Aug 24, 1979 - was purchased from Aldrich Chemical Co., Milwaukee, ... 65881), Mallinckrodt Chemical Works, St. Louis, Mo.; ... Analytical procedures.
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Nov. 1979, p. 579-583 0066-4804/79/1 1-0579/05$02.00/0
Vol. 16, No. 5
Rapid Assay for Determination of Trimethoprim and Sulfamethoxazole Levels in Serum by Spectrofluorometry DIANE M. LICHTENWALNER, BYUNGSE SUH,* BENNETT LORBER, AND ALAN M. SUGAR Section of Infectious Diseases, Temple University Health Sciences Center, Philadelphia, Pennsylvania 19140 Received for publication 24 August 1979
A rapid spectrofluorometric method for determining the levels of both trimethoprim and sulfamethoxazole from the same specimen of serum is described. The method involves stepwise extraction of the specimen first with chloroform at an alkaline pH (pH 9.0) for trimethoprim followed by n-butyl chloride at an acidic pH (pH 2.0) for sulfamethoxazole. To quantitate trimethoprim, the chloroform layer was subjected to fluorometry by exciting the specimen at 295 nm and measuring the relative intensity at 330 nm. To determine sulfamethoxazole levels, the n-butyl chloride layer was subjected to fluorometry by exciting the specimen at 285 nm and measuring the relative intensity at 330 nm. Relative intensities were linear (r > 0.99) over the concentration ranges of 0.5 to 40 jig/ml for trimethoprim and 1 to 400 tig/ml for sulfamethoxazole. Values obtained by this spectrofluorometric procedure were in excellent agreement with those obtained by a conventional fluorometric assay for trimethoprim and a colorimetric assay for sulfamethoxazole. Elevated levels of endogenous metabolic products and numerous other drugs, including a number of antimicrobial agents, did not interfere with the method. Although salicylates interfere with the determination of sulfamethoxazole, an appropriate correction can be made. This method can also be used to determine the drug levels in cerebrospinal fluid.
Trimethoprim-sulfamethoxazole (TMPSMZ) is a well-recognized combination preparation and its clinical application is continuously increasing. Although it is best known as an agent for the treatment of chronic urinary tract infections, it has also been used for prophylaxis of Pneumocystis carinii pneumonia in immunosuppressed pediatric populations with promising results (7). It may also become a useful prophylactic antibiotic for hospitalized granulocytopenic patients (6). The preparation is usually administered by an oral route, although a parenteral preparation of TMP-SMZ is under investigation (5). Side effects are relatively few, and blood level determinations are usually not mandatory. Blood levels, however, depend upon the efficiency of gastrointestinal absorption, which may be significantly compromised in patients with shock or congestive heart failure or in those receiving vasoconstrictors (1, 4). Under these circumstances, blood level measurements may become important to ensure adequate antimicrobial activity. To date, TMP has been assayed by the fluorometric method of Schwartz et al. (11). This procedure involves chloroform extraction of tissue fluids at a basic pH, followed by back extraction into dilute sulfuric acid and oxidation by potassium permanganate to trimethoxyben-
zoic acid. The fluorescence of trimethoxybenzoic acid in chloroform is used to quantitate TMP. Nonacetylated sulfonamides including SMZ have usually been measured by the colorimetric method of Bratton and Marshall (2). We report a rapid and simple spectrofluorometric assay method which is easily applicable to the measurement of both TMP and SMZ. MATERLA1S AND METHODS Chemicals. SMZ (lot 754017) and TMP (lot 150) were supplied by Hoffmann-LaRoche, Inc., Nutley, N.J. Acetylsalicylic acid (ASA), bilirubin (lot 127C0446), creatinine phosphate (lot 55C-0163), and human hemoglobin type IV (lot 118C-4045) were obtained from Sigma Chemical Co., St. Louis, Mo. Spectroquality n-butyl chloride (analyzed by infrared spectroscopy and gas chromatography, 97%; lot HC030087) was purchased from Aldrich Chemical Co., Milwaukee, Wis., and spectrophotometric-grade chloroform (lot HAN) was from Mallinckrodt, Inc., St. Louis, Mo.
