Toxic Alcohol and Metabolite Concentrations
Serum methanol, formate, ethylene glycol, oxalate, and isopropanol concentrations (as appropriate) would be the ideal tests to perform when toxic alcohol poisoning is suspected shortly after exposure. However, these concentrations are most commonly measured by gas chromatography with or without mass spectrometry confirmation, methodologies that are not available in most hospital laboratories on a 24 hour basis, if at all. In fact, in many hospitals these are only available as “send out” tests, so results arrive too late for early clinical decision making.133 Enzymatic assays for methanol, formic acid, ethylene glycol, and glycolic acid have been developed,27,232,249 and these may lead to more readily available clinical tests. However, the commercial product is currently approved for veterinary use only. This veterinary test is effective for confirming the qualitative presence of ethylene glycol in human poisoning, although false positives may occur with propylene glycol.157 In a murine model, a commercially available ethanol in saliva point of care test can detect the presence of a low concentration of methanol but not ethylene glycol.96 Unfortunately, it would not distinguish between methanol and ethanol, limiting the clinical utility of this test. A group in Finland described a point of care breath test for methanol, using a portable Fourier transform infrared (FT-IR) analyzer similar to the “breathalyzers” used by law enforcement agents.147 Although analyzers like this are used to check for methanol as a combustion product in industry, they are not yet approved for medical use in the United States. Once approved, they would be useful for early clinical decision making because they are easy to use and provide a rapid result. They also can provide continuous monitoring of concentrations, a feature that would be very helpful during hemodialysis. Unfortunately, this methodology could not be used to detect ethylene glycol because of its low volatility.
Patients presenting late after ingestion may already have metabolized all parent compound to toxic metabolites and thus may have low or no measurable toxic alcohol concentrations. Fortuitously, the enzymatic assay for ethylene glycol is also capable of detecting glycolic acid, although as mentioned, this assay is approved only for veterinary use. Some authors have actually advocated for routine testing for glycolic acid in addition to testing for the parent compound when ethylene glycol poisoning is suspected.198 Serum and urine oxalate concentrations may also be determined,233 although their clinical utility is unclear. Similarly, a formate concentration may be valuable when a patient presents late after methanol ingestion.115,184 Formate was detected in blood samples from 97% of patients who died of methanol poisoning in one series; all of these patients also had detectable blood or vitreous methanol concentrations.126 Clearly, a low or undetectable toxic alcohol concentration must be interpreted within the context of the history and other clinical data, such as the presence of acidosis and end-organ toxicity, with glycolate and formate concentrations as potentially valuable additions.
Samples must be handled correctly for accurate toxic alcohol results. Particularly with the more volatile alcohols methanol and isopropanol, concentrations may be falsely low if the sample tubes are not airtight. This commonly results in low concentrations if alcohol concentrations are done as “add on” tests to samples already opened for electrolyte or osmol determinations.
Other alcohols such as benzyl alcohol and propylene glycol are not routinely assessed for by gas chromatography. Thus these xenobiotics present a much greater diagnostic challenge than methanol and ethylene glycol. Enzymatic assays for methanol or ethylene glycol would also fail to detect these, although false positive ethylene glycol tests may occur if propylene glycol is present. Thus a high index of suspicion is critical to establishing the diagnosis in these cases. If suspected on the basis of history, specific toxic alcohol testing should be performed.
Once alcohol concentrations are obtained, their interpretation represents a further point of controversy. Traditionally, a methanol or ethylene glycol concentration greater than 25 mg/dL has been considered toxic, but the evidence supporting this as a threshold is often questioned. In a case series of methanol poisoned patients from the 1950s, a methanol concentration of 52 mg/dL was the lowest associated with vision loss.23 This may have been the origin of the 25 mg/dL threshold, incorporating a 50% reduction as a margin of safety. However, the patient with the 52 mg/dL concentration presented 24 hours after his initial ingestion, and therefore was much more severely poisoned than suggested by his serum concentration at that point. In fact, almost all reported cases of methanol poisoning involve patients with delayed presentations who already have a metabolic acidosis.141 The only reported patient who went untreated after presenting early with an elevated methanol concentration (45.6 mg/dL) never developed acidosis or end-organ toxicity.32,141 A systematic review found that 126 mg/dL was the lowest methanol concentration resulting in an acidosis in a patient who arrived early after ingestion and met the inclusion criteria. The authors concluded that the available data are currently insufficient to apply a 25 mg/dL treatment threshold in a patient presenting early after ingestion without acidosis.141 However, until better data are available demonstrating the safe application of a higher concentration, it seems prudent to use a conservative concentration such as 25 mg/dL as a threshold for treatment.
