Summary
We evaluated the accuracy of a point-of-care test designed to measure adherence to isoniazid (INH) preventive therapy in a hospital setting in Rio de Janeiro, Brazil. Patients on treatment with daily INH and patients not receiving INH were included. Sensitivity and specificity of the test were 84%/98% at the first minute, and 95%/98% at the fifth minute, respectively. Among smokers, sensitivity and specificity was reduced (80%/89% at the fifth minute, respectively), but only 17% smoked. This test accurately detected INH metabolites 24 hours following directly observed INH intake, though sensitivity and specificity may be compromised by tobacco smoke exposure.
Keywords: latent tuberculosis infection, isoniazid, adherence, Arkansas method
Introduction
Treatment of latent tuberculosis (TB) infection with isoniazid (INH) preventive therapy (IPT) as a TB control strategy has been limited predominantly to developed countries, though in settings with high rates of HIV co-infection, this strategy is being re-visited (1). Ensuring adherence to IPT is a significant challenge. Although measuring patient adherence is essential, it is often limited to patient report (2).
The Arkansas color method of detecting INH metabolites in the urine relies on a colorimetric assay test performed in a laboratory (3,4,5,6). A recently developed test, IsoScreen (GFC Diagnostics Limited, Oxfordshire, UK), utilizes the reagents of the Arkansas method in a patented plastic testing device (SafeTube), allowing test evaluation in the field. However, IsoScreen has only been tested in small groups of patients (7).
The objective of this study was to determine, under routine conditions, the sensitivity and specificity of the IsoScreen test in measuring adherence to IPT among an out-patient hospital population with high prevalence of TB in Rio de Janeiro, Brazil.
Methods
We recruited two patient groups: patients currently on treatment with daily INH were classified as the INH group; and individuals attending the hospital for initial evaluation for active or latent TB therapy initiation were considered the control group. For the INH group, a first visit was a routinely scheduled appointment in which the subjects took INH under supervision, and a second visit was scheduled after 24 hours for urine specimen collection. The control group had only one visit to provide the urine specimen. All patients were ≥18 years. Exclusion criteria were non-daily INH use for the INH group, and INH intake within the last 30 days for the controls. Urine specimens were aspirated using the syringe connected to the device, then injected into the dried chemicals within the device and shaken. The urine color was recorded one and five minutes after mixing. The test was considered positive when urine turned green or dark blue and negative if no change in color. The McNemar test was used to analyze the difference between test sensitivities at first and fifth minutes.
This project was approved by the Clementino Fraga Filho University Hospital and the Johns Hopkins School of Medicine institutional review boards.
Results
From June 2007 to July 2008, a total of 146 subjects were included in the study. Two patients were excluded from the INH group because they did not return to provide a urine specimen. There were 94 patients in the INH group and 50 in the control group.
In the INH group, 65 (69%) patients were on IPT while 29 (31%) were on active TB treatment. The median time from INH ingestion and test performance was 23.4 hours (IQR: 23.0–24.2 hours). Table 1 shows test accuracy. Sensitivity was 10.3% greater at the fifth minute (94.7%) versus the first minute (84.4%; p<0.01), and specificity was 98% at both readings. The difference between sensitivities at first and fifth minutes was due to nine specimens that became blue after the first minute. Table 2 shows demographic characteristics of individuals with false test results. Among smokers, the sensitivity and specificity decreased to 80% and 88.8% at fifth minute, respectively, though only 17% (24/144) were smokers.
Table 1.
Isoscreen test accuracy at one and five minutes after test performance
| TnQTable1 | INH group | Control group | Sensitivity | Specificity | INH group | Control group | Sensitivity | Specificity |
|---|---|---|---|---|---|---|---|---|
| Color | (95% CI) | (95% CI) | (95% CI) | (95% CI) | ||||
| yellow | 14 | 49 | 5 | 49 | ||||
| green | 8 | 0 | 8 | 0 | ||||
| blue | 68 | 1 | 84.4%* | 98.0% | 81 | 1 | 94.7* | 98.0 |
| missing | 4 | 0 | (78.5 – 90.3) | (95.7 – 100) | 0 | 0 | (91.0 – 98.4) | (95.7 – 100) |
| Total | 94 | 50 | 94 | 50 | ||||
Sensitivity/Specificity calculated with blue and green reflecting a positive result
The difference between sensitivity at fifth and first minutes was 10.3% (p<0.01).