Other chemicals and antimicrobial agents, listed by their generic names, lot numbers where available, and suppliers, were: sulfisoxazole (lot 449064) and diazepam (lot 633116), Hoffmann-LaRoche, Inc.; acetaminophen (lot 738-7467) and erythromycin base (lot 75138CD), Abbott Laboratories, North Chicago, Ill.; chloramphenicol (lot 380031) and phenytoin (lot 415781), Parke, Davis & Co., Detroit, Mich.; sodium salicylate (lot 62164), Merck & Co., Inc., Rahway, N.J.; urea (ultra pure, lot ZZ1473), Schwartz/Mann, Or579
580
LICHTENWALNER ET AL.
angeburg, N.J.; gentamicin (lot GRA-2), Roerig-Pfizer, New York, N.Y.; digitoxin, Nutritional Biochemicals Corp., Cleveland, Ohio; phenobarbital, USP (lot 65881), Mallinckrodt Chemical Works, St. Louis, Mo.; and reagent-grade phosphoric acid (lot 527958), J. T. Baker Chemical Co., Phillipsburg, N.J. Analytical procedures. (i) Spectrofluorometric assay of TMP and SMZ. To 2-ml serum samples in 15-ml graduated, ground-glass stoppered centrifuge tubes was added 0.3 ml of 0.1 M glycine-NaOH buffer (pH 9.5). The solutions were mixed on a Vortex for a few seconds. After 5 min at room temperature, 4.5 ml of chloroform was added, and the tubes were shaken with a to-and-fro motion of approximately 100 motions per min on a mechanical shaker (Dubnoff metabolic shaking incubator, model P/S, Precision Scientific Co., Chicago, Ill.) for 5 min. The samples were then centrifuged for 10 min (2,000 x g) at room temperature in an International centrifuge, model PR-6. Four-milliliter aliquots of the organic solvent layer (lower) were transferred to clean, ground-glass stoppered, 15-ml, graduated centrifuge tubes and used to determine the relative intensity of fluorescence of TMP. Two milliliters of the chloroform-extracted aqueous phase was transferred to clean 15-ml, graduated, ground-glass stoppered centrifuge tubes to which 4.5 ml of 2.0 M phosphoric acid-NaOH buffer (pH 1.95) was added. The solutions were mixed on a Vortex for a few seconds. After 5 min at room temperature, 4.5 ml of n-butyl chloride was added, and the tubes were gently shaken on a mechanical shaker for 7.5 min. The samples were then centrifuged for 10 min (2,000 x g) at room temperature in an International centrifuge. Four-milliliter aliquots of the organic solvent layer (upper) were transferred to clean, ground-glass stoppered, 15-mi, graduated centrifuge tubes and were used to determine the relative intensity of fluorescence of SMZ. Chloroform extracts were excited at 295 nm, and the relative intensities at 330 nm were measured to determine the levels of TMP, using a Perkin-Elmer MPF-3 spectrofluorometer (Perkin-Elmer, Norwalk, Conn.). To determine SMZ levels, the n-butyl chloride extracts were excited at 285 nm and the relative intensities at 330 nm were measured. The slit widths for both the excitation and emission were set at 10 nm, and the sensitivity of the recorder was set at 1.0. Standards were prepared with pooled normal human serum to contain 0, 0.5, 1, 2.5, 5, 10, 20, and 40 ug of TMP per ml and 0, 1, 2.5, 5, 10, 20, 40, 60, 80, 100, 200, and 400 yg of SMZ per ml. Sample standard curves for TMP and SMZ are presented in Fig. 1 and 2, respectively. A linear relationship existed between both TMP levels and intensity (r > 0.99) and SMZ levels and intensity (r > 0.99) in the concentration ranges of 0.5 to 40 lAg of TMP per ml and 1 to 400 ,ug of SMZ per ml, respectively. The coefficients of variation for TMP and SMZ were determined using ten samples of each of three different concentrations. The average coefficient of variation for TMP (1 to 20 jLg/ ml) was 3.3%. Individual coefficients of variation of 4.04, 2.13, and 3.85% were obtained for 1, 10, and 20 ug of TMP per ml, respectively. The average coefficient of variation for SMZ (10 to 100 Ag/ml) was 2.2%. Individual coefficients of variation of 4.8, 0.63, and 1.26% were obtained for 10, 50, and 100 ,ug of SMZ per
ANTIMICROB. AGENTS CHEMOTHER.