Because of the problems with obtaining and interpreting actual serum concentrations, many surrogate markers have been used to assess the patient with suspected toxic alcohol poisoning. The initial laboratory evaluation should include serum electrolytes, including calcium, blood urea nitrogen, serum creatinine concentrations, urinalysis, measured serum osmolality, and a serum ethanol concentration. Blood gas analysis with a lactate concentration is also helpful in the initial evaluation of ill appearing patients.
For a full discussion of the anion gap concept, refer to Chap. 19. As previously discussed, anion gap elevation is a hallmark of toxic alcohol poisoning. In fact, the possibility of methanol or ethylene glycol poisoning is often first considered when patients present with an anion gap acidosis of unknown etiology, frequently with no history of ingestion. Unless clinical information suggests otherwise, it is important to exclude metabolic acidosis with elevated lactate concentration and ketoacidosis, which are the most common causes of anion gap acidosis, before pursuing toxic alcohols in these patients. This is because of the extensive evaluation required and expensive, potentially invasive course of therapy to which they are otherwise committed. However, elevated lactate concentrations may be present in the setting of both methanol and ethylene glycol poisoning.163,170,219
The unmeasured anions in toxic alcohol poisoning are the dissociated organic acid metabolites discussed above. The acidosis takes time to develop, sometimes up to 16 to 24 hours for methanol. Thus the absence of an anion gap elevation early after reported toxic alcohol ingestion does not exclude the diagnosis. If ethanol is present in the body, the development of acidosis will not begin to occur until enough ethanol has been metabolized that it can no longer effectively inhibit ADH (see Ethanol Concentration, below).
A potential early surrogate marker of toxic alcohol poisoning is an elevated osmol gap (the principles and the calculations are discussed in detail in Chap. 19). However, it is important to recognize that osmol gap elevation is neither sensitive nor specific for toxic alcohol poisoning. Since a baseline osmol gap is generally not available when evaluating a patient (with rare exceptions),113 and a normal osmol gap ranges from –14 to +10 osmols, so-called “normal” osmol gaps cannot exclude toxic alcohol poisoning.108 For example, in a patient with a baseline osmol gap of –10, a current gap of +5 potentially represents a methanol concentration of 47 mg/dL or an ethylene glycol concentration of 93 mg/dL, values that might require hemodialysis. Inversely, a moderately elevated osmol gap (+10 to +20) is not necessarily diagnostic of toxic alcohol poisoning because other disorders such as alcoholic ketoacidosis and metabolic acidosis with elevated lactate concentration, may raise the osmol gap.214 Furthermore, mean osmol gaps vary within populations over time, further limiting their utility.142 However, a markedly elevated osmol gap (>50) is difficult to explain by anything other than a toxic alcohol.
Further complicating matters, the anion gap and osmol gap have a reciprocal relationship over time. This is because soon after ingestion, the alcohols present in the serum raise the osmol gap but do not affect the anion gap because metabolism to the organic acid anion has not yet occurred. As the alcohols are metabolized to organic acid anions, the anion gap rises while the osmol gap falls, because the metabolites are negatively charged particles that have already been accounted for in the calculated osmolarity by doubling of the sodium. Thus patients who present early after ingestion may have a high osmol gap and normal anion gap, while those who present later may have the reverse.111,117 Figure 109–4 depicts a more intuitive visual representation of this process.
The reciprocal relationship of anion gap and osmol gap over time (hours). Note that patients presenting early may have a normal anion gap while patients who present late may have a normal osmol gap.
One retrospective and one prospective study have attempted to evaluate the performance characteristic of the osmol gap as a diagnostic test. Although in both cases, the osmol gap performed fairly well, the studies were small, 20 patients with toxic alcohol poisoning in the retrospective study and 28 patients with methanol poisoning in the prospective study, and the prospective study identified three patients with significant poisoning and acidosis but “normal” osmol gaps, defined in the study as less than 25.111,162 Therefore, these data do not eliminate the concern that a patient with significant poisoning could be missed by relying on the osmol gap alone to exclude poisoning.
A serum ethanol concentration is an important part of the assessment of the patient with suspected toxic alcohol poisoning. As discussed in Chap. 19, the ethanol concentration is necessary to determine the calculated osmolarity. In addition, because ethanol is the preferred substrate of ADH (4:1 over methanol and 8:1 over ethylene glycol), a significant concentration would be protective if coingested with a toxic alcohol. In fact, ethanol concentrations near 100 mg/dL virtually preclude toxic alcohols as the cause of an unknown anion gap metabolic acidosis because the presence of such a concentration should have prevented metabolism to the organic acid. A possible exception would be ingestion of ethanol several hours after ingestion of a toxic alcohol.106 If a breath alcohol analyzer is used to determine ethanol concentration, a false positive ethanol value may be obtained if significant methanol concentrations are present, and the machine may not indicate that an interfering substance is present (as it does with acetone).40 Therefore, even if a prehospital breath alcohol analyzer indicates a significant ethanol concentration, this should be confirmed by determining the serum ethanol concentration.
