Adaptations of the Saker-Solomons test: simple, reliable colorimetric field assays for chloroquine and its metabolites in urine.

Adaptations of the Saker-Solomons test: simple, reliable colorimetric field assays for chloroquine and its metabolites in urine* D.L. Mount,' B.L. Nahlen,2 L.C. Patchen,l & F.C. Churchill3 Two field-adapted colorimetric methods for measuring the antimalarial drug chloroquine in urine are described. Both are modifications of the method of Saker and Solomons for screening urine for phencycli- dine and other drugs of abuse, using the colour reagent tetrabromophenolphthalein ethyl ester. One method is semiquantitative, detecting the presence of chloroquine (Cq) and its metabolites in urine with a 1 pg/ml detection limit; it is more sensitive and reliable than the commonly used Dill-Glazko method and is as easy to apply in the field. The second method uses a hand-held, battery-operated filter photometer to quantify Cq and its metabolites with a 2 pg/ml detection limit and a linear range up to 8 pg/ml. The first method was validated in the field using a published quantitative colorimetric method and samples from a malaria study in Nigeria. The second method was validated in the laboratory against high-performance liquid chromatographic results on paired samples from the Nigerian study. Both methods may be used in remote locations where malaria is endemic and no electricity is available. A number of colorimetric methods for detecting and estimating the concentrations of chloroquine (Cq) and its metabolites (primarily desethylchloroquine (DECq)) in urine have been published in the past 40 years for use in malaria field studies (1-5). Because of its simplicity the Dill-Glazko method (1) is most commonly used, but recent publications have con- cluded that the test is neither sufficiently sensitive nor reliable to assess effectively the history of Cq use (6-8). A simple, more sensitive and reliable field assay than the Dill-Glazko method is therefore needed for monitoring chemoprophylaxis and treat- ment compliance, investigating drug-use patterns, and screening prospective subjects for in-vivo and in- vitro Cq resistance studies based on evidence of recent drug use. Quantitative measurements of Cq + metabolites in urine can be a useful adjunct to drug sensitivity studies in the field. Three methods (1-3) are semi- quantitative at best, while two recently published methods (4, 5) permit quantitation of Cq + metabolites. The latter methods require more equipment than the former, but may be performed * From the Division of Parasitic Diseases, Center for Infectious Diseases, Centers for Disease Control, Atlanta, GA 30333, USA. Requests for reprints should be sent to Dr F.C. Churchill at this address. ' Research Chemist, Control Technology Branch, Division of Parasitic Diseases. 2 Epidemiologic Intelligence Service Medical Officer, Malaria Branch, Division of Parasitic Diseases. 3 Chief, Control Technology Branch, Division of Parasitic Dis- eases. Reprint No. 4979 without mains electricity using a hand-held battery- operated, filter photometer (5). This paper describes an adaptation of a method developed by E. G. Saker and E. T. Solomons for assay of drugs of abuse in urine (9) which permits facile qualitative or quantitative determination of Cq + metabolites in the field. Cq was partitioned from buffered urine (pH 8) into a chloroform solu- tion of tetrabromophenolphthalein ethyl ester (TBPEE) to give a purple colour, the intensity of which is proportional to the sum of the concentra- tions. of Cq and its metabolites. We quantified Cq + metabolites by comparing the colour intensity between samples and standards either by eye (Saker- Solomons, Cq Method I) or by using a hand-held, battery-operated filter photometer (Saker-Solomons, Cq Method II). These two methods are abbreviated S-S/CqI and S-S/CqII, respectively. We validated S-S/CqI with the Haskins MMII assay (5) during a malaria study in Nigeria and S-S/CqII with a refer- ence high-performance liquid chromatographic (HPLC) method (10) in the laboratory using urine samples from the same study. Both Saker-Solomons tests may be applied without mains electricity. Materials