30-
_20 z0
-
10 0
-,
0 25 5 10 20 30 TIP CONCENTRATION IN
40
pg/ml
FIG. 1. Standard curve for serum TMP levels by spectrofluorometry. The maximum relative intensity was measured at 330 nm with excitation at 295 nm. The relative intensity obtained from the chloroform layer of a serum specimen without TMP was used to correct for background fluorescence (1.54 ± 0.02 relative intensity units).
150-
- 100-
0
/
m
0 25 50
100
200
SMZ CONCENTRATION IN pgi/l FIG. 2. Standard curve for serum SMZ levels by spectrofluorometry. The maximum relative intensity was measured at 330 nm with excitation at 285 nm. The relative intensity obtained from the n-butyl chloride layer of a serum specimen without SMZ was used to correct for background fluorescence (18.125 ± 0.125 relative intensity units). ml, respectively. When different concentrations of each drug alone were analyzed by the stepwise extraction procedure, there was no evidence that SMZ is detected in the TMP assay and vice versa. The ex-
ASSAY FOR TMP AND SMZ LEVELS IN SERUM
VOL. 16, 1979
traction efficiencies for TMP and SMZ were 96.1 and 95.8%, respectively. The spectrofluorometric assay for TMP described by Schwartz and co-workers (11) and the colorimetric assay technique of Bratton and Marshall without hydrolysis (2) were used as the reference methods. (ii) Determination of serum SMZ levels in the presence of ASA. Quantitation of serum SMZ levels was interfered with by ASA and salicylate (SL) ion. The fluorescent nature of SL is well established (3). Under the present assay conditions, both ASA and SL when excited at 285 nm give maximum fluorescence emission at 450 nm. The emission peak is broad and exhibits some intensity over background at 330 nm. To determine the degree of interference at 330 nm due to the presence of SL, as well as to establish a correction factor for the presence of SL, pooled normal human serum was adjusted to contain either 0, 50, 100, 200, or 400 ug of SL per ml or 0, 20, 50, or 100 jg of SMZ per ml, or a combination of SMZ and SL. Samples were extracted and analyzed as described above, with excitation at 285 nm and measurement of emission peaks at both 330 and 450 nm. Because the emission spectrum of SMZ revealed no fluorescence over background at 450 nm (Fig. 3), it was possible to establish a ratio between the emission peaks at 330 and 450 nm, when only SL was present in the serum. SL levels were also determined by the colorimetric method of Trinder (12). (iii) Other assay methods. Serum cholesterol levels were determined by the method of Levy (8), and 80-
581
triglycerides were estimated according to the procedure of Loeffler and McDougald (10).
RESULTS Determination of TMP and SMZ levels in human serum specimens. Ten blood specimens were obtained from patients on TMP-SMZ therapy, and the levels of TMP and SMZ were determined by both the new spectrofluorometric method and the conventional assays described above. Serum specimens were used, and the results are shown in Table 1. Excellent correlation was demonstrated between the assay techniques (r = 0.995 for TMP and 0.997 for SMZ). In five patients, we obtained cerebrospinal fluid and performed the assays, and the results are included in Table 1. Effect of endogenous metabolic products on the assay procedure. Bilirubin (500 ,ug/ ml), urea (400,ug/ml), creatinine phosphate (150 ,ug/ml), and hemoglobin (150 Ag/ml) were added to serum specimens to determine whether they would interfere with the assay procedure. TMP and SMZ levels studied in the presence of these compounds ranged from 0 to 10 and 0 to 100 ltg/ ml, respectively. There was no evidence that these metabolic products interfered with the assay procedure. Elevated serum levels of cholesterol (405 mg/dl) and triglycerides (338 mg/ dl) did not interfere with the assay procedure. TABLE 1. TMP and SMZ levels in serum and cerebrospinal fluid specimens as deternined by a new spectrofluorometric method and conventional methods TMP levels
SMZ levels
(JAg/mI)
Specimen -40 Serum
D
20-
c
A
0I 300
400 WAVELENGTH
500 IN
560
nm
FIG. 3. Fluorescence emission spectra ofSMZ and SL in n-butyl chloride extracts of human sera. Curve A represents the n-butyl chloride extract of normal pooled human serum; curve B, that of serum containing SMZ (50 pug/ml); curves C, D, and E, those of serum specimens containing SL in concentrations of 50, 100, and 200 jig/ml, respectively. The samples were excited at 285 nm, and the fluorescence emission was scanned from 300 to 560 nm. SMZ and SL exhibited emission maxima at 330 and 450 nm, respectively.
