Available online at www.annclinlabsci.org Annals of Clinical & Laboratory Science, vol. 45, no. 3, 2015 323 The iDigoxin Assay is More Sensitive than LOCI Assay for Rapid Detection of Convallatoxin, the Active Cardiac of Lily of The Valley

Jamie M. Everett1, Yumi A. Kojima1, Barbara Davis2, Amer Wahed1, and Amitava Dasgupta1

1Department of Pathology and Laboratory Medicine, University of Texas-Houston Medical School, Houston TX and 2Department of Pathology, Baylor St Luke’s Medical Center (TMC), CHI St Luke’s Health System, Houston, TX, USA

Abstract. Objective. Lily of the valley is a poisonous plant due to the presence of the con- vallatoxin. We compared two immunoassays (LOCI digoxin assay and iDigoxin assay) for rapid detection of convallatoxin if present in human serum. Materials and Methods. Aliquots of a drug free serum pool and a digoxin serum pool were supplemented with microliter amounts of lily of the valley extract or nanogram to microgram quantities of convallatoxin, followed by measurement of apparent digoxin concentrations using the LOCI and iDigxoin assays. Results. Apparent digoxin concentrations were observed when aliquots of a drug free serum pool were supplemented with convallatoxin or lily of the valley extract using both assays but apparent digoxin concentrations were significantly higher using the iDigoxin assay. In addition, the interference of convallatoxin in serum digoxin measurement was also significantly higher using iDigxoin assay compared to the LOCI digoxin assay. Conclusions. The iDigxoin assay is more sensitive in detecting convallatoxin in human serum.

Key words: Lily of the valley, Convallatoxin, LOCI Digoxin assay, iDigoxin assay.

Introduction component in many herbal extracts used as a tonic for heart [4]. Convallatoxin, the major cardiac gly- The lily family is composed of 280 to 300 genera coside present in lily of the valley, has -like made up of 4000 to 4600 different species but only properties by inhibition of the Na, K-ATPase there- 90 genera representing approximately 525 species by causing a positive inotropic effect [5]. are found in North America. Lilies are popular dec- orative plants and are also found in floral arrange- Convallatoxin has structural similarity with digox- ments. The Convallaria genus, commonly called in, and we reported earlier rapid detection of con- “lily of the valley” and used for decoration, is a vallatoxin in human serum using luminescent oxy- toxic plant [1]. Lily of the valley grows in the cooler gen channeling technology based digoxin assay climate of North America and also in Europe and (LOCI digoxin assay) [6]. Because iDigoxin assay certain parts of Asia. The entire plant is toxic, con- on the Architect i1000 or i2000 analyzer is also taining cardiac causing digitalis-like tox- widely used in clinical laboratory, we compared icity. The principle toxic cardiac glycoside found in iDigoxin assay with LOCI digoxin assay for detec- lily of the valley is convallatoxin. There are several tion of convallatoxin in vitro in human serum and reports of lily of the valley poisoning [2]. Symptoms here we report our findings. of digitalis-like in a family after accidental ingestion of lily of the valley plant have also been Materials and Methods reported [3]. Lily of the valley extract is used in herbal medicine despite known toxicity of the plant Convallatoxin was purchased from the Sigma Chemical since the 16th century and is still used as a Company (St. Louis, MO). Lily of the valley herbal ex- tract (Wise Woman Herbals, Creswell, OR) was ob- Address correspondence to Dr. Amitava Dasgupta, Department of tained from a local herbal store in Houston, TX. LOCI Pathology and Laboratory Medicine, University of Texas-Houston digoxin assay kits were purchased from Siemens Medical School, 6431 Fannin, MSB 2.292, Houston, TX 77030, USA; phone: 713 500 5407; fax: 713 500 0730; e mail: Amitava. Diagnostics (Deerfield, IL) and assays were run on a [email protected] Vista 1500 auto analyzer (Siemens Diagnostics)

0091-7370/15/0300-323. © 2015 by the Association of Clinical Scientists, Inc. 324 Annals of Clinical & Laboratory Science, vol. 45, no. 3, 2015

Table 1. Effect of supplementing aliquots of drug free serum pool with convallatoxin or lily of the valley extract on LOCI digoxin assay

Specimen Apparent Digoxin, ng/mL, Mean (SD), n=3 LOCI Digoxin iDigoxin

Drug free serum None detected None detected +50 ng/mL convallatoxin 0.56 (0.003) 1.18 (0.018) +100 ng/mL convallatoxin 0.86 (0.008) 1.98 (0.047) +250 ng/mL convallatoxin 1.63 (0.008) 6.89 (0.066) +500 ng/mL convallatoxin 2.43 (0.016) 13.43 (0.028) +5 µl/mL Lily of the valley herb extract 2.18 (0.007) 1.95 (0.049) +10 µl/mL Lily of the valley herb extract 3.39 (0.009) 3.62 (0.011) +20 µl/mL Lily of the valley herb extract 6.81 (0.034) 9.33 (0.004)

Table 2. Effect of supplementing aliquots of digoxin pool with convallatoxin or lily of the valley extract on serum digoxin measurement using the LOCI digoxin assay and Architect digoxin assay.