and methodsa A hand-held, battery-operated filter photometerb equipped with a 565-nm filter was used for colori- metric quantification. a Use of trade names and commercial sources is for identifica- tion only and does not constitute endorsement by the Public Health Service or by the U.S. Department of Health and Human Services. b Chemetrics Inc., Calverton, VA, USA. Bulletin of the World Health Organization, 67 (3): 295-30 (1989) 295 D.L. Mount et al. Reagents and standards Chloroquine and desethylchloroquine standards available from earlier studies were donated.c Chloro- form,d TBPEE,e K2HP04- 3H20,' and KH2PO4' were obtained commercially. All other chemicals were of reagent grade or better. The pH-8 buffer contained 324 g of K2HPO4- 3H20 and 10 g of KH2PO4 in 1 litre of water. The solution of 50 mg TBPEE dissolved in 100 ml of chloroform was shaken with 10 ml of 2 mol/l aqueous HCI. After the separation of phases, the aqueous phase was aspira- ted by Pasteur pipette, leaving 0.05% TBPEE-in- chloroform reagent solution. Sunlight and high ambient temperatures promote decomposition of the reagent in solution, which may be monitored by noting the decrease in reagent colour intensity with time. To minimize decomposition the solution should be stored in the dark and refrigerated, using a fresh solution every 3 weeks. If the solution is stored at room temperature, fresh solutions should be made every 3-4 days. Procedure for the S-S/Cql assay Pre-fild preperations. Screw-cap conical centrifuge tubes (15 ml) were charged with 1 ml of pH-8 phos- phate buffer and 0.2 ml of the TBPEE solution. Field asay. Urine (2 ml) from each subject was trans- ferred to a pre-charged centrifuge tube. The tubes were capped, vigorously hand-shaken for about 15 seconds, and kept standing for about 15 minutes to allow phase separation before the result was read. A yellow-green colour of the chloroform layer indicates a negative test for Cq + metabolites; a red to purple colour of the organic layer indicates a positive result with the shade depending on the concentration of Cq + metabolites. We estimated Cq concentrations by comparing the sample tubes with tubes of control urine fortified with 0, 1, 2, or 3 ug/ml Cq. Procedure for the S-S/CqlI assay Pre-fleid preperations. The 0.05% TBPEE-in-chloro- form solution was diluted to 0.02% with chloroform; screw-cap culture tubes (13 x 100mm) were charged with 2 ml of the diluted (0.02%) TBPEE solution and 1 ml of pH-8 phosphate buffer. Field assy. Urine (2 ml) from each subject was trans- ferred to a pre-charged culture tube. The tubes were capped and inverted 30 times by hand; many tubes c Sterling-Winthrop, Rensselaer, NY, USA. d J.T. Baker, Phillipsburg, NJ, USA. * Eastman Organic Chemicals, Rochester, NY, USA. 'Aldrich Chemical Co., Milwaukee, WI, USA. may be grasped and inverted simultaneously. The tubes were kept standing for about 15 minutes and the chloroform phase was clarified using a glass- wool swab (5). The percentage transmittance of the chloroform layer was read using the filter photome- ter (565-nm filter). Each value was recorded, con- verted to absorbance, and compared with a freshly prepared standard curve to yield the concentration values. In practice, the filter photometer was zeroed with the blank, and measurements on standards and samples were made relative to this value. The absorbance of the blank was 0.21 when measured relative to chloroform and is due largely to the colour of the TBPEE-in-chloroform solution. Procedure for liquid chromatographic assy The liquid chromatographic method has been pre- viously described (10); 100-jul samples of urine were taken for HPLC analysis. Determination of cross-reactivity of other antimalarials to S-S/Cqii Blank urine was spiked to give individual 5 jg/ml concentration standards for each of 9 antimalarial drugs or metabolites. We analysed each standard by S-S/CqII, subtracting the blank absorbance, and compared the resulting value with that from a 5 4ug/ml Cq standard. Comparison of the 5-SICqI, Dill-Glazko, and Haskins MMII tests during field trials The S-S/CqI method was evaluated in Oban (Cross River State) and Igbo Ora (Oyo State), Nigeria, in connection