Diagnostic accuracy of a urine dipstick for detecting albuminuria in hypertensive patients

Background: Screening for albuminuria is generally recommended among patients with hypertension. While the urine dipstick is commonly used for screening urine albumin, there is little evidence about its diagnostic accuracy among these patients in Thailand. This study aimed to assess the diagnostic accuracy of a dipstick in Thai hypertensive patients for detecting albuminuria. Methods: This study collected the data of 3,067 hypertensive patients, with the results of urine dipstick and urine albumin-to-creatinine ratio (ACR) from random single spot urine being examined in the same day at least once, at Lampang Hospital, Thailand, during 2018. For ACR, a reference standard of ≥ 30 mg/g was applied to indicate the presence of albuminuria. Results: The sensitivity, specificity, positive predictive value (PPV), and negative predictive value of the trace result from dipsticks were 53.6%, 94.5%, 86.5%, and 75.5%, respectively. The area under the receiver operating characteristic curve of the dipstick was 0.748. Conclusion: Using the dipstick for screening albuminuria among hypertensive patients should not be recommended for mass screening due to its low sensitivity. In response to high PPV, a trace threshold of the dipstick may be used to indicate presence of albuminuria.


Introduction
Strong evidence has indicated that the presence of albuminuria in hypertensive patients is associated with the development of chronic kidney disease (CKD), which increases the risk of cardiovascular-related morbidity and mortality 1,2 . Early detection of CKD is important as either angiotensin-converting-enzyme inhibitor drugs or angiotensin II receptor blocker drugs can be added to a patient's treatment regimen to slow down the progress of the disease and thus reduce all-cause mortality.
Detection of albumin in urine plays an important role in diagnosing CKD in the early stages. Regarding the detection of albumin in urine, urine albumin-to-creatinine ratio (ACR) has widely been recommended to be used in diagnosing albuminuria, which is defined as the amount of urine albumin divided by urine creatinine ≥ 30 mg/g [≥ 3 mg/mmol] 3,4 .
Despite the recommendations, performing ACR in all patients with hypertension is not always applicable, particularly in a primary care unit in rural or outreach areas where the necessitated resources may be unavailable. Practically, the urine dipstick is a test that has widely been used to identify the presence of albumin in the urine due to its low cost and high accessibility.
Although using the urine dipstick is pragmatic, existing literature has not affirmed the accuracy of the test. Previous research has revealed a variety of diagnostic accuracy of the urine dipstick, compared with ACR. While some studies suggest that the dipstick is inappropriate for screening albuminuria 5-8 , others conclude that trace albuminuria from a dipstick can be used to indicate the presence of urine albumin 9,10 .
Owing to result inconsistencies, it is still arbitrary as to whether or not positive findings of albumin from a urine dipstick could be used to confirm presence of albuminuria. Additionally, there is as yet no evidence to demonstrate if diagnostic results would be consistent across populations. Therefore, this study aimed to assess the diagnostic accuracy of a dipstick in Thai hypertensive patients for detecting albuminuria.

Participants
This analysis is based on retrospective data from patients who visited Lampang Hospital from January to December 2018. The study included patients aged 18 years and over who were diagnosed with hypertension, ICD10 code "I10-14", with the results of urine dipstick and ACR from random single spot urine being examined in the same day at least once. Laboratory results from the last visit were used if multiple results of a urine dipstick and ACR on the same day were presented within the same patient. Patients with the urine results containing white blood cell more than 5-10 cells per high power field were suspected of having urinary tract infections, and thus were excluded from the study.
This study protocol was approved by the Ethics Committee at Lampang Hospital (No.79/62). Consent of the patients to use their data in the study was waived by the ethical committee due to the retrospective nature of the study.
Reference standard and index test ACR was a reference standard to indicate the level of urine albumin. Evaluation of ACR was performed at Lampang Hospital using the immunoturbidimetric essay by AU5800/ DxC700AU. The result of ACR ≥30 mg/g indicates the presence of albuminuria 11,12 .
This study employed the urine dipstick, "URiSCAN 9 SG" and the analyzer "URiSCAN SUPER+", as an index test.
Interpretation of the results were based on the color changes on the indicator tetrabromophenol blue in the presence of urine albumin. A positive reaction is indicated by a color change to yellow or green, reflecting the albumin results of negative, trace, 1+, 2+,3+, and 4+.