Cerebrospinal fluid
New New Conventional method method method, 1.5 2.0 1.3 9.1 ND 1.0 2.3 5.0 4.9 18.8 2.2
1.3 2.0 1.3 7.9 ND 1.2 2.5 4.9 5.1 19.9 2.3
0.8 1.8 0.2 3.0 7.7
0.7 1.8 0.3 3.4
ND" 74.6 15.0 194.3
27.0d 59.6
50.6" 152.3 157.4 322.3
49.0" 77.0 166.7 38.6 142.6 195.0
(,ug/ml) Conventional methodb
ND 73.4 16.4 198.3 26.5 56.3 49.9 152.9 151.4 319.0 46.4 79.0 162.8 40.7 141.5
191.2 7.8 a The spectrofluorometric method of Schwartz and co-workers (11). 'The colorimetric method of Bratton and Marshall (2). 'ND, Not determined. d The specimens contained SL; appropriate corrections have been made for SL as described in the text.
582
ANTIMICROB. AGENTS CHEMOTHER.
LICHTENWALNER ET AL.
Effect of common therapeutic agents on the assay procedure. To determine potential interference with the spectrofluorometric assay for TMP and SMZ due to the presence of other therapeutic agents, many commonly used drugs were added to the system. The drugs and concentrations tested included carbenicillin (200 ug/ ml), chloramphenicol (30 pg/ml), erythromycin (20 ,g/ml), gentamicin (20 ug/ml), acetaminophen (50 ug/ml), ASA (500 pg/ml), diazepam (1 pug/ml), digitoxin (25 ng/ml), phenobarbital (50 ,ug/ml), phenytoin (20 ,ug/ml), and sodium salicylate (200 pg/ml). TMP and SMZ levels tested again ranged from 0 to 10 and 0 to 100 pg/ml, respectively. There was no evidence of any significant interference by the presence of drugs tested except for SL (ASA and sodium salicylate) in the determination of SMZ. Effect ofSL on the assay procedure. When it was shown that SL interfered with the quantitation of SMZ, the problem was further pursued to determine whether the method could be used to assay for levels of SMZ when SL was present. When excited at 285 nm, the maximum intensity of fluorescence of SL in n-butyl chloride occurs at 450 nm. The emission peak is broad and exhibits some intensity over background at 330 nm (Fig. 3). When SL was present in serum at 0, 50, 100, 200, or 400 pg/ml, the ratio between the intensity of fluorescence at 330 nm and that at 450 nm was calculated for each SL level. The average of these ratios was shown to be constant at 0.145. From this ratio, the following equation was constructed to determine SMZ levels when SL was present: relative intensity (RI)Sz = RItZSL-(RIa x 0.145). To determine the validity of this equation, pooled normal human serum was adjusted to contain known amounts of SL, SMZ, or SL and SMZ as described in Materials and Methods. The above equation was used to calculate SMZ levels in which both SMZ and SL were present. There was excellent agreement between the measured and known amounts of SMZ in samples in which both were present. DISCUSSION This study describes alternative methods for quantitating serum levels of TMP and SMZ using a single instrumentation, spectrophotofluorometry. A stepwise differential fractionation procedure has been used to extract TMP and SMZ into their respective organic solvent layers at pH 9.0 and 2.0. This technique is based on the fact that TMP is a weak base (pK. 7.3), whereas SMZ is a weak acid (pK. 5.6) (13). The TMP assay procedure is a modification of the method of Schwartz et al. (11), eliminating the