Table 2.
Demographic characteristics of patients with false test results at one and five minutes after test performance
| Total | Age mean (SD) | Gender male n (%) | Race white n (%) | Smoker n (%) | Use of other medications n (%) | |
|---|---|---|---|---|---|---|
| Urine color at 1st minute | ||||||
| INH group | ||||||
| Yellow (false-negative) | 14 | 45 (10.3) | 9 (64) | 5 (36) | 3 (21) | 6 (43) |
| Control group | ||||||
| Green (false-positive) | 0 | 0 | 0 | 0 | 0 | 0 |
| Blue (false-positive) | 1 | 28 | 1 (100) | 0 | 1 (100) | 0 |
| Urine color at 5th minute | ||||||
| INH group | ||||||
| Yellow (false-negative) | 5 | 50 (5.3) | 3 (60) | 2 (40) | 3 (60) | 4 (80) |
| Control group | ||||||
| Green (false-positive) | 0 | 0 | 0 | 0 | 0 | 0 |
| Blue (false-positive) | 1 | 28 | 1 (100) | 0 | 1 (100) | 0 |
Regular medication use other than TB drugs
Discussion
We evaluated a point-of-care test for measuring adherence to IPT, and our findings show high sensitivity and specificity (≥95%) with mean time from INH ingestion and test results of approximately 24 hours. Based on these results, this test is being utilized in the THRio study in Rio de Janeiro, Brazil investigating the adherence of HIV-infected patients receiving IPT (8).
Adherence to IPT has been reported to be enhanced by measures such as provision of incentives and treatment supervision (2,9). However, due to the high number of patients who would be recommended to receive IPT, these measures are prohibitively expensive, especially in developing countries where the limited resources available are directed towards detection and treatment of active disease. The IsoScreen test might be a practical alternative for this purpose since it can be performed without strict laboratory conditions and provides an immediate confirmation of patient compliance. In a previous study, non-adherence to IPT was found in 28.5% of patients submitted to the Arkansas method. However, 83% of patients who were tested negative at the first visit became adherent in the second visit, suggesting that biological acknowledgement of their non-adherence may have changed their behavior (6).
In our study, in five patients for whom INH ingestion was confirmed, the test was negative. Four of them were on regular treatment with other drugs, although none were receiving the same drugs. Although we did not assess acetylator status, the false-negative results could be due to faster drug metabolization, as previously reported, or it may actually reflect variability in INH metabolites detected in urine or inherent test inaccuracy (10,11). Test accuracy may also have been compromised by tobacco smoke exposure similarly to another study evaluating the test spectrophotometrically. In that study, while the urine of non-smokers taking INH turned deep blue, the presence of nicotine and its metabolites in urine of smokers taking INH produced a green color after reaction, resulting in the blue-green color (3).
A limitation of our study is that it was unblinded. Since the study was conducted under routine operational conditions, the responsible nurse was aware of which subjects were under current INH treatment and which subjects were being investigated for TB (controls). Because a green color (also considered as positive in this study) may not definitively represent a positive result, we presented the results stratified by color (Table 1). If green was not considered to be a positive result, then sensitivity at 5 minutes would have been 86%, while specificity would have remained unchanged. We are confident that our technician was readily able to differentiate between yellow and green, and that the high sensitivity reported (95%) suggests strong utility for this test in operational settings. Another limitation is that the test accuracy was only measured within 24 hours of drug intake, thus we cannot confidently extrapolate our results beyond this period of time. In programmatic settings, patients may miss a dose 24 hours prior to providing a urine specimen, but may still be strongly adherent to their regimen or take IPT the day before the urine test simply to “appear” to be adherent. Thus, appropriate counseling for test objective and an oral history of adherence should always be included when using a biological assay for measuring adherence.
In conclusion, the IsoScreen test accurately detected INH metabolites 24 hours following directly observed INH intake. As an objective test, it might be used repeatedly as a follow-up method to better characterize adherence and potentially increase adherence among patients receiving IPT or treatment for active TB.