with in-vivo and in-vitro drug sensitivity studies during July and August of 1987. Thick blood smears from consecutive children <5 years of age who were brought to the local health clinics at the two study sites were examined for the presence of Plasmodium parasites. Children whose blood smears were positive for P. falciparum were treated with 25 mg/kg Cq in three doses (10, 10 and 5 mg/kg) over 3 days. A urine sample was collected from each study patient before treatment on day 0 and at follow-up visits on days 2, 7, and 14. Urine samples were analysed by the S-S/CqI, Dill-Glazko, and Haskins MMII assays at the local health clinic laboratories in Oban and Igbo Ora. Urine specific gravities were measured and recorded. The quantitative Haskins MMII test was interpreted according to the percentage of light transmittance using a filter photometer. A urine sample was con- sidered positive by S-S/CqI if the colour changed to purple in the chloroform layer, by Dill-Glazko if the reagent colour changed to rose, and by Haskins 296 WHO Bulletin Vol. 67 1989 Adaptations of the Saker-Solomons test for chloroquine In urine MMII if there was < 70% transmittance (corresponding to a concentration of > 1 ug/ml Cq + metabolites (5)). Comparison of the S-S/CqII and HPLC asay results A 3-ml sample from each urine collection in the study in Nigeria was refrigerated and transported to the Centers for Disease Control laboratories in Atlanta for later analysis by S-S/CqIl and HPLC. Results Comparison of the S-S/Cql, DiII-Glazko, and Haskins MMII tests In the field Table 1 shows the results of 115 urine samples tested by both S-S/CqI and Haskins MMII. The sensitivity of the S-S/CqI test is 95% and its specificity 99% relative to Haskins MMII. There is no significant difference in results obtained by the S-S/CqI and the Haskins MMII tests. The detection limit for S-S/Cql is 1 pg/ml. Table 2 shows poor agreement of results from the Dill-Glazko test and the Haskins MMII test; 19 out of 119 samples gave negative results for the former and positive results for the latter. Comparison of the S-S/Cqll and HPLC assays In the laboratory Fig. 1 shows a standard curve for the S-S/CqII assay prepared by fortifying the control urines at 0, 2, 4, 6, and 8 pg/ml with Cq. A result from assay of a 12 Mg/ml standard is also plotted to illustrate the limits of linearity of the method. The standard curve is essentially linear to a concentration of approx- imately 8 pg/ml; at higher concentrations the absorbance response per unit of concentration falls. The correlation coefficient for the curve shown in Fig. 1 is 0.9896 over the range of 0 to 8 pg/ml but is 0.977 if the 12 ug/ml data point is included in the Table 1: S-S/CqI urine test compared with the Haskins MMiI (n = 115) Haskins MMII Positive' Negative S-S/Cql: Positive 41 1 Negative 2 71 Sign Test:b P = 0.50 Sensitivity = 0.95 Specificity = 0.99 Pos. predictive value = 0.98 Neg. predictive value = 0.97 Positive Haskins MMII result ccrresponds to .70% transmit- tance. bNon-parametric test for related samples. Table 2: Dili-Glazko urine test compared with the Haskins MMII (n = 119) Haskins MMII Positive Negative Dill-Glazko: Positive" 28 4 Negative 19 68 Sign Test:b P = 0.001 Sensitivity = 0.60 Specificity = 0.94 Pos. predictive value = 0.88 Neg. predictive value = 0.78 Positive Haskins MMII result corresponds to .70% transmit- tance. b Non-parametric test for related samples. calculation. The deviation from linearity above 8 pg/ml is apparent. The detection limit for this test is 2 pg/ml. Table 3 compares the results from urine samples by the S-S/CqII assay with those tested by the HPLC assay. The HPLC values represent the sum of concentrations found for Cq and its metabolites Fig. 1. Calibration curve for the S-S/CqlI assay. The sandard curve Is for the range 0-8 pg/ml (see text). 8.