Covariates
Demographic characteristics including age and sex were collected for use in the analysis. Body mass index was calculated by weight in kilograms divided by squared height in centimeters 13 . Glomerular filtration rate (GFR) was estimated using the formula eGFR = 141 × min(S Cr /κ, 1) α × max(S Cr /κ, 1) -

Statistical analysis
Chi-squared test and t-test were applied to explore the association between the presence of albuminuria from ACR and covariates, with a significance level of 0.05. Sensitivity, specificity, positive predictive value, and negative predictive value of the dipstick were calculated, with 95% confidence intervals.
The area under the receiver operating characteristic curve was approximated to demonstrate the test performance 16 . Subgroup analyses using the trace threshold of dipstick were performed to elucidate the diagnostic accuracy of the test among subgroups. Statistical analyses were performed using STATA version 13 17 .

Results
A total of 3,067 hypertensive patients matched the study criteria and were included in the analysis (Table 1). Approximately 39.8% of the samples presented with albuminuria. The mean age of the patients was 63.7 year, with ~40% being men. Diabetes appeared among 73.7% of the patients; 17.7% of them had eGFR <60 ml/min/1.73m 2 . Albuminuria was present in 24.5% of those with negative result from the dipsticks. Distribution of albumin-creatinine ratios with respect to results of urine dipsticks were exhibited in Figure 1.

Amendments from Version 2
The 2nd version of the manuscript provided additional information in the method section for better clarity. Furthermore, a reference was revised.
Any further responses from the reviewers can be found at the end of the article Figure 1. Distribution of albumin-creatinine ratios stratified by results of dipsticks.

REVISED
Body mass index, mean±SD 25.6±5.1 25.7±4.7 0.592 Table 2 demonstrated the sensitivity, specificity, positive and negative predictive values of urine dipstick in detecting albuminuria. It is seen that sensitivity of 53.6% was observed when the trace threshold was applied, whereas cutoff of ≥2+ and higher yields 100% test specificity. The area under the receiver operating characteristic curve was 0.7482 ( Figure 2).
Comparing diagnostic accuracy of the dipstick, it appears that sensitivity, specificity, along with positive and negative  predictive values were approximately the same in all subgroups (Table 3).

