- Email: [email protected]
Photometric determination of human serum bromide levels—a convenient biomonitoring parameter for methyl bromide exposure
Toxicology Letters 107 (1999) 155 – 159
Photometric determination of human serum bromide levels — a convenient biomonitoring parameter for methyl bromide exposure Michael Mu¨ller a,*, Peter Reinhold a, Martina Lange a, Marc Zeise b, Uwe Ju¨rgens c, Ernst Hallier a a
Department of Occupational and Social Medicine, Georg-August-Uni6ersity, Waldweg 37, D-37073 Go¨ttingen, Germany b Uni6ersidad de Santiago de Chile, Facultad Tecnologica, Programa de Gestion Tecnologica Agraria, A6. Ecuador 3769 -Casilla 442, Correo 2, 71770 -3, Chile c Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V., Bethel, Maraweg 13, D-33617 Bielefeld, Germany Accepted 31 January 1999
Abstract Methyl bromide is one of the most important pesticides for the control of insects, fungi and nematodes. Serum bromide has been proposed as a biomonitor for occupational exposure to methyl bromide. Therefore, a novel, sensitive photometric method was developed for the determination of serum bromide at concentrations relevant for such exposure. Further possible applications are monitoring of intoxication victims and halothane narcosis. Using the method we have established a mean serum bromide level of 4.13 9S.D. 1.05 mg/l (n/64) in a group of healthy female and male volunteers not knowingly exposed to bromide or bromine containing organics. Serum of a subject accidently exposed to methyl bromide revealed a bromide level of 11.5 mg/l serum, while two individuals exposed to methyl iodide had no elevated levels. A group of 30 agricultural workers showed a mean serum bromide level of 15.33 9 S.D. 1.90 mg/l at the end of the methyl bromide application season. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Methyl bromide; Human physiological serum bromide level; Photometric determination; Occupational exposure; Biomonitoring
1. Introduction The halocarbon methyl bromide is one of the most important pesticides for the control of in* Corresponding author. Tel.: + 49-551-394950; fax: +49551-396184. E-mail address: [email protected] (M. Mu¨ller)
sects, fungi and nematodes. Due to the extraordinary neurotoxicity of the compound, its use is subjected to stringent legal regulations in most countries. Serum bromide has been proposed as a biomonitor for occupational exposure to methyl bromide and other haloalkanes containing bromide, such as halothane, bromoethane and dibro-
0378-4274/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 4 2 7 4 ( 9 9 ) 0 0 0 4 2 - 9
156
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
moethane. However, the analytical methods developed hitherto have focussed on monitoring therapeutic concentrations of the antiepileptic drug bromide, which are above 600 mg/l (Boenigk et al., 1985). Therefore, a novel, sensitive photometric method was developed for the determination of serum bromide at concentrations relevant for occupational exposure. Further possible applications are monitoring of intoxication victims and halothane narcosis.
2. Materials and methods All chemicals were of analytical grade purity and purchased from Aldrich Chemical (Steinheim, FRG) or Sigma (Deisenhofen FRG). The solutions described herein were prepared fresh prior to use. Five sera of healthy volunteers were pooled and subsequently used as a standard serum pool. Photometric determinations were run on a Uvikon 922 two beam instrument (Kontron, Neufahrn, FRG). The new photometric method is based on the oxidation of bromide to bromine, which reacts with phenolsulfonphthalein to yield 3%,3%%,5%,5%%-tetrabromophenolsulfonphthalein (Fig. 1). The reaction has been described previously (Lange and Vejdelek, 1980), but major changes were necessary to improve sensitivity: A total of 100 ml distilled water or 100 ml diluted bromide standard solution were added to 400 ml human serum (pool/sample). For calibration, the 50 mM sodium bromide standard stock solution was diluted to contain 12.5; 25; 50; 100;
Fig. 1. Reaction of phenolsulfonphthalein with bromine generated from bromide to yield 3%,3%%,5%,5%%-tetrabromophenolsulfonphthalein.