- 0.7 0.61 0.51 a 0 0 0 4: 0.41 0.31 0.2 0.1 oI 0 0 S X V 0 0 0.0 2 0.13 4 0.33 6 0.45 8 0.53 S Y=0.069 X+0. n=n5; r=0.981 12 0.67 I I xl -0.17 1.71 4.61 6.28 7.51 )12 916 2 4 6 8 10 12 Chloroquine (pg/ml) 14 WHO Bulletin Vol. 67 1989 I - 2917 D.L. Mount et al. Fig. 2. S-S/Cqll results compared with high-performance liquid chromatographic results (data from Table 3). S-S/Cqll results combine the contributions from DECq and Cq; for HPLC values of Cq + DECq above 8 pg/mi the corresponding S-S/Cqll values are low, as expected from the nonlInearity of the S-S/Cqll method above 8 pg/mi (see text). - 14 cm > 12 (U CA n = 10 en10 cnC,) 6 (U c4 ._ ac ._ C 2 0 o L O 0 2 4 6 8 10 12 14 16 18 Chloroquine + Desethylchloroquine in urine by HPLC (pg/ml) i 20 Table 3: Concentrations of chloroquine and metabolite In urine analysed by Saker-Solomons test and by high- performance liquid chromatographic (HPLC) assay HPLC results (pug/ml) Sample S-S/Cqll identifier' value (pg/mI) Cq DECq Cq + DECq 22-D14 2.58 2.58 0.92 3.50 49-D14 2.87 1.49 0.86 2.35 108-D14 0.00 0.54 0.24 0.78 122-D14 2.89 1.27 0.58 1.85 134-D14 5.19 3.27 0.68 3.95 144-D14 0.26 0.88 0.53 1.41 148-D14 0.11 0.71 <0.2 0.71 232-D14 0.84 1.01 0.24 1.25 174-D07 4.75 3.02 1.14 4.16 201-D07 4.03 2.87 0.82 3.69 241-D07 5.48 4.02 1.75 5.77 321-D07 2.58 1.90 0.34 2.24 337-D07 6.06 6.15 1.13 7.28 342-D07 4.90 3.90 1.16 5.06 21-102b 9.86 8.09 3.59 11.7 112-D02b 11.7 12.9 3.46 16.4 127-D02b 12.6 18.3 1.47 19.8 The second portion of the sample identifier refers to the number of days (14, 7 and 2) after treatment when the sample was taken. b When measured by HPLC, these samples contained a sum of Cq and DECq beyond the range of linearity for S-S/Cqll. (primarily DECq) since Cq and DECq were shown to provide essentially equivalent responses to both Saker-Solomons tests. A standard curve over the range of 0 to 8 ug/ml Cq was used to calculate the S-S/CqII results. Least-squares comparison of S-S/ CqII test results to the summed HPLC data for the range of 0 to 8 pg/mI of Cq (Fig. 2) yields the sta- tistical result y = 0.989x - 0.075; n = 14, r = 0.9274. Cross-reactivity of other antimalarlals to S-SlCqll The results of the cross-reactivity study are shown in Table 4. Table 4. Relative sensitivities of S-S/Cqll to some common antimalarial drugs and metabolites Drug (X) Absorbance,/Absorbancec, Chloroquine 1.0 Quinine 0.74 Proguanil 0.98 Cycloguanil 1.15' Mefloquine 0.62 Desethylamodiaquine 0.12 Sulfadoxine 0.0 Primaquine 0.0 Pyrimethamine 0.31 8 The test resulted in an amber colour rather than the character- istic purple colour. Discussion The S-S/Cql method meets all the criteria of an acceptable field test-sensitivity, reliability, simpli- city, rapidity, and low cost-and proved to be as easily taught to and performed by field technicians as the Dill-Glazko test. An average of 50 specimens WHO Bulletin Vol. 67 1989 0 298 Adaptations of the Saker-Solomons test for chloroquine In urine were processed in 1 hour, including time for labelling and recording. S-S/CqI is more sensitive than the Dill-Glazko test and therefore more effectively excludes patients who recently ingested Cq. It has been argued that Cq concentrations in urine of less than 1 ug/ml (the detection limit of S-S/CqI) indicate corresponding blood concentrations of less than 100 ng/ml, the therapeutic threshold concentration for Cq-sensitive strains of P. falciparum (11). The nega- tive predictive value (97%) for the S-S/CqI test indi- cates that this test is useful during in-vivo drug sensitivity studies for including only persons with a urine Cq concentration of < 1 gg/ml (Table 1). The high positive predictive value (98%) indicates that this test is acceptable for documenting drug absorp- tion in patients on day 2 or day 7 after Cq treatment. Table 2 shows that both positive and negative pre- dictive values of the Dill-Glazko test are substan- tially lower than those of S-S/CqI relative to the Haskins MMII test used for validation. The S-S/CqI test does not detect drugs such as caffeine, nicotine, aspirin, paracetamol (aceta- minophen), or antibiotics which might be in common use and give false positive results (9, 12). However, the Saker-Solomons Cq methods, like the bromthymol blue (4) and Haskins tests (4, 5), give positive results for the antimalarials quinine and proguanil; relative sensitivities of the S-S/Cq tests to these and other antimalarials are shown in Table 4. Even so, in applications in in-vivo and in-vitro Cq sensitivity trials, such imperfect specificity for Cq would