Screening Tests for Porphobilinogen Are Insensitive

Transcription

1 Screening Tests for Porphobilinogen Are Insensitive The Problem and Its Solution WILLIAM E. SCHREIBER, M.D., AZIM JAMANI, R.T., AND MORRIS R. PUDEK, PH.D. Standard screening tests for porphobilinogen (PBG) do not turn positive until the concentration of PBG exceeds 1020 times the upper limit of normal. The authors have developed a screening procedure that uses a new ionexchange resin to separate PBG from interfering substances in urine. On addition of Ehrlich's reagent, the color is more intense than that produced in the WatsonSchwartz test, and its spectrum more closely resembles that of pure PBG. By measuring the absorbance of this solution at 555 nm, it is possible to discriminate between urine samples with 9 /*mol/l (2 mg/l) of PBG and those with no detectable PBG. The new screening test had positive results in two patients with latent acute intermittent porphyria; the WatsonSchwartz test had negative results in both cases. This procedure is easy to perform, has much greater sensitivity than the Watson Schwartz test, and uses objective spectrophotometric data to separate positive from negative results. (Key words: Porphobilinogen; WatsonSchwartz test; Ionexchange resin; Spectrophotometry; Porphyria; Screening test) Am J Clin Pathol 1989;92: THE BIOCHEMICAL SIGNATURE of the three neurologic porphyrias (acute intermittent porphyria, variegate porphyria, and coproporphyria) is an increase in the urinary excretion of porphobilinogen (PBG) during acute attacks. 5 A positive screening test result for PBG is usually followed by a quantitative analysis to confirm the diagnosis of acute porphyria. 8 However, a negative test result in the presence of signs and symptoms suggesting an acute attack rules out porphyria as a potential diagnosis. The screening test for PBG is therefore the single most important analysis in defining this type of porphyria. Most screening tests are based on the WatsonSchwartz test. 15 Urine is reacted with pdimethylaminobenzaldehyde, which forms a red color in the presence of PBG. Extraction of this mixture with one or more organic solvents removes most interfering substances (mainly urobilinogen), and the presence of a pink or red color in the aqueous layer is regarded as a positive test result. 6 A variant of this procedure, the Hoesch test, has the advantage of not reacting with urobilinogen. 7 Like other screening Received December 5, 1988; received revised manuscript and accepted for publication May 8, Address reprint requests to Dr. Schreiber: Division of Clinical Chemistry. Department of Pathology, Vancouver General Hospital, 855 W. 12th Avenue, Vancouver, British Columbia, Canada V5Z 1M9. Division of Clinical Chemistry, Department of Pathology, Vancouver General Hospital and University of British Columbia, Vancouver, British Columbia, Canada tests, these methods are designed to give a positive result when a threshold value is exceeded. Their major drawback is the need for subjective interpretation of a visual end point. During the development of a sensitive liquid chromatographic procedure for PBG, 4 we were shocked by the discovery that urine samples with up to 88 jtmol/l (20 mg/l) of PBG gave a negative screening test result by our inhouse method. 12 The reference interval for urine PBG excretion is less than 4.5 jumol (1.0 mg) per day. 6 By comparison, aqueous solutions of PBG at concentrations as low as 11 /umol/l (2.5 mg/l) gave a positive screening test result. Similar results were recently described by Deacon. 2 These data indicate that urine contains substances that interfere with color development and/or detection in conventional screening assays for PBG. To confirm our results, we sent ten urine specimens with varying PBG concentrations to four hospital laboratories in Vancouver and asked them to perform screening tests for PBG. We also decided to develop a screening test that (1) removes interfering substances in urine more completely, and (2) gives an objective readout of positive or negative results. To this end, we have modified a quantitative procedure that separates PBG from other components of urine with a new anionexchange resin. 3 The PBGcontaining eluate is then reacted with pdimethylaminobenzaldehyde, and the absorbance of this solution is measured in a spectrophotometer. We have developed objective spectrophotometric criteria that differentiate positive from negative results and increase the sensitivity of the test by fivefold to tenfold. The procedure is rapid and easy to perform and may be applied to random urine specimens. Materials and Methods Materials Porphobilinogen was purchased from Porphyrin Products (Logan, UT 84321). Paradimethylaminobenzalde 644 Downloaded from