200 nmol bromide/100 ml, respectively. The serum proteins in each sample were precipitated by the addition of 100 ml trichloroacetic acid (30%), followed by an overnight incubation at 37°C and centrifugation at 8000× g for 30 min. A 50 ml phenolsulfonphthalein solution (25 mg dissolved in 105 ml 1 M sodium hydroxide and diluted with distilled water to a final vol. of 200 ml) were added to 100 ml of the serum supernatant from the serum pool, subsequently diluted with 550 ml phosphate buffer (3 g KH2PO4, 0.2 g Na2HPO4 × 2 H2O/100 ml, pH 5.5) and mixed to yield the photometric reference blank. For calibration, 50 ml phenolsulfonphthalein solution, 400 ml phosphate buffer and 75 ml chloramine-T solution (0.5%) were added to 100 ml of the serum supernatant from the serum pool and mixed. The reaction was terminated after 1 min by the addition of 75 ml sodium thiosulfate solution (2.5%). Absorption was determined at 589 nm versus the photometric reference blank. This determination was carried out twice and the resulting absorption values were averaged. Samples containing serum pool and defined amounts of bromide were assayed according to the same procedure. The mean basal serum bromide absorption value from the first determination was subtracted from the mean absorption values with bromide addition, thus yielding a calibration curve corrected for the physiological serum pool bromide level (Fig. 2). Patient sera were analyzed following the procedure for the serum pool sample without bromide addition. Serum bromide levels were calculated with the mean absorptions obtained using the calibration curve. The linear range of the method covered 2–40 mg bromide/l serum with a limit of detection of 1 mg/l. Patient sera from different sources were acquired and analyzed: for validation of the method five sera of epilepsy patients treated with potassium bromide were supplied by Dr. Ju¨rgens (Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V., Bethel, FRG). Three sera of persons accidently exposed to methyl bromide and methyl iodide were sent for analysis from the Poison Center of the University of Paris (France), and 30 sera of agricultural workers exposed to
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
157
Fig. 2. Calibration curve for bromide determination. Each point represents the mean absorption value of two separate determinations corrected for the mean physiological absorption value of the serum pool. Each bromide concentration was calibrated five times to demonstrate variations under routine measurement conditions. Table 1 Interlaboratory validation of the method Patient sera
Bromide (g/l) gold chloride (standard method)
Bromide (g/l) 3%,3%%,5%,5%-tetrabromophenolsulfonphthalein
Recovery (%)
1 2 3 4 5
1.35 1.35 2.39 1.15 1.55
1.15 1.29 2.24 1.00 1.45
85.2 95.6 93.7 87.0 93.5
methyl bromide were supplied by Dr. M. Zeise (University of Santiago de Chile, Chile). A control group of 64 healthy female and male volunteers was recruited from laboratory and hospital personnel in Go¨ttingen and their sera were analyzed. All work was done in accordance with the University of Go¨ttingen Medical Ethics Committee guidelines based on German and International law (Declaration of Helsinki).
