Increased Serum Formate in the Diagnosis of Methanol Poisoning

0 downloads 0 Views 253KB Size Report
formate was a sensitive indicator of methanol poisoning. Our results proved .... units), Conversion factors from mmol/L to mg/dL for formic acid and lactic acid areĀ ...
Journal of Analytical Toxicology, Vol. 29, September 2005

]CaseReport

Increased Serum Formate in the Diagnosis of Methanol Poisoning Knut Erik Hovda TM, Petter Urdal 2, and Dag Jacobsen 1

1Departmentof Acute Medicine and 2Departmentof Clinical Chemistry, Ullevaal University Hospital, N0-0407 Oslo, Norway

Abstract Early diagnosis is essential for successful treatment in methanol poisoning. Methanol detection by gas chromatography is not available in most hospitals. Methanol increases the osmolal gap in serum and its metabolite formate increases the anion gap. The sensitivity of these indirect diagnostic methods is not good at low concentrations of methanol or formate. We therefore studied the usefulness of formate measurement in diagnosing methanol poisoning. In 15 patients poisoned with methanol, serum formate was measured enzymatically on a Cobas Mira analyzer using formate dehydrogenase and nicotinamid adenine dinucleotid. Day-to-day coefficient of variation was 5%, and the upper reference limit was 2 mg/dL (0.4 mmol/L). Methanol was detected in all 15 patients of whom 14 had elevated serum formate concentrations. Anion gap was increased in 11 of 11, and osmolal gap in 11 patients of 15 examined. Metabolic acidosis was present in 12 of 15 patients, but pH was below 7.30 in only 9 of them. Four patients with no symptoms had formate concentrations in the range 2-38 mg/dL (0.5-8.3 mmol/L), indicating that increased serum formate was a sensitive indicator of methanol poisoning. Our results proved formate analyzes to be a simple, sensitive, and specific way of diagnosing methanol poisoning. Confounders are patients admitted early, or concomitant ethanol ingestion, and therefore no acidosis. This problem may, however, be omitted by repeated formate analysis in patients developing metabolic acidosis.

Introduction

Early diagnosis is essential for successful treatment of patients poisoned with methanol. The direct detection of methanol usually requires a gas chromatographic (GC) method, and such methods are not available on a 24 h basis in most hospitals. Methanol may be estimated indirectly through calculation of the osmolal and anion gap (1-3) because methanol increases the osmolality of serum and its metabolite formate increases the anion gap. Increase in both osmolal and anion gap indicates either methanol and/or ethylene glycol poisoning. However, serum methanol concentrations of less * Author to whom correspondenceshould be addressed.E-mail: [email protected] [email protected].

586

than approximately 65 mg/dL (20 mmol/L) might not increase the osmolal gap above the reference range (2,4), and formate must increase several times above its upper reference limit of 2 mg/dL (0.4 retool/L) before the anion gap is significantly increased. The sensitivity of these indirect methods is therefore not very good at such low concentrations. A third way of diagnosing methanol poisoning is through measurement of serum formate. This toxic metabolite of methanol may be measured directly using an enzymatic method (5-7). In the present report, we have addressed serum formate measured enzymaticallyas a fast, sensitive, secure, and easy way of obtaining the methanol poisoning diagnosis during a large outbreak.

Experimental

In a recent methanol outbreak in Norway, a total of 51 patients were hospitalized because they consumed contraband spirit containing 20% methanol and 80% ethanol. Of these, 15 patients were admitted to our hospital where formate measurements could be performed (Table I). Patients 1-10 presented with dyspnea and visual disturbances, and patients 11-15 presented without symptoms. Diagnosis was in most cases initially based on anamnesis, clinical features, and calculation of the anion and osmolal gaps. The patients were treated with alkalizing buffer, antidote (fomepizole), and hemodialysis (11 of 15). Three patients died (cases 1, 3, and 7), and two were discharged with permanent visual and/or cerebral sequelae (cases 2 and 6). One (case 6) died one year later.

Methods

Formate was measured enzymatically on a Cobas Mira analyzer (Roche Diagnostics, Basel, Switzerland) using formate dehydrogenase (Roche) and nicotinamid adenine dinucleotid (Sigma, St. Louis, MO). The method is essentially the second step of a previously published method for oxalate in urine (5), modified by preparing the reagents in a 10 g/L pH 7.4 phos-

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.