serve the designated purpose of excluding patients with appreciable concentrations of antima- larial drugs in their bodies that could adversely affect the outcome of the studies. Urine-Cq tests are also useful for community investigations of drug-use practices. For example, an individual taking 25 mg/kg Cq in response to a febrile episode will be found positive for approx- imately 10 days using a urine-Cq method with a 1 Mg/ml detection limit; similarly, individuals taking chemoprophylactic Cq (300 mg weekly) will remain positive for up to 5 days after the most recent dose (13). The S-S/CqI assay is adequately sensitive and reliable to determine compliance with therapeutic or chemoprophylactic regimens. Although the results of a qualitative urine test are sufficient for the purposes of many field studies, quantitative tests can help to estimate the Cq con- centration in blood as well (4, 5, 13). Finger-stick blood taken at the time of the urine sample can be examined later in a field laboratory (14) or a more fully equipped laboratory (10). Selection of blood samples for analysis would depend on the parasito- logical and urine assay results. S-S/CqII is a reliable method for quantifying Cq + metabolites in urine in the range of 0 to 8 ug/ml, the test results correlating highly with the sum of Cq and DECq concentration in urine as determined by HPLC. In summary, for measuring Cq in urine, the S-S/CqI assay yields results which correlate excel- lently with the Haskins MMII test, one of the most sensitive colorimetric tests to date, and is as simple to run as the Dill-Glazko test, but is more sensitive and reliable. The average cost of reagents per urine test in our experience is US$ 0.02 for the S-S/CqI test, $0.005 for the Dill-Glazko test, and $0.05 for the Haskins MMII test. The companion method, S-S/CqII, provides the option of quantifying Cq + metabolites but demands slightly greater effort and with a modest compromise in sensitivity. As demonstrated by our study, the S-S/Cq methods can be used in field studies instead of the Dill-Glazko method. Acknowledgements The authors thank Dr O.J. Ekanem, Director, National Malaria Service, Nigeria, for permission to publish their findings. They are also grateful to the staff in the health centres in Igbo Ora and Oban for collecting specimens, and to Dr Katrina Hedberg, Dr Norman B. Molta and Mr James E. Imoni for assisting in the field laboratory testing of urine. This work was supported in part by the Agency for International Development, Africa Child Survival Initiative-Combatting Childhood Communicable Dis- eases Project-698-0421 and in part by the World Health Organization through a WHO Contract Technical Agree- ment and the UNDP/World Bank/WHO Special Prog- ramme for Research and Training in Tropical Diseases. Resume Adaptations du test de Saker-Solomons: methodes colorimetriques simples et flables pour le dosage sur le terrain de la chloroquine et de ses metabolites dans les urines Deux methodes de detection de la chloroquine (Cq), un antipaludique, et de ses metabolites (principalement la desethylchloroquine (DECq)) dans les urines, adaptees au travail sur le terrain, ont e mises au point et validees. Ces methodes sont des variantes du test de Saker-Solomons (S-S) qui sert au depistage dans les urines de cer- tains medicaments susceptibles de donner lieu a des abus. A partir de l'urine tamponnee (pH 8), la chloroquine est extraite dans une solution d'ester ethylique de tetrabromophenolphtaleine (TBPEE) WHO Bulletin Vol. 67 1989 299 D.L. Mount et al. dans le chloroforme, qui vire au violet. L'intensite de la coloration est proportionnelle a la somme des concentrations de la Cq et de ses metabolites. Dans la premiere methode (S-S/Cql), on compare a l'oeil nu les echantillons et les temoins; c'est une methode aussi simple a appliquer que le test de Dill-Glazko, mais elle est plus fiable et plus sensible, avec une concentration limite decelable de 1,0 pg/ml. La seconde methode (S-S/Cqll) permet au moyen d'un photometre a filtre (565 nm) alimente par piles de quantifier sur le terrain la Cq et ses metabolites avec une concentration limite de 2,0 pg/mI. Le proguanil, la quinine et la