3. Results and discussion The novel photometric method was evaluated for sensitivity and specificity. For an interlaboratory validation, bromide levels of sera of epilepsy
patients were determined by the gold chloride photometric standard method (Lange and Vejdelek, 1980) at the Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V. (Bethel, FRG). This method is used on a routine basis for bromide levels above 75 mg/l. Serum bromide levels determined in the patients ranged from 1.35 to 2.39 g bromide/l. As the novel method is about 30-fold more sensitive, the patient sera were diluted with serum pool and analyzed. After correction for the serum pool dilution we had recoveries of 85–96% as compared to the gold chloride method (Table 1). The method was also evaluated for interference by other halogen ions. As expected, fluoride and chloride ions did not interfere, whereas iodide
158
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
gave false positive readings. Due to its rapid absorption in humans, iodide levels in human sera range from 5–25 mg/l serum (Ingbar, 1989). These concentrations are 200 – 40 times lower than the limit of detection of this method, thus rendering any interference from iodide to be highly unlikely in routine analysis. There is limited data available on physiological bromide levels. Using our method we have established a mean serum bromide level of 4.139 S.D. 1.05 mg/l (n/64) in a group of female and male healthy caucasian volunteers not knowingly exposed to bromide or bromine containing organics (Fig. 3). This is in excellent agreement with a recent report on physiological bromide levels of 5.3 91.4 mg/l determined in whole blood by wavelength-dispersive X-ray fluorescence spectrometry in a group of 183 healthy Australian individuals (Olszowy et al., 1998). Moreover, a urine excretion of 6.992.2 mg bromide/l (24 h urine collection, n/30) using an ion-selective electrode measurement method has been described as background level in humans (Angerer, 1978). Serum of a subject accidently exposed to methyl bromide in France revealed a bromide level of 11.5 mg/l serum, whereas two individuals exposed
to methyl iodide had no elevated serum bromide levels (4. 7 and 3.7 mg /l). Thus, no interference of iodide with the bromide determination under real exposure conditions was detected, confirming the specificity of the method for the bromide ion. In a group of 30 agricultural workers from Chile we determined a mean serum bromide level of 15.339S.D. 1.90 mg/l at the end of the methyl bromide application season (Fig. 4). This value is 3–4 times higher than the mean physiological bromide level found in the European samples. These data are in good agreement with a Japanese study on methyl bromide exposure of plant quarantine fumigators (Tanaka et al., 1991). While in the Japanese investigation the mean urinary bromide concentration of 379 non-methyl bromide exposed workers was 6.39 2.5 mg/l with 95% confidence limits of 10 mg/l; 44.6% of 251 workers exposed to methyl bromide had urinary bromide concentrations exceeding 10 mg/l. Thus, assuming parallel serum and urinary bromide levels as indicated by our data and the reference values, the agricultural workers in Chile have to be regarded as a significantly exposed group. In conclusion, we have developed a method for a convenient photometric biomonitoring of hu-
Fig. 3. Determination of physiological bromide levels in a group of 64 healthy female and male Caucasians (mean serum bromide level: 4.13 9 S.D. 1.05 mg/l).
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
159
Fig. 4. Determination of bromide levels in a group of 30 agricultural workers from Chile exposed to methyl bromide (mean serum bromide level: 15.33 9S.D. 1.90 mg/l).
man serum bromide levels. Monitoring of bromide levels may be of interest as an indicator for intoxication or, due to the long biological half time life of the bromide ion in blood of 12 – 20 days (So¨remark, 1960; Angerer, 1978), for chronic exposure to bromine containing halocarbons, bromide or bromine itself. The method is sufficiently sensitive and specific to distinguish between physiological levels and environmental and occupational exposure.
References Angerer, J., 1978. Bromid. In: Henschler, D. (Ed.), Deutsche Forschungsgemeinschaft, Analytische Methoden zur Pru¨fung gesundheitsscha¨dlicher Arbeitsstoffe, Analysen in bi-
.
ologischem Material, VCH Verlagsgesellschaft mbH, Weinheim. Boenigk, H.-E., Lorenz, J.H., Ju¨rgens, U., 1985. Aktuelle Erfahrungen mit bromiden zur Behandlung generalisierter Epilepsien. In: Kruse, R. (Ed.), Epilepsie 84, EinhornPresse Verlag, Reinbek, pp. 316 – 325. Ingbar, S.H., 1989. Krankheiten der Schilddru¨se. In: Straub, P.W. (Ed.), Harrison, Prinzipien der Inneren Medizin, Schwabe and Co. AG Verlag, Basel, p. 2043. Lange, B., Vejdelek, Z.J., 1980. Brom und Bromide, Photometrische Analyse. Weinheim, Verlag Chemie, pp. 331 – 335. Olszowy, H.A., Rossiter, J., Hegarty, J., Geoghegan, P., 1998. Background levels of bromide in human blood. J. Anal. Toxicol. 22, 225 – 230. So¨remark, R., 1960. The biological half-life of bromide ions in human blood. Acta Physiol. Scand. 50, 119 – 123. Tanaka, S., Abuku, S., Seki, Y., Imamiya, S., 1991. Evaluation of methyl bromide exposure on the plant quarantine fumigators by environmental and biological monitoring. Ind. Health 29, 11 – 21.