Journal of Analytical Toxicology, Vol. 29, September2005

phate buffer and doubling the sample fraction and is similar to that published by others (6). Pure sodiumformate (Merck, Darmstadt, Germany) was used to prepare a standard of 46 mg formate per liter phosphate buffer and two control sera. Day-to-day coefficient of variation was 5%, and the upper reference limit was 2 mg/dL (0.4 mmol/L). Methanol in serum was measured by GC using a headspace injector (Fisons GC 8000; Carlos Erba Instruments, Rodano, Italy) [sensitivity 4 mg/dL (1.3 mmol/L) and day-to-day coefficient of variation 5%]. Calibrators and controls were made by dilution of 100% methanol (Merck). Anion and osmolal gaps were calculated using standard formulas, reference intervals being 5-21 mmol/L (anion gap), and -9-19 mOsm/kgH20 (osmolal gap) (4).

take of methanol, where metabolism is blocked by antidote (either ethanol taken concomitantly or through fomepizole treatment), or the patient is admitted very early, the serum formate levels might be within the reference interval (7), as also seen in patient 15 (Table I). This patient drank small amounts of methanol and concomitantly also ethanol. Supplemental analyzes after 4 and 8 h would in these cases most probably be sufficient to verify or exclude the diagnosis of methanol poi50,

y = 0.712x + 17,221 / 40 ,

~

30

Results

Methanol was detected in all 15 patients of whom 14 had serum formate concentrations above the reference interval (Table I). Anion gap was increased in 11 of 11, and the osmolal 101 gap was increased in 11 of 15 patients examined. Metabolic acidosis was present in 12 of 15 patients, but the pH was below 7.30 in only 9 of them (Table I). There was no correlation between serum methanol and formate concentrations (Pearson's 0 10 20 30 40 50 r - -0.14). There was, however, a good correlation between Formate + lactate (mmol/L) serum methanot and the osmolal gap, calculated by linear regression: y = 0.95x + 14.45, R2 = 0.86 (n = 15). The correlation Figure 1. Correlation between the anion gap and formate and lactate (SI between the anion gap and serum formate and lactate was also units), Conversion factors from mmol/L to mg/dL for formic acid and good (y = 0.71x + 17.22, t72 _ 0.79, n = 6) (Figure 1), but the lactic acid are 4.6 and 9.0, respectively. anion gap and serum formate did not correlate that well (y = 0.89x + 21.22, R 2 = 0.37, n = 11). All of these equations were calculated Table I. Biochemical Findings in Blood Drawn on Admission using SI units. Conversion factors Gender/ BD* MetOH Form Lact EtOH OG AG from mmol/L to mg/dL for methanol, Case Age pH (mmol/L) (mg/dt)* (mg/dL) t (mg/dL) t (mg/dL) t (mOsm/kgH20) (mmol/L) formic acid, lactic acid, and ethanol are 3.2, 4.6, 9.0, and 4.6, respectively. I

Discussion

Most methanol-poisoned patients are unaware of their poisoning and believe they are drinking ethanol. For this reason they are seldom admitted before symptoms like dyspnoea and visual disturbances develop (3). In all our 10 patients with clinical features, the formate concentration was above 46 mg/dL (10 mmol/L). Four patients (cases 11-14) with no symptoms had formate concentrations in the range 2-38 mg/dL (0.5-8.3 mmol/L), indicating that increased serum formate is a sensitive indicator of methanol poisoning. In cases with a small in-

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

M63 M41 F31 F54 M49 M62 F46 M35 M70 M35 M44 M69 M53 F59 M67

7.13 6.79 6,34 7.31 7.15 6.63 6.62 7.27 7.38 7.28 7.29 7.33 7.5 7.4 7.4

24 29 ND 11 25 28 ND 18 7 13 15 9 5 1 2

75 104 28 50 100 248 244 27 67 108 30 329 450 26 13

148 97 53 94 80 72 63 54 51 49 38 32 15 2 1

I

ND 127' 189 64 ND 114 130 32 ND ND ND ND 31 8 11

I

0 0 0 0 0 0 0 40~ 40 90~ 0 0 40 0 0

I

56 61 49 16 22 113 94 24' 28' 50' 9 101 138' 13 6

I

ND 40 48 33 42 35 ND 28 27 25 31 24 23 ND ND

*Abbreviations:ND, not determined;BD, basedeficit;MetOH,methanol;Form,formate;Lact,lactate;EtOH,ethanol;OG, osmolalgap;and AG, aniongap. Conversionfactorsfrommmol/Lto rag/alLfor methanol,formicacid, lacticacidand ethanolare 3.2, 4.6, 9.0,and 4.6, respectively. *Measured1.5 h later. wMeasuredafterethanolwasgivenat local hospital. Osmolalityfromethanolis subtracted.