Specimen Apparent Digoxin, ng/mL, Mean (SD), n=3 LOCI Digoxin iDigoxin

Digoxin Pool 1.56 (0.064) 1.46 (0.045) +50 ng/mL convallatoxin 1.08 (0.004)* 2.10 (0.002)^ +100 ng/mL convallatoxin 1.30 (0.022)* 3.20 (0.001)^ +250 ng/mL convallatoxin 1.98 (0.003)^ 10.61 (0.016)^ +500 ng/mL convallatoxin 2.69 (0.010)^ 15.80 (0.034)^ +5 µl/mL Lily of the valley herb extract 2.80 (0.015)^ 3.24 (0.014)^ +10 µl/mL Lily of the valley herb extract 3.74 (0.006)^ 7.58 (0.032)^ +20 µl/mL Lily of the valley herb extract 6.86 (0.040)^ 11.54 (0.056)^

*Significantly less than digoxin value of the digoxin pool by independent t-test, two tailed (p<0.05) ^Significantly greater than the value of the digoxin pool by independent t-test, two tailed (p<0.05) following the manufacturer’s protocol. The LOCI tubes followed by evaporation of the solvent under a Digoxin assay utilizes a specific mouse monoclonal anti- gentle stream of nitrogen and reconstitution with an ali- body against digoxin and requires no sample pre-treat- quot of drug free serum. The apparent digoxin concen- ment prior to assessment. The analytical measurement trations were then measured using the LOCI digoxin range of this assay is from 0.06 to 5.0 ng/mL of serum assay and iDigoxin assay. Each measurement was per- digoxin concentration with a lower limit of quantifica- formed in triplicate and results were expressed as the tion of 0.06 ng/mL. The iDigoxin assay kits were pur- mean and one standard deviation. chased from Abbott Laboratories (Abbott Park, IL) and assays were run on the Architect I system analyzer Digoxin serum pools were prepared by combining de- (Architect i2000) also available from the Abbott identified serum specimens from patients receiving di- Laboratories. The assay is linear up to a serum digoxin goxin that were submitted for therapeutic drug monitor- concentration of 4.0 ng/mL and lower limit of quantifi- ing. The specimens were used after performing and cation is 0.3 ng/mL. Standard solutions of convallatoxin reporting all results to the ordering clinicians, and after were made in absolute ethanol, followed by further dilu- holding these specimens for one week as required by our tion with ethanol to prepare working standards. laboratory protocol and according to guidelines of the University of Texas-Houston Institutional Review Board. Aliquots of digoxin-free serum pools were supplemented Aliquots of the digoxin pool were supplemented with with either convallatoxin (50 ng/mL to500 ng/mL) or various amounts of pure convallatoxin or lily of the val- liquid extract of lily of the valley (5 µl/mL to 20 µl/mL). ley extract, with measurement of digoxin concentrations Microliter quantities of convallatoxin ethanol working before and after supplementation using both LOCI and solution or lily of the valley extract were added to dry test iDigoxin assay. Lily of the Valley and Digoxin immunoassays 325