Photometric determination of human serum bromide levels — a convenient biomonitoring parameter for methyl bromide exposure Michael Mu¨ller a,*, Peter Reinhold a, Martina Lange a, Marc Zeise b, Uwe Ju¨rgens c, Ernst Hallier a a
Department of Occupational and Social Medicine, Georg-August-Uni6ersity, Waldweg 37, D-37073 Go¨ttingen, Germany b Uni6ersidad de Santiago de Chile, Facultad Tecnologica, Programa de Gestion Tecnologica Agraria, A6. Ecuador 3769 -Casilla 442, Correo 2, 71770 -3, Chile c Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V., Bethel, Maraweg 13, D-33617 Bielefeld, Germany Accepted 31 January 1999
Abstract Methyl bromide is one of the most important pesticides for the control of insects, fungi and nematodes. Serum bromide has been proposed as a biomonitor for occupational exposure to methyl bromide. Therefore, a novel, sensitive photometric method was developed for the determination of serum bromide at concentrations relevant for such exposure. Further possible applications are monitoring of intoxication victims and halothane narcosis. Using the method we have established a mean serum bromide level of 4.13 9S.D. 1.05 mg/l (n/64) in a group of healthy female and male volunteers not knowingly exposed to bromide or bromine containing organics. Serum of a subject accidently exposed to methyl bromide revealed a bromide level of 11.5 mg/l serum, while two individuals exposed to methyl iodide had no elevated levels. A group of 30 agricultural workers showed a mean serum bromide level of 15.33 9 S.D. 1.90 mg/l at the end of the methyl bromide application season. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Methyl bromide; Human physiological serum bromide level; Photometric determination; Occupational exposure; Biomonitoring
1. Introduction The halocarbon methyl bromide is one of the most important pesticides for the control of in* Corresponding author. Tel.: + 49-551-394950; fax: +49551-396184. E-mail address: [email protected] (M. Mu¨ller)
sects, fungi and nematodes. Due to the extraordinary neurotoxicity of the compound, its use is subjected to stringent legal regulations in most countries. Serum bromide has been proposed as a biomonitor for occupational exposure to methyl bromide and other haloalkanes containing bromide, such as halothane, bromoethane and dibro-
0378-4274/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 4 2 7 4 ( 9 9 ) 0 0 0 4 2 - 9
156
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
moethane. However, the analytical methods developed hitherto have focussed on monitoring therapeutic concentrations of the antiepileptic drug bromide, which are above 600 mg/l (Boenigk et al., 1985). Therefore, a novel, sensitive photometric method was developed for the determination of serum bromide at concentrations relevant for occupational exposure. Further possible applications are monitoring of intoxication victims and halothane narcosis.
2. Materials and methods All chemicals were of analytical grade purity and purchased from Aldrich Chemical (Steinheim, FRG) or Sigma (Deisenhofen FRG). The solutions described herein were prepared fresh prior to use. Five sera of healthy volunteers were pooled and subsequently used as a standard serum pool. Photometric determinations were run on a Uvikon 922 two beam instrument (Kontron, Neufahrn, FRG). The new photometric method is based on the oxidation of bromide to bromine, which reacts with phenolsulfonphthalein to yield 3%,3%%,5%,5%%-tetrabromophenolsulfonphthalein (Fig. 1). The reaction has been described previously (Lange and Vejdelek, 1980), but major changes were necessary to improve sensitivity: A total of 100 ml distilled water or 100 ml diluted bromide standard solution were added to 400 ml human serum (pool/sample). For calibration, the 50 mM sodium bromide standard stock solution was diluted to contain 12.5; 25; 50; 100;
Fig. 1. Reaction of phenolsulfonphthalein with bromine generated from bromide to yield 3%,3%%,5%,5%%-tetrabromophenolsulfonphthalein.