potassium permanganate oxidation step. This modification did not affect the sensitivity or accuracy of the method. The two methods compared favorably (see Table 1), with a correlation coefficient (r) of 0.995. The assay of SMZ by fluorometry has not been previously described, and the colorimetric method of Bratton and Marshall without hydrolysis (2) has been the most widely used technique for the determination of nonacetylated sulfonamide levels. This new spectrofluorometric assay has been shown to be quite satisfactory for the measurement of SMZ levels in clinical specimens such as serum and cerebrospinal fluid. The results obtained from both the new and standard methods were in close agreement, as demonstrated in Table 1, with a correlation coefficient (r) of 0.998. The standard colorimetric method of Bratton and Marshall without hydrolysis measures only nonacetylated sulfonamide concentrations. Since the correlation coefficient between the new and standard methods is 0.998, it can be inferred that the new method also measures nonacetylated and therefore microbiologically active SMZ. This method has only been evaluated for SMZ, but it may prove to be useful for measurements of other sulfa drugs. As described in Results, elevated levels of endogenous metabolic products and commonly used pharmacological agents did not interfere with the assay of TMP, although the SMZ assay was interfered with by the presence of SL. The interference from SL can easily be overcome by correcting for SL, using the formula shown in Results. Furthernore, we have previously demonstrated that the fluorometric method can satisfactorily be used to measure SL levels in clinical specimens if desired (9). ACKNOWLEDGMENTS We thank Tully Speaker for his kindness in permitting us to use the spectrofluorometer and Karen Schuck for manu-
script preparation. LITERATURE CITED 1. Benet, L Z., A. Greither, and W. Meister. 1976. Gastrointestinal absorption of drugs in patients with cardiac failure, p. 33-50. In L Z. Benet (ed.), The effect of disease states on drug pharmacokinetics. American Pharmaceutical Association/Academy of Pharmaceutical Sciences, Washington, D.C. 2. Bratton, A. C., and E. K. Marhall, Jr. 1939. A new coupling component for suifanlamide determination. J. Biol. Chem. 128:537-660. 3. Elevitch, F. A. 1973. Determination of salicylates in serum, p. 151-153. In L. R. Cawley (ed.), Fluorometric techniques in clinical chemistry. Little, Brown and Co., Boston. 4. FIngl, E., and D. M. Woodbury. 1975. Absorption of drugs, p. 5-9. In L S. Goodman and A. Gilman (ed.),
The pharmacological basis of therapeutics. Macmillan
VOL. 16, 1979
ASSAY FOR TMP AND SMZ LEVELS IN SERUM
Publishing Co., Inc., New York. 5. Grose, W. E., G. P. Bodey, and T. L. Loo. 1979. Clinical pharmacology of intravenously administered trimethoprim-sulfamethoxazole. Antimicrob. Agents Chemother. 15:447-451. 6. Gurwith, M. J., J. L. Brunton, B. A. Lank, G. K. M. Harding, and A. R. Ronald. 1979. A prospective controlled investigation of prophylactic trimethoprim/ sulfamethoxazole in hospitalized granulocytopenic patients. Am. J. Med. 66:248-256. 7. Hughes, W. T., S. Kuhn, S. Chaudhary, S. Feldman,
M. Verzosa, R. J. A. Aur, C. Pratt, and S. L George. 1978. Successful chemoprophylaxis of Pneumocystis carinii pneumonitis. N. Engl. J. Med. 297:1419-1426. 8. Levy, A. 1972. Triglycerides by nonane extraction and colorimetry, manual and automated. Ann. Clin. Lab.
583
Sci. 6:474-479. 9. Lichtenwalner, D. M., B. Suh, B. Lorber, and A. M. Sugar. 1979. New rapid assay for nafcillin in serum by 10. 11.
12. 13.
spectrofluorometry. Antimicrob. Agents Chemother. 16:210-213. Loeffler, H. H., and C. H. McDougald. 1963. Estimation of cholesterol in serum by means of improved technics. Am. J. Clin. Pathol. 39:311-315. Schwartz, D. E., B. A. Koechlin, and R. E. Weinfeld. 1969. Spectrofluorometric methods for the determination of trimethoprim in body fluids. Chemotherapy 14(Suppl.):S22-S29. Trinder, P. 1954. Rapid determination of salicylates in biological fluids. J. Biol. Chem. 57:301-303. Wormser, G. P. 1978. Trimethoprim-sulfamethoxazole. I. Description. N.Y. State J. Med. 78:1915-1921.