587

Journal of Analytical Toxicology, Vol. 29, September 2005

soning. Over more than 20 years, we have experienced only one methanol patient in whom serum formate remained within its reference interval. This patient had no symptoms, was admitted early, and treated promptly with ethanol for days (8), which emphasizes the sensitivity of the present method in diagnosing methanol poisoning. There was no correlation between serum methanol and serum formate. Most authors describe the same ]ack of correlation (7,9,10). Theoretically, one might expect an inverse correlation between serum methanol and serum formate since the increase in formate concentration results from methanol metabolism. There are different reasons for not observing this inverse correlation. There are large individual differences regarding the elimination of both methanol and formate. The inhibition of the metabolism of methanol by an antidote would also alter the correlation between methanol and formate (10), as would the increase of formate metabolism by treatment with folinic acid (3). In our experience, a markedly increased serum formate as seen here is still both a sensitive and specific indicator of methanol poisoning. Increased formate may be seen following accidental exposure to formaldehyde, but such an exposure would most probably be known to the patient and would therefore seldom pose a diagnostic problem to the clinician. The formate assay used here is an established one; it is easily adaptable to most analyzers and requires only commercially available reagents. The reagents once dissolved may be stored for months at -70~ Being a referral hospital for acute intoxications, we need both a methanol and a formate assay. We now use the formate assay on a 24 h basis to screen for poisoning of an unknown agent, allowing us to use our GC methanol method only during daytime for analysis of sera with increased formate, or when formate is normal but there is a high suspicion of methanol poisoning. Such a combined use of the two assays is scientifically sound and efficient, allowing us to run the more labor-intensive and time-consuming GC method in a more limited and specific way. The formate assay may also well be used by hospitals lacking the GC method because it will allow them to reliably detect the methanol-poisoned patients.

588

Acknowledgment This study was supported by a grant from the Norwegian Directorate for Health and Social Affairs, Department for Emergency Medicine and Preparedness.

References 1. D. Jacobsen, J.E. Bredesen, I. Eide, and J. Ostborg. Anion and osmolal gaps in the diagnosis of methanol and ethylene glycol poisoning. Acta Med. Scand. 212:17-20 (1982). 2. K.E. Hovda, O.H. Hunderi, N. Rudberg, S. Froyshov, and D. Jacobsen. Anion and osmolal gaps in the diagnosis of methanol poisoning: clinical study in 28 patients. Intensive Care Med. 30: 1842-1846 (2004). 3. D.G. Barceloux, G.R. Bond, E.P. Krenzelok, H. Cooper, and J.A. Vale. American Academy of Clinical Toxicology practice guidelines on the treatment of methanol poisoning. J. Toxicol. Clin. Toxicol. 40:415-446 (2002). 4. L. Aabakken, K.S. Johansen, E.B. Rydningen, J.E. 8redesen, S. Ovrebo, and D. Jacobsen. Osmolal and anion gaps in patients admitted to an emergency medical department. Hum. Exp. Toxicol. 13:131-134 (1994). 5. P. Urdal. Enzymic assayfor oxalate in unprocessedurine, as adapted for a centrifugal analyzer. Clin. Chem. 30:911-913 (1984), 6. K.H. Schaller and G.T. Triebig. Formate determination with formate dehydrogenase, In Methods of Enzymatic Analysis, H.U. Bergmeyer, Ed. Verlag Chemie, Weinheim, Germany, 1984, pp 668-672. 7. W. Kerns, C. Tomaszewski, K. McMartin, M. Ford, and J. Brent. Formae kinetics in methanol poisoning. J. Toxicol. Clin. Toxicol. 40:137-143 (2002). 8. D. Jacobsen, S. Ovrebo, E. Amesen, and P.N. Paus. Pulmonary excretion of methanol in man. Scand. J. Lab. Invest. 43:377-379 (1983). 9. J. Brent, K. McMartin, S. Phillips, C. Aaron, and K. Kulig. Fomepizole for the treatment of methanol poisoning. N. Engl. J. Med. 344: 424-429 (2001). 10. O.M. Sejersted, D. Jacobsen, S. Ovrebo, and H. Jansen. Formate concentrations in plasma from patients poisoned with methanol. Acta Med. Scand. 213:105-110 (1983). Manuscript received September 29, 2004; revision received December 17, 2004.