200 nmol bromide/100 ml, respectively. The serum proteins in each sample were precipitated by the addition of 100 ml trichloroacetic acid (30%), followed by an overnight incubation at 37°C and centrifugation at 8000× g for 30 min. A 50 ml phenolsulfonphthalein solution (25 mg dissolved in 105 ml 1 M sodium hydroxide and diluted with distilled water to a final vol. of 200 ml) were added to 100 ml of the serum supernatant from the serum pool, subsequently diluted with 550 ml phosphate buffer (3 g KH2PO4, 0.2 g Na2HPO4 × 2 H2O/100 ml, pH 5.5) and mixed to yield the photometric reference blank. For calibration, 50 ml phenolsulfonphthalein solution, 400 ml phosphate buffer and 75 ml chloramine-T solution (0.5%) were added to 100 ml of the serum supernatant from the serum pool and mixed. The reaction was terminated after 1 min by the addition of 75 ml sodium thiosulfate solution (2.5%). Absorption was determined at 589 nm versus the photometric reference blank. This determination was carried out twice and the resulting absorption values were averaged. Samples containing serum pool and defined amounts of bromide were assayed according to the same procedure. The mean basal serum bromide absorption value from the first determination was subtracted from the mean absorption values with bromide addition, thus yielding a calibration curve corrected for the physiological serum pool bromide level (Fig. 2). Patient sera were analyzed following the procedure for the serum pool sample without bromide addition. Serum bromide levels were calculated with the mean absorptions obtained using the calibration curve. The linear range of the method covered 2–40 mg bromide/l serum with a limit of detection of 1 mg/l. Patient sera from different sources were acquired and analyzed: for validation of the method five sera of epilepsy patients treated with potassium bromide were supplied by Dr. Ju¨rgens (Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V., Bethel, FRG). Three sera of persons accidently exposed to methyl bromide and methyl iodide were sent for analysis from the Poison Center of the University of Paris (France), and 30 sera of agricultural workers exposed to
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
157
Fig. 2. Calibration curve for bromide determination. Each point represents the mean absorption value of two separate determinations corrected for the mean physiological absorption value of the serum pool. Each bromide concentration was calibrated five times to demonstrate variations under routine measurement conditions. Table 1 Interlaboratory validation of the method Patient sera
Bromide (g/l) gold chloride (standard method)
Bromide (g/l) 3%,3%%,5%,5%-tetrabromophenolsulfonphthalein
Recovery (%)
1 2 3 4 5
1.35 1.35 2.39 1.15 1.55
1.15 1.29 2.24 1.00 1.45
85.2 95.6 93.7 87.0 93.5
methyl bromide were supplied by Dr. M. Zeise (University of Santiago de Chile, Chile). A control group of 64 healthy female and male volunteers was recruited from laboratory and hospital personnel in Go¨ttingen and their sera were analyzed. All work was done in accordance with the University of Go¨ttingen Medical Ethics Committee guidelines based on German and International law (Declaration of Helsinki).
3. Results and discussion The novel photometric method was evaluated for sensitivity and specificity. For an interlaboratory validation, bromide levels of sera of epilepsy
patients were determined by the gold chloride photometric standard method (Lange and Vejdelek, 1980) at the Biochemisches Labor der Gesellschaft fu¨r Epilepsieforschung e.V. (Bethel, FRG). This method is used on a routine basis for bromide levels above 75 mg/l. Serum bromide levels determined in the patients ranged from 1.35 to 2.39 g bromide/l. As the novel method is about 30-fold more sensitive, the patient sera were diluted with serum pool and analyzed. After correction for the serum pool dilution we had recoveries of 85–96% as compared to the gold chloride method (Table 1). The method was also evaluated for interference by other halogen ions. As expected, fluoride and chloride ions did not interfere, whereas iodide
158
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