2 Vol. 92 No. 5 hyde was from J. T. Baker Co. (Phillipsburg, NJ 08865). Polybenzimidazole (PBI) resin was obtained from the Celanese Corporation (Charlotte, NC 28232). Plastic Econocolumns (10 ml) were from BioRad Laboratories (Richmond, CA 94804). All other chemicals and solvents were reagent grade. Sensitivity Study SCREENING TEST FOR PORPHOBILINOGEN 645 This elution procedure gives an average recovery of about 90%. A second elution with 2.5 ml of 0.5 mol/l formic acid increases the recovery to 95100%, 4 but it reduces the sensitivity of the screening test. Urobilinogen is not retained by the column and is effectively separated from PBG. Liquid chromatographic analysis of the eluate from the PBI resin indicates no substances other than PBG that react with pdimethylaminobenzaldehyde. 4 One of us collected his own urine over a sixhour period without a preservative. Pure PBG was added to this urine to produce samples with PBG concentrations of 0, 11, 22, 44, 110, 221, 332, and 442 Mmol/L (0, 2.5, 5, 10, 25, 50, 75, and 100 mg/l). A 24hour urine specimen was collected from each of two patients with latent acute intermittent porphyria. The amount of PBG in each sample was verified by a liquid chromatographic procedure. 4 Each sample was randomly assigned a letter (A through J), and aliquots were stored in test tubes covered with aluminum foil at 20 C until analysis (within one week). The full set often samples was analyzed by a single person at each of the four participating laboratories. The screening methods were based on the WatsonSchwartz test and carried out as described in the corresponding references. 612 IonExchange Procedure The PBI resin was prepared for use by suspending one volume of resin in two volumes of 1 mol/l NAOH for 5 minutes, followed by washing with 40 volumes of deionized water in a glass fritted funnel. The resin was then resuspended in two volumes of 1 mol/l acetic acid for 5 minutes, followed by a second wash with 40 volumes of deionized water. The resin was stored in a glass bottle containing 5 mmol/l sodium acetate, ph 5.0. Individual columns were prepared by packing 2.5 ml of resin into a plastic Econocolumn and washing with 5 mmol/l sodium acetate, ph 5.0. Samples were prepared by adding 0.5 ml of urine to 8.0 ml of deionized water and adjusting the ph to 5.0 with 1 mol/l acetic acid (approximately one drop). The diluted urine was passed through the column, which retains PBG. The column was washed with 2.5 ml of water, followed by 2.2 ml of 0.5 mol/l formic acid (to displace the water volume in the column). Another 2.5 ml of 0.5 mol/l formic acid was then added to the column, and the PBGcontaining eluate was collected. Two milliliters of the eluate was reacted with an equal volume of Ehrlich's reagent (2.0 g of pdimethylaminobenzaldehyde dissolved in a mixture of 25 ml of concentrated HC1 and 75 ml of glacial acetic acid). Maximum color development takes place at about 5 minutes, and a pink or red color is considered a positive test result. Urine specimens that do not contain PBG appear yellow. Spectrophotometry Spectrophotometry analyses were performed with a PerkinElmer Lambda Array 3840 UV/VIS Spectrophotometer, 7300 professional computer, and PR 200 printer (PerkinElmer Corporation, Oakbrook, IL 60521). Visible spectra were obtained of the aqueous layer from the WatsonSchwartz test and of the eluates from the ionexchange column after reaction with Ehrlich's reagent. The blank contained equal volumes of Ehrlich's reagent and 0.5 mol/l formic acid. Absorbances at 525 and 555 nm were measured from the spectra and the A 5 25:A 5 55 ratio calculated for each specimen. The concentration of PBG was calculated as follows: A 555 X 5 X 4.42 PBG (jimol/l) = 01[4 = A 555 X where A555 = absorbance at 555 nm 5 = dilution factor for urine 4.42 = conversion factor from mg/l to /nmol/l = absorbance of 1 mg/l of PBG when reacted with Ehrlich's reagent Results Sensitivity of WatsonSchwartz Test Urine samples spiked with increasing concentrations