
- 18.El Moussaoui N, van Andel E, van der Beek MT, Vlot JA, Bilsen MP, van Nieuwkoop C, et al. Validation of urinary biomarkers for accurate diagnosis of urinary tract infections in older adults across primary care, hospitals and long-term care facilities in the Netherlands and UK (UTI-GOLD): a multicentre observational study protocol. BMJ Open. 2025;15(6):e103311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bosmans JE, Coupe VMH, Knottnerus BJ, Geerlings SE, Moll van Charante EP, Ter Riet G. Cost-effectiveness of different strategies for diagnosis of uncomplicated urinary tract infections in women presenting in primary care. PLoS ONE. 2017;12(11):e0188818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fenwick EA, Briggs AH, Hawke CI. Management of urinary tract infection in general practice: a cost-effectiveness analysis. Br J Gen Pract. 2000;50(457):635–9. [PMC free article] [PubMed] [Google Scholar]
- 21.Little P, Turner S, Rumsby K, Warner G, Moore M, Lowes JA, et al. Dipsticks and diagnostic algorithms in urinary tract infection: development and validation, randomised trial, economic analysis, observational cohort and qualitative study. Health Technol Assess. 2009;13(19):iii–iv, ix–xi, 1–73. [DOI] [PubMed]
- 22.Lammers RL, Gibson S, Kovacs D, Sears W, Strachan G. Comparison of test characteristics of urine dipstick and urinalysis at various test cutoff points. Ann Emerg Med. 2001;38(5):505–12. [DOI] [PubMed] [Google Scholar]
- 23.Advani SD, North R, Turner NA, Ahmadi S, Denniss J, Francis A, et al. Performance of Urinalysis Parameters in Predicting Urinary Tract Infection: Does One Size Fit All? Clin Infect Dis. 2024;79(3):600–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Moragas A, Monfà R, García-Sangenís A, Llor C. Accuracy of leukocyte esterase and nitrite tests for diagnosing bacteriuria in older adults: a systematic review and meta-analysis. Clin Microbiol Infect. 2026;32(1):19–29. [DOI] [PubMed]
- 25.van den Broek D, Keularts IM, Wielders JP, Kraaijenhagen RJ. Benefits of the iQ200 automated urine microscopy analyser in routine urinalysis. Clin Chem Lab Med. 2008;46(11):1635–40. [DOI] [PubMed] [Google Scholar]
- 26.World Medical Association. Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191–4. [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
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.