Acknowledgments
This work has been supported, in part, by a fellowship/grant from the Fogarty International Center/USNIH: Grant # U2RTW006885 ICOHRTA. Marcus Conde was supported by Brazilian CNPq grant PQ-2006 (309552/2006-0). Jonathan Golub and Richard Chaisson were supported by the National Institutes of Health grants AI066994 and AI001637, respectively.
Footnotes
Authorship
Renata L. Guerra participated in data management, data analysis and interpretation, and drafted the manuscript. Marcus B. Conde, Anne Efron, Jonathan E. Golub and Richard E. Chaisson designed the study, participated in data interpretation, and reviewed the report. Carla Loredo e Gisele Bastos participated in data collection and data management. All authors approved the report final version.
Conflict of interest
The authors have no conflict of interest. The study sponsors had no involvement in the study design, data collection, analysis and interpretation, writing of the manuscript, and in the decision to submit the manuscript for publication.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Churchyard G, Fielding K, Charalambous S, et al. Efficacy of secondary isoniazid preventive therapy among HIV-infected Southern Africans: time to change policy? AIDS. 2003;17(14):2063–2070. doi: 10.1097/00002030-200309260-00007. [DOI] [PubMed] [Google Scholar]
- 2.American Thoracic Society. Targeted tuberculosis testing and treatment of latent tuberculosis infection. Am J Respir Crit Care Med. 2000;161:5221–5247. [Google Scholar]
- 3.Schraufnagel DE, Stoner R, Whiting E, Snukst-Torbeck G, Werhane MJ. Testing for isoniazid. An evaluation of the Arkansas method. Chest. 1990;98(2):314–316. doi: 10.1378/chest.98.2.314. [DOI] [PubMed] [Google Scholar]
- 4.Elizaga J, Friedland JS. Monitoring compliance with antituberculous treatment by detection of isoniazid in urine. The Lancet. 1997;350:1225–1226. doi: 10.1016/S0140-6736(05)63457-5. [DOI] [PubMed] [Google Scholar]
- 5.Perry S, Hovell MF, Blumberg E, et al. Urine testing to monitor adherence to TB preventive therapy. J Clin Epidemiol. 2002;55:235–238. doi: 10.1016/s0895-4356(01)00470-x. [DOI] [PubMed] [Google Scholar]
- 6.Eidlitz-Markus T, Zeharia A, Baum G, Mimouni M, Amir J. Use of the Urine Color Test to Monitor Compliance With Isoniazid Treatment of Latent Tuberculosis Infection. Chest. 2003;123:736–739. doi: 10.1378/chest.123.3.736. [DOI] [PubMed] [Google Scholar]
- 7.Whitfield R, Cope GF. Point-of-care test to monitor adherence to anti-tuberculosis treatment. Ann Clin Biochem. 2004;41:411–413. doi: 10.1258/0004563041731637. [DOI] [PubMed] [Google Scholar]
- 8.Moulton LH, Golub JE, Durovni B, et al. Statistical design of THRio: a phased implementation clinic-randomized study of a tuberculosis preventive therapy intervention. Clin Trials. 2007;4(2):190–9. doi: 10.1177/1740774507076937. [DOI] [PubMed] [Google Scholar]
- 9.White MC, Tulsky JP, Reilly P, McIntosh HW, Hoynes TM, Goldenson J. A clinical trial of a financial incentive to go to the tuberculosis clinic for isoniazid after release from jail. Int J Tuberc Lung Dis. 1998;2:506–512. [PubMed] [Google Scholar]
- 10.Ellard GA, Jenner PJ, Downs PA. An evaluation of the potential use of isoniazid, acetylisoniazid and isonicotinic acid for monitoring the self-administration of drugs. Br J Clin Pharmac. 1980;10:369–381. doi: 10.1111/j.1365-2125.1980.tb01773.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hanifa Y, Mngadi K, Lewis J, Fielding K, Churchyard G, Grant AD. Evaluation of the Arkansas method of urine testing for isoniazid in South Africa. Int J Tuberc Lung Dis. 2007;11:1232–1236. [PubMed] [Google Scholar]