of PBG were analyzed by four different hospital laboratories (Table 1). None of the laboratories detected 44 jumol/l (10 mg/l), and results were mixed at 110 jtmol/ L (25 mg/l). Urine specimens from two women with acute intermittent porphyria (latent phase) were also analyzed, and they were uniformly reported as having negative results. Comparison of WatsonSchwartz and IonExchange Methods The same urine samples used in the sensitivity study were treated by both WatsonSchwartz 6 and ionexchange procedures. We visually compared the colors produced by each method. The WatsonSchwartz test shows a weak pink color that begins to intensify at 110 /nmol/l (25 mg/ L) and turns magenta at 221 /umol/l (50 mg/l) and above. Downloaded from

3 646 SCHREIBER, JAMANI, AND PUDEK A.J.C.P. November 1989 Table 1. Sensitivity of the WatsonSchwartz Test PBG Concentration* (^mol/l) A tr Laboratory B C D tr = positive; = negative; tr = trace. Laboratories A and C used the method described in Labbe and Lamon 6 ; laboratories B and D used the method described in Rimington. 12 The lower two samples are from patients with acute intermittent porphyria, latent phase. Confirmed by liquid chromatography. 4 The conversion factor from SI to conventional units is 4.42 (imol/l = 1.00 mg/l. The tubes appear turbid rather than clear. By contrast, the ionexchange extract is yellow in the absence of PBG and begins to turn pink at 44 /imol/l (10 mg/l). The change in color is obvious at 110 /*mol/l (25 mg/l), and the classic magenta color appears at 221 jtmol/l (50 mg/ L). These tubes are clear, and the ionexchange procedure generates a more intense color. Spectrophotometric scans of these samples were also performed (Fig. 1). The absolute absorbance of the WatsonSchwartz samples was higher, but the spectra of the ionexchange samples more closely resembled the spectrum of pure PBG. Moreover, the change in absorbance between 555 nm (peak absorbance of PBG) and 620 nm (baseline absorbance of PBG) was greater for the ionexchange samples. When pure PBG dissolved in water was carried through the WatsonSchwartz procedure, the spectrum of this material had approximately twice the change in absorbance between 555 nm and 620 nm as did a similarly treated urine sample. Thus, the Watson Schwartz test does not remove interfering substances in urine that hinder spectrophotometric identification of PBG. Spectrophotometric Analysis of IonExchange Eluates Twenty urine specimens (24hour collections), all with PBG concentrations of less than 1.5 jtmol/l (0.3 mg/l) by liquid chromatography, were spiked with PBG (4,420 jtmol/l) to final concentrations of 9 and 22 jimol/l (2 and 5 mg/l). Each urine sample was run through an ionexchange column, reacted with Ehrlich's reagent, and scanned within 10 minutes in a spectrophotometer (Fig. 2). Values for the A 5 25:A 5 55 ratio appear in Table 2. At 22 /jmol/l (5 mg/l), the characteristic PBG spectrum was always recognizable, and the A 52 5 :A 555 ratio was close to that expected for the PBGEhrlich's reagent condensation product. At 9 /tmol/l (2 mg/l), the spectrum was more difficult to recognize, but a "hump" was usually visible at nm. The A 5 25:A 5 55 ratio varied more widely, but it did not exceed In urine with no detectable PBG, the spectrum was illdefined and the A 525 : A 555 ratio was greater than The concentration of PBG in each sample was also calculated from spectrophotometric data (Table 2). Urine specimens with no detectable PBG had a small but finite absorbance at 555 nm because of other chromogenic compounds that bound to the ionexchange resin. However, the calculated PBG concentration was always less than 4 ^mol/l (0.9 mg/l). When PBG was added to a final concentration of 9 /umol/l (2 mg/l), the values calculated from absorbance readings agreed fairly well and ! WAVELENGTH (nm) WAVELENGTH (nm) FIG. 1. Spectrophotometric scans of a urine sample treated by the WatsonSchwartz (A, left) and ionexchange (B, right) procedures. Both tests were performed on the same urine sample containing 110 jimol/l (25 mg/l) of PBG. The absolute absorbance of the WatsonSchwartz aqueous layer is much higher, but the change in absorbance between 555 nm (peak) and 620 nm (baseline) is lower than with the ionexchange procedure. The spectrum of the ionexchangetreated material is also smoother and more closely resembles the spectrum of pure PBG. Downloaded from