gave false positive readings. Due to its rapid absorption in humans, iodide levels in human sera range from 5–25 mg/l serum (Ingbar, 1989). These concentrations are 200 – 40 times lower than the limit of detection of this method, thus rendering any interference from iodide to be highly unlikely in routine analysis. There is limited data available on physiological bromide levels. Using our method we have established a mean serum bromide level of 4.139 S.D. 1.05 mg/l (n/64) in a group of female and male healthy caucasian volunteers not knowingly exposed to bromide or bromine containing organics (Fig. 3). This is in excellent agreement with a recent report on physiological bromide levels of 5.3 91.4 mg/l determined in whole blood by wavelength-dispersive X-ray fluorescence spectrometry in a group of 183 healthy Australian individuals (Olszowy et al., 1998). Moreover, a urine excretion of 6.992.2 mg bromide/l (24 h urine collection, n/30) using an ion-selective electrode measurement method has been described as background level in humans (Angerer, 1978). Serum of a subject accidently exposed to methyl bromide in France revealed a bromide level of 11.5 mg/l serum, whereas two individuals exposed
to methyl iodide had no elevated serum bromide levels (4. 7 and 3.7 mg /l). Thus, no interference of iodide with the bromide determination under real exposure conditions was detected, confirming the specificity of the method for the bromide ion. In a group of 30 agricultural workers from Chile we determined a mean serum bromide level of 15.339S.D. 1.90 mg/l at the end of the methyl bromide application season (Fig. 4). This value is 3–4 times higher than the mean physiological bromide level found in the European samples. These data are in good agreement with a Japanese study on methyl bromide exposure of plant quarantine fumigators (Tanaka et al., 1991). While in the Japanese investigation the mean urinary bromide concentration of 379 non-methyl bromide exposed workers was 6.39 2.5 mg/l with 95% confidence limits of 10 mg/l; 44.6% of 251 workers exposed to methyl bromide had urinary bromide concentrations exceeding 10 mg/l. Thus, assuming parallel serum and urinary bromide levels as indicated by our data and the reference values, the agricultural workers in Chile have to be regarded as a significantly exposed group. In conclusion, we have developed a method for a convenient photometric biomonitoring of hu-
Fig. 3. Determination of physiological bromide levels in a group of 64 healthy female and male Caucasians (mean serum bromide level: 4.13 9 S.D. 1.05 mg/l).
M. Mu¨ller et al. / Toxicology Letters 107 (1999) 155–159
159
Fig. 4. Determination of bromide levels in a group of 30 agricultural workers from Chile exposed to methyl bromide (mean serum bromide level: 15.33 9S.D. 1.90 mg/l).
man serum bromide levels. Monitoring of bromide levels may be of interest as an indicator for intoxication or, due to the long biological half time life of the bromide ion in blood of 12 – 20 days (So¨remark, 1960; Angerer, 1978), for chronic exposure to bromine containing halocarbons, bromide or bromine itself. The method is sufficiently sensitive and specific to distinguish between physiological levels and environmental and occupational exposure.
References Angerer, J., 1978. Bromid. In: Henschler, D. (Ed.), Deutsche Forschungsgemeinschaft, Analytische Methoden zur Pru¨fung gesundheitsscha¨dlicher Arbeitsstoffe, Analysen in bi-
.
ologischem Material, VCH Verlagsgesellschaft mbH, Weinheim. Boenigk, H.-E., Lorenz, J.H., Ju¨rgens, U., 1985. Aktuelle Erfahrungen mit bromiden zur Behandlung generalisierter Epilepsien. In: Kruse, R. (Ed.), Epilepsie 84, EinhornPresse Verlag, Reinbek, pp. 316 – 325. Ingbar, S.H., 1989. Krankheiten der Schilddru¨se. In: Straub, P.W. (Ed.), Harrison, Prinzipien der Inneren Medizin, Schwabe and Co. AG Verlag, Basel, p. 2043. Lange, B., Vejdelek, Z.J., 1980. Brom und Bromide, Photometrische Analyse. Weinheim, Verlag Chemie, pp. 331 – 335. Olszowy, H.A., Rossiter, J., Hegarty, J., Geoghegan, P., 1998. Background levels of bromide in human blood. J. Anal. Toxicol. 22, 225 – 230. So¨remark, R., 1960. The biological half-life of bromide ions in human blood. Acta Physiol. Scand. 50, 119 – 123. Tanaka, S., Abuku, S., Seki, Y., Imamiya, S., 1991. Evaluation of methyl bromide exposure on the plant quarantine fumigators by environmental and biological monitoring. Ind. Health 29, 11 – 21.