meflo- quine donnent egalement des resultats positifs avec ces epreuves. La methode S-S/Cql a ete validee au cours d'une etude effectuee sur le terrain au Nigeria en comparant ses resultats et ceux du test de Dill- Glazko aux resultats obtenus avec le dosage de Haskins MMII, une methode colorimetrique clas- sique. Les resultats de la S-S/Cql concordaient avec ceux de la methode de reference, contraire- ment a ceux de 1'epreuve de Dill-Glazko. La seconde methode a ete validee au laboratoire en comparant ses resultats avec ceux obtenus pour des echantillons identiques analyses par chroma- tographie liquide a haute performance (HPLC). Jusqu'a 8 pg/mI les valeurs obtenues avec la S-S/ Cqll, comparees a la somme des concentrations de Cq et de DECq obtenues par HPLC, suivent une droite de regression lineaire donnee par l'equation y = 0,989x- 0,075 (n = 14), avec un coefficient de correlation de 0,9274. On avait besoin d'une methode de terrain pour rechercher la presence de la Cq et de ses metabolites chez les sujets susceptibles de partici- pera des essais in vitro de pharmacosensibilite et pour verifier l'observance du traitement au cours de ces essais et en chimioprophylaxie. La me- thode S-S/Cql repond a ces criteres et, bien qu'un peu plus complexe et legerement moins sensible, la methode S-S/Cqll permet de doser la Cq et ses metabolites. On a demontre l'utilite de ces variantes de la methode de Saker-Solomons pour le dosage de la chloroquine lors de l'analyse d'echantillons d'urine provenant d'un eslai realise sur le terrain au Nigeria. Ces deux epreuves peuvent etre pratiqu6es en l'absence de raccordement au secteur. References 1. Lellveld, J. & Kortmann, H. The eosin colour test of Dill and Glazko: a simple field test to detect chloro- quine in urine. Bulletin of the World Health Organiz- ation, 42: 477-479 (1970). 2. Haskins, W.T. A simple qualitative test for chloro- quine in urine. American journal of tropical medicine and hygiene, 7: 199-200 (1958). 3. Wilson, T. & Edeson, J.F.B. Studies on the chemo- therapy of malaria. Ill. Treatment of acute malaria with chloroquine. Medical journal of Malaya, 9: 115-131 (1954). 4. Bergqvist, Y. et al. Determination of chloroquine and its metabolites in urine: a field method based on ion- pair extraction. Bulletin of the World Health Organiz- ation, 63: 893-898 (1985). 5. Mount, D.L. et al. Colorimetric and thin-layer chro- matographic methods for field assay of chloroquine and its metabolites in urine. Bulletin of the World Health Organization, 65: 615-623 (1987). 6. Verdler, F. et al. Unreliability of Dill-Glazko test in detecting chloroquine in urine. Lancet, 1: 1281-1283 (1985). 7. Rombo, L. et al. Reliability of Dill-Glazko test. Lancet, 1: 1509 (1985). 8. Rombo, L. et al. Evaluation of three qualitative tests for the detection of chloroquine in urine: agreement with plasma concentrations determined with liquid chromatography. Annals of tropical medicine and parasitology, 80: 293-298 (1986). 9. Saker, E.G. & Solomons, E.T. A rapid inexpensive presumptive test for phencyclidine and certain other cross-reacting substances. Journal of analytical toxi- cology, 3: 220-221 (1979). 10. Patchen, L.C. et al. Analysis of filter-paper-absorbed, finger-stick blood samples for chloroquine and its major metabolite using high-performance liquid chromatography with fluorescence detection. Journal of chromatography, 278: 81-89 (1983). 11. Mount, D.L. et al. Field-adapted method for high- performance thin-layer chromatographic detection and estimation of chloroquine and desethylchloro- quine in urine. Journal of chromatography, 423: 261-269 (1987). 12. Rio, J.G. & Hodneft, C.W. Evaluation of a colorimetric screening test for basic drugs in urine. Journal of analytical toxicology, 5: 267-269 (1981). 13. Steketee, R.W. et al. Field application of a colorimet- ric method of assaying chloroquine and its metabo- lites in urine. Bulletin of the World Health Organization, 66: 485-490 (1988). 14. Mount, D.L. et al. Field-adapted method for high- performance, thin-layer chromatographic detection and estimation of chloroquine in finger-stick blood. Journal of chromatography, 428: 196-202 (1988). 300 WHO Bulletin Vol. 67 1989