4 Vol. 92 No. 5 SCREENING TEST FOR PORPHOBILINOGEN 647 IM.8641 A URIHEUHTREflTEO IM.8573 A ^\ S WAVELENGTH (nm) 548 5(8 WAVELENGTH (nm) (8 588 (88 WAVELENGTH (nm) «C ( WAVELENGTH (nm) FIG. 2. Spectrophotometric scans of a 24hour urine sample containing 0 {A, upper, left), 9 (B, upper, right), and 22 (C, lower, left) Mmol/L (0, 2, and 5 mg/l) of added PBG, analyzed by the ionexchange method. Compare these scans with that of an aqueous solution of pure PBG at 22 jjmol/ L (5 mg/l) after reaction with Ehrlich's reagent (D, lower, right). were always more than 7 /umol/l (1.6 mg/l). At 22 jimol/ L (5 mg/l), the calculated concentrations were again close to the added amount, although slightly lower. Table 2. A 5 25:A 5 55 Ratios and Calculated PBG Concentrations at Three Different Levels of Added PBG (24hour urine specimens)* Added PBG (/jmol/l) 9 Mean SD Range Mean SD Range 22 Mean SD Range SD = standard deviation. n = 20. A525:A 55 5 Ratio Calculated PBG (jimol/l) These data suggested that we could reliably detect PBG at a concentration of 9 jimol/l (2 mg/l). We therefore selected 30 random urine specimens and spiked each to a final concentration of 9 /umol/l (2 mg/l) PBG. The ionexchange procedure was carried out on each urine specimen, both spiked and untreated, and the spectrophotometric data were recorded (Table 3). As expected, the A 5 25:A 5 55 ratio was lower in the spiked urine specimens, although there was some overlap with the untreated urine specimens. By comparison, the calculated PBG was always greater than 8 jimol/l (1.8 mg/l) in spiked samples and less than 4 jtmol/l (0.9 mg/l) in untreated samples. Patients with Acute Intermittent Porphyria We applied the ionexchange screening procedure to the 24hour urine specimens from the two patients with latent acute intermittent porphyria (Table 4). One of these patients (patient 1) had clearly positive results from the A525:A 555 ratio and calculated PBG concentration. The other patient (patient 2) had indeterminate values, but the calculated PBG was greater than expected for a urine sample containing no PBG. Downloaded from

5 648 SCHREIBER, JAMANI, AND PUDEK. A.J.C. P. November 1989 Table 3. A 52S :A 555 Ratios and Calculated PBG Concentrations in Untreated and Spiked Random Urine Specimens* Added PBG Calculated PBG (jimol/l) A 5 25:A 5 55 Ratio (^mol/l) 0 Mean SD Range Mean SD Range SD = standard deviation. n = 30. Discussion A screening test for porphobilinogen is the first step in identifying patients with a neurologic porphyria. These disorders can mimic surgical illness, and acute attacks are precipitated by a variety of drugs and toxins. 5 It is imperative, therefore, to rule out this diagnosis before proceeding with other therapies. Previous studies suggest that the WatsonSchwartz and Hoesch tests routinely give positive results at PBG concentrations greater than /tmol/l (69 mg/l). 7 ' 8 '" However, these findings are disputed in a more recent article. 2 Our own study indicates that the traditional WatsonSchwartz test is an insensitive and unreliable indicator of increased PBG in urine. There are two reasons for this. First, urine may contain substances that inhibit or mask color development when Ehrlich's reagent is added. 14 These substances will vary from one urine specimen to the next. Second, the subjective evaluation of a pink or red color depends on the observer rather than an objective standard. We have addressed the problem by using a new ionexchange resin to remove interfering substances from urine. We find that this provides a more distinct color change (yellow to pink) and generates a more intense color than does the WatsonSchwartz test. Our technologists can routinely detect 44 jtmol/l (10 mg/l), and experienced handlers can see a pink color at 22 ^mol/l (5 mg/ L). We also looked at two spectrophotometry parameters, Table 4. Porphobilinogen Screening Data in Two Patients with Acute Intermittent Porphyria, Latent Phase Calculated PBG Patient A 525 :A 5 55 Ratio (MITIOI/L) the A 52 5: A555 ratio and the PBG concentration calculated from the absorbance at 555 nm. After reaction with Ehrlich's reagent, pure PBG shows a major peak at 555 nm and a lesser peak at 525 nm (Fig. 2D). The ratio of absorbances at these wavelengths has been used as a criterion for identifying PBG. 10 This ratio proved to be a useful guide for separating PBGpositive