Statistical analyses were performed using the two-tailed reduced to 1.08 ng/mL using the LOCI assay but student t-test. A statistically significant difference was significantly increased to 2.10 ng/mL using the considered at 95% confidence interval or higher iDigoxin assay (Table 2). (p<0.05). Discussion Results Accidental poisoning by lily of the valley in both Significant apparent digoxin concentrations were animals and humans has been reported. In addi- observed when aliquots of drug free serum pool tion, it is possible that a person taking digoxin were supplemented with either convallatoxin or liq- may also take lily of the valley herbal supplement uid extract of lily of the valley using both LOCI due to increased and widespread use of herbal and iDigoxin assay, indicating that the both digox- remedies. Therefore, a rapid diagnosis of poison- in assays could be used for rapid detection of con- ing with lily of the valley could be of high utility. vallatoxin as well as digoxin like immunoreactive The data presented here indicate that the iDigoxin components of lily of the valley herbal extract. assay is more sensitive than LOCI digoxin assay However, apparent digoxin concentrations were for indirect determination of the presence of con- significantly higher using iDigoxin assay indicating vallatoxin in the serum if the patient is not receiv- that this assay is more sensitive than LOCI digoxin ing digoxin. Digoxin is a cardiac glycoside with a assay for rapid detection of convallatoxin if present narrow therapeutic range. Therefore, it is impor- in human serum. For example, when an aliquot of tant to accurately measure serum digoxin level in drug free serum pool was supplemented with 100 patients taking digoxin. Immunoassays are widely ng/mL of convallatoxin, the mean observed appar- used for digoxin measurement in clinical laborato- ent digoxin level was 0.86 ng/mL using LOCI assay ries but digoxin assays also suffer from various in- but the value was 1.98 ng/mL using the iDigoxin terferences. Jones and Morris analyzed digoxin assay. When another aliquot of the drug free serum values in 36 plasma samples by sending aliquots to pool was supplemented with lily of the valley ex- two different laboratories using different digoxin tract to achieve a final concentration of 20 µl/mL, immunoassays (CEDIA DRI digoxin, Microgenics the observed mean apparent digoxin concentration Corporation, Fremont, CA and DGNA digoxin was 6.81 ng/mL using the LOCI assay, but 13.43 assay, Dade Behring, now Siemens Diagnostics, ng/mL using the iDigoxin assay (Table 1). Deerfield, CA) [8]. Those authors observed clini- cally significant discordance in 39% of these sam- Because cross-reactivity of a substance should be ples and commented that digoxin-like immunore- tested in the presence of the primary analyte [7], active substances interference may explain only the effect of adding convallatoxin or lily ofthe val- some of the discordance. Various other endoge- ley extract to aliquots of a digoxin pool to test cross nous and exogenous factors are known to interfere reactivity of convallatoxin in the presence of the with serum digoxin measurements using immu- primary analyte digoxin was investigated. Bi- noassays [9]. It is possible that a patient receiving directional (negative interference with smaller digoxin is also taking lily of the valley extract for amounts of convallatoxin and positive interference self-treating a heart condition. Our results indi- with higher amounts of convallatoxin) interference cate that neither LOCI nor iDigoxin assay are of convallatoxin with serum digoxin measurement suitable for therapeutic drug monitoring of digox- using the LOCI digoxin assay was observed. in in patients receiving digoxin and any herbal However, with iDigoxin assay, only positive inter- remedies containing lily of the valley extract. ference was observed regardless of convallatoxin Because of known toxicity of convallatoxin we are concentration of amount of lily of the valley extract unable to conduct human study demonstrating in added to aliquots of serum digoxin pool. For ex- vivo detection of convallatoxin using digoxin as- ample, in concentration of digoxin in the digoxin says. Our preliminary studies must be confirmed pool was 1.56 ng/mL. In the presence of 50 ng/mL in case studies involving patients poisoned with of convallatoxin, the digoxin value was significantly lily of the valley. 326 Annals of Clinical & Laboratory Science, vol. 45, no. 3, 2015

Acknowledgment We thank Chi M Fan for technical assistance.

References

1. Fitzgerald KT. Lily toxicity in the cat. Top Companion Anim Med 2010; 25: 213-217. 2. Alexandre J, Foucault A, Coutance G, Scanu P et al. Digitalis intoxication induced by an acute accidental poisoning of lily of the valley. Circulation 2012; 125: 1053-1055. 3. Edgerton PH. Symptoms of digitalis like toxicity in a family after accidental ingestion of lily of the valley plant. J Emerg Nurs 1989; 15: 220-223. 4. Haass LF. Convallari majalis (lily of the valley) (also known as our lady’s tears, ladder to heaven). J Neurol Neurosurg Psychiatry 1995; 59: 367. 5. Choi DH, Kang DG, Cui X, Cho KW et al. The positive ino- tropic effect of the aqueous extract of Convallaria keiskei in beating rabbit atria. Life Sciences 2006; 79: 1178-1185. 6. Welsh KJ, Huang RS, Actor JK, Dasgupta A. Rapid detection of the active cardiac glycoside convallatoxin of the lily of the valley using LOCI digoxin assay. Am J Clin Pathol 2014; 142; 307-312. 7. Miller JJ, Valdes R. Methods for calculating cross reactivity in immunoassays.J Clin Immunoassay 1992; 15: 97-100. 8. Jones TE, Morris RG. Discordant results from “real-world” patient samples assayed for digoxin. Ann Pharmacother 2008; 42: 1797-1803. 9. Dasgupta A. Therapeutic drug monitoring of digoxin: impact on endogenous and exogenous digoxin-like immunoreactive substances. Toxicol Rev 2006; 25: 273-281.