Discussion
Existing studies have manifested a wide range of positive predictive values (PPVs) of urine dipsticks among patients with hypertension, ranging from 27 to 82 6,18 . However, none have been conducted in a Thai population. Results of this study, exploring the diagnostic accuracy of the dipstick in a Thai population, not only illustrates the outcomes in this specific population, but can also be used in comparison with results from other populations for a better understanding of test accuracy.
Previous research has documented the differences in sensitivity and specificity of the dipstick across populations.
A Japanese study showed sensitivity, specificity, and PPV of 37.1%, 97.3%, and 71.4%, respectively 10 . Another study conducted in Australian adults showed sensitivity, specificity, and PPV of 69.4%, 86.8%, and 27.1%, respectively 6 . One possible explanation for the difference in diagnostic accuracy of the dipstick was owing to differences in the characteristics It should be noted that false positive results of the dipstick could come from highly alkaline urine and contamination of antiseptics. Moreover, urine specimens used in this study came from random spot urine collection, which may be subjected to false positive results. Likewise, false negative results may have occurred due to excessive hydration before collecting the urine specimen, which leads to a decrease in concentration of urine albumin and subsequently a smaller chance of detecting albuminuria.
Such low sensitivity of 53% from the urine dipstick indicates that almost half of the patients with albuminuria cannot be identified using just the urine dipstick. It is also seen that among patients with a negative albumin result from the dipstick, albuminuria was found in nearly a quarter of them. This outcome well aligns with previous studies asserting low sensitivity of the dipstick in detecting albuminuria 5,8,10 . Given strong evidence indicating the high probability of cases being undetected, using the dipstick alone should not be recommended for use in screening of albuminuria among hypertensive patients.
Results from the study revealed a rather high predictability of the dipstick in detecting urine albumin. Concerning the dipstick cutoffs, applying the trace threshold yields a PPV of 86.5%, compared with 98.2% and 100% using the 1+ and 2+ thresholds, respectively. Though a rather high chance of predicting albuminuria once hypertensive patients have these results of trace or higher from the dipstick, it should be borne in mind that albuminuria may be overly diagnosed with the application of the trace threshold, compared with using the higher cutoffs.
Although excellent PPV can be achieved when employing higher thresholds of the dipstick, drawbacks remain when the recommendation for using the high threshold is applied due to fewer patients being applicable. Considering the trade-off between PPV and applicability of the dipstick results, the trace threshold may be recommended for indicating the presence of albuminuria in hypertensive patients.
Even though the KDIGO guidelines 3 have recommended the use of ACR to indicate the presence of albuminuria, this is proven to be rather costly and not readily available in some regions. Limitations, regarding the availability and costs of ACR, may arise when considering the application of ACR for routine screening of hypertensive patients. Nonetheless, evidence has demonstrated a low sensitivity of urine dipsticks, which should not be recommended for screening albuminuria. Hence, ACR is deemed the option for screening albuminuria in the setting where resources are available.

Conclusion
While existing evidence is controversial to whether the urine dipstick should be recommended for screening albuminuria in hypertensive patients, results from this study demonstrated that the dipstick has such low sensitivity in detecting albumin in urine in the Thai population. These results suggest that the urine dipstick not be recommended for screening urine albumin in patients with hypertension. In contrast, results of trace or higher yields high PPV, indicates a very high possibility of the presence of microalbuminuria.

Jhonatan Mejia
Universidad Nacional del Centro del Peru, Huancayo, Peru The authors aimed to determine the diagnostic accuracy of a urine dipstick for detecting albuminuria in hypertensive patients. They developed a retrospective study in 3067 hypertensive patients from a Thai hospital concluding that a urine dipstick should not be used for mass screening, but a trace result could be used to detect albuminuria. The manuscript is well-written, with an adequate design, statistical analysis, and conclusions. Nevertheless, I consider that some minor changes should be considered to enhance the reproducibility and readability.

Methods
For further risk of bias evaluation, it will be helpful to clarify if the reference and index test were performed with the same random urine sample. Also, the authors should specify if there was a big gap of time (< 8, ≥ 8 hours) between reference and index test evaluation, along with the preservation method for long intervals. ○ It should be literally stated if the test was performed with the analyser or manually, as well as the ACR or albumin cut-off values for "trace, 1+, 2+,3+, and 4+".

Results
I believe the term "≥ Negative" is confusing. It would be better to consider "> Negative" and specify the prevalence of albuminuria in another