from PBGnegative specimens, but there was some overlap between the two groups (Tables 2 and 3). However, the calculated PBG concentration separated normal urines from those with 9 jimol/l (2 mg/l) in every case. We can reliably detect this level of PBG in random or 24hour urine specimens with the use of the above criterion. We tested the new method on 24hour urine specimens from two patients with acute intermittent porphyria in the latent phase (Table 4). One patient had clearly positive results, whereas the other gave values intermediate between positive and negative. Because it is a screening test, we would regard both patients as having positive results and proceed to quantitative measurements. These patients had negative results by the WatsonSchwartz test (Table 1). It is generally recognized that PBG excretion increases significantly during acute attacks, but the degree of increase is highly variable. Some authors report very high levels of PBG in urine, even during the latent phase of acute intermittent porphyria. 13 Other articles indicate that some patients will not excrete enough PBG to be detected by the WatsonSchwartz test. 1 ' 9 Therefore, the danger exists that a subgroup of patients excreting relatively small amounts of PBG will have negative results and remain undiagnosed. This makes a strong argument for a screening test that turns positive at or near the upper reference limit for PBG. Technically, the new screening test is no more demanding than the WatsonSchwartz test, and interpretation of the result is much easier. Preparation of the ionexchange resin is straightforward, and minicolumns can be packed with resin and stored in the refrigerator until use. The procedure takes about 20 minutes from sample preparation through spectrophotometric scanning. If a scanning spectrophotometer is not available, a spectrophotometer that measures absorbance in the visible region is sufficient, so virtually any clinical laboratory can perform the assay. In summary, this new screening procedure offers five to ten times the sensitivity of the WatsonSchwartz test and eliminates subjective judgments of positive or negative results. Acknowledgments. The authors thank Dr. J. Ramirez of the Celanese Corporation for providing the PBI resin and S. Alexander and P. Tait for typing the manuscript. Downloaded from

6 Vol. 92 No. 5 SCREENING TEST FOR PORPHOBILINOGEN 649 References 1. Ackner B, Cooper JE, Gray CH, Kelly M, Nicholson DC. Excretion of porphobilinogen and 5aminolevulinic acid in acute porphyria. Lancet 1961;1: Deacon AC. Performance of screening tests for porphyria. Ann Clin Biochem 1988;25: Galbraith RA, Sassa S, Kappas A. A comparison of the utility of Dowex resin and polybenzimidazole Aurorez resin in the determination of urinary porphobilinogen concentrations. Clin Chim Acta 1987;164: Jamani A, Pudek M, Schreiber WE. Liquidchromatographic assay of urinary porphobilinogen. Clin Chem 1989;35: Kappas A, Sassa S, Anderson KE. The porphyrias. In: Stanbury JB, Wyngaarden JB, Fredrickson DS, Goldstein JL, Brown MS, eds. The metabolic basis of inherited disease. 5th ed. New York: McGrawHill, 1983: Labbe RF, Lamon JM. Porphyrins and disorders of porphyrin metabolism. In: Tietz NW, ed. Fundamentals of clinical chemistry. 3rd ed. Philadelphia: WB Saunders, 1987: Lamon J, With TK, Redeker AG. The Hoesch test: bedside screening for urinary porphobilinogen in patients with suspected porphyria. Clin Chem 1974;20: Lamon JM, Frykholm BC, Tschudy DP. Screening tests in acute porphyria. Arch Neurol 1977;34: Mahood WH, Killough JH. Acute intermittent porphyria: a clinical and laboratory study of a large family. Ann Intern Med 1966;64: Moore DJ, Labbe RF. A quantitative assay for urinary porphobilinogen. Clin Chem 1964; 10: Pierach CA, Cardinal R, Bossenmaier I, Watson CJ. Comparison of the Hoesch and the WatsonSchwartz tests for urinary porphobilinogen. Clin Chem 1977;23: Rimington C. Association of Clinical Pathologists Broadsheet No. 20 (new series), November Stein JA, Tschudy DP. Acute intermittent porphyria: a clinical and biochemical study of 46 patients. Medicine 1970;49: Taddeini L, Kay IT, Watson CJ. Inhibition of the Ehrlich's reaction of porphobilinogen by indican and related compounds. Clin Chim Acta 1962;7: Watson CJ, Schwartz S. A simple test for urinary porphobilinogen. Proc Soc Exp Biol Med 1941;47:393. Downloaded from