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Partly

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: urinary tract infections, medical philosophy I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Author Response 07 Oct 2021
Win Techakehakij, Lampang Hospital, Thailand, Amphur Muang, Thailand We could not find research in this area presenting the detailed information of diagnostic test parameters, sensitivity, specificity, PPV, and NPV, respecting to each dipstick cutoff, including the Neg. We thus appreciate that presenting this information would be beneficial to the audiences who want to find out in details about this issue.
Moreover, while it is known that prevalence of the disease in the sample population directly affects PPV, this formation is generally missed in most papers. Providing the PPV information at the Neg cutoff clearly demonstrates the prevalence of proteinuria in the sample and may help audiences gain a better understanding of this research.
already performed test in that particular study population, then it should be made sufficiently clear in the manuscript. I would like to maintain that discussing the utility of the high PPV of the test after recommending against the use of the test is confusing.
Introduction: Reference no.4 is a study comparing 24-hour urinary albumin estimation versus urine spot ACR in patients with diabetic patients. How are the results from this study relevant? 2.
Methods: "Laboratory results from the last visit were used if multiple results of a urine dipstick and ACR on the same day were presented within the same patient." Do the authors mean that the last one of the multiple tests done on the same day was considered for inclusion in the study? 3.
The method used to exclude patients with UTI should be mentioned in the manuscript. 4.
Results: In my opinion, giving the PPV for ≥ negative test result is not intuitive (in Table 2). This could be modified as > Negative.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: urinary tract infections, medical philosophy I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Author Response 23 Sep 2021
Win Techakehakij, Lampang Hospital, Thailand, Amphur Muang, Thailand Abstract: If the authors' point is that the high PPV is helpful in interpreting the result of an already performed test in that particular study population, then it should be made sufficiently clear in the manuscript. I would like to maintain that discussing the utility of the high PPV of the test after recommending against the use of the test is confusing.
We rechecked throughout the manuscript to ensure emphasis of the Thai population was appropriately addressed. ○ As stated in the previous response, implication of the use of sensitivity and PPV information are different. Thus, we briefly provided the key messages of both issues for the audiences. ○ Introduction: Reference no.4 is a study comparing 24-hour urinary albumin estimation versus urine spot ACR in patients with diabetic patients. How are the results from this study relevant?
We removed the reference no.4. ○ Methods: "Laboratory results from the last visit were used if multiple results of a urine dipstick and ACR on the same day were presented within the same patient." Do the authors mean that the last one of the multiple tests done on the same day was considered for inclusion in the study?
It is correct that the last UA results were used in the study if multiple tests were done on the same day. This is as, in reality, there is sometimes errors in performing laboratory test or in reporting results, which require a re-test. Thus, we decidedly chose the last result to minimize these errors.

○
The method used to exclude patients with UTI should be mentioned in the manuscript.
We added "Patients with the urine results containing white blood cell more than 5-10 cells per high power field were suspected of having urinary tract infections, and thus were excluded from the study" in the method section. ○ Results: In my opinion, giving the PPV for ≥ negative test result is not intuitive (in Table 2). This could be modified as > Negative.
We suggest to keep the ≥ negative results as providing the results of "≥ negative" is beneficial to audiences to know the prevalence of albuminuria among the study population. ○ Competing Interests: No competing interests were disclosed.

Version 1
The manuscript is clear and well documented. The rationale is well established. The authors applied appropriate methods for data analysis and the results were convincing. However, to improve paper readability, minor changes are required. I suggest the following: The introduction did not discuss the heterogeneity of various accuracy of urine dipstick.

○
The authors selected the laboratory results from the last visit. It would be great if you can provide the rationale for this selection method.

○
The logic for calculating the sample size is missed in this manuscript. The authors may add the calculation of sample size or power analysis. ○ Table 1: Regarding rounding decimals, the "63.52" should be "63.5". ○ Figure 1: Please explain or discuss the outlier (i.e., ACR ~ 5000+ in the Negative group). Thank you for the opportunity of reading and evaluating this paper.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate?
Concerning the implication, this research mentioned the interpretation and applicability of the results in the discussion part: "results from this study demonstrated that the dipstick has such low sensitivity in detecting albumin in urine. These results suggest that the urine dipstick not be recommended for screening urine albumin in patients with hypertension. In contrast, results of trace or higher yields high PPV, which indicates a very high possibility of the presence of microalbuminuria.".
across populations is not clear enough. Does this apply only to automated analyzer results?
The authors state that proteinuria will be over-diagnosed with trace threshold. This point is also not substantiated since a test with a PPV of 98-100% will not have a high false-positivity rate.
diagnostic test is considered. "≥Trace", for instance, means that any results with albuminuria higher than the trace level, trace,1+,…,4+, were counted. A 100% sensitivity when applying the negative threshold is explained by the fact that the negative result is the lowest possible outcome of dipsticks. Using the negative threshold, all the samples, with or without albuminuria, would be identified as having albuminuria, resulting in a 100% sensitivity.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com