Both methanol and ethylene glycol poisoning can result in elevated lactate concentrations, for different reasons. Formate, as an inhibitor of oxidative phosphorylation, can lead to anaerobic metabolism and resultant lactate elevation. Additionally, metabolism of all alcohols results in an increased NADH/NAD+ ratio, which favors the conversion of pyruvate to lactate. Furthermore, hypotension and organ failure in severely poisoned patients can also produce an elevated lactate concentrations. However, lactate production by these mechanisms tends to result in serum concentrations no greater than 5 mmol/L.
In ethylene glycol poisoning, the glycolate metabolite may also cause a false positive lactate elevation when measured by some analyzers, particularly with whole blood arterial blood gas analyzers. The Radiometer ABL series (625, 700, 725, 825, 835) is most widely reported to result in a false positive lactate; other specific models implicated to varying degrees include: Beckman LX 20, Bayer/Chiron Rapidlab series (860, 865), Roche Modular, Architect c8000, Vitros Fusion 5.1, Cobas Integra, GEM Premier 4000 and Hitachi 911 analyzers, but not the Vitros 950 or Vitros 250 or the Beckman Coulter DxC-800 chemistry analyzer.35,45,53,71,163,170,175,191,197,213,256 In such cases, the degree of lactate elevation directly correlates with the concentration of glycolate present,163,170 and the artifact results from the lack of specificity of the lactate oxidase enzyme used in these machines,170,175,197,256 although direct oxidation of glycolate at the analyzer anode is also suggested as a possible mechanism.222 Thus the presence of a “lactate gap” might also be used to diagnose ethylene glycol poisoning in hospitals where lactate assays are available with and without sensitivity to glycolate, or two lactate assays with different sensitivities to lactate.222,246 Ingestion of propylene glycol can also result in elevated lactate concentrations, but in this case, it is not a false positive lactate but rather an accurate measurement of a metabolite of propylene glycol.127,128
Serum glucose concentration is generally obtained as part of routine laboratory analysis. Hyperglycemia, defined as serum glucose greater than 140 mg/dL (7.77 mmol/L) in nondiabetic patients, portended a greater risk of death after methanol poisoning, with an odds ratio of 6.5 in one retrospective study.210 This has not yet been prospectively validated.
The urine may provide information in the assessment of the patient with suspected ethylene glycol poisoning. Calcium oxalate monohydrate (spindle-shaped) and dihydrate (envelope-shaped) crystals may be seen when the urine sediment is examined by microscopy, although this finding is neither sensitive nor specific.79,118,174 In fact, calcium oxalate crystals were present in the urine of only 63% (12 of 19) of patients with proven ethylene glycol ingestion in one series.33
Some brands of antifreeze contain fluorescein to facilitate the detection of radiator leaks. If one of these products is ingested and the urine is examined with a Woods lamp within the first 6 hours, there may be urinary fluorescence.255 Gastric aspirate may also demonstrate fluorescence.56 False positive fluorescence may result from examining the urine in glass or plastic containers due to the inherent fluorescence of these materials, so if this test is performed, an aliquot of the urine should be poured onto a piece of white gauze or paper. Recent work has suggested a lack of utility of this test. Almost all children had urinary fluorescence, and there was poor interrater agreement in determining fluorescence of specimens.43,189
The evaluation of patients with known or suspected ethylene glycol poisoning should also include serum calcium and creatinine concentrations. Patients with methanol poisoning and abdominal pain also warrant an assessment of liver function tests and serum lipase because of the possibility of associated hepatitis and pancreatitis.
Although characteristic brain CT and MRI abnormalities are frequently reported in the setting of methanol poisoning, it is unclear what role they have in the routine evaluation of these patients. The presence of putaminal hemorrhage or insular subcortex white matter necrosis was associated with a greater odds ratio of death (8 and 10, respectively) in one study of patients with methanol poisoning.237 However, in the absence of neurological abnormalities on physical examination, routine CT or MRI are probably not indicated.
Diagnostic Testing and Risk Assessment
Increases in both anion gap and osmolar gap may be useful for risk stratification in methanol poisoning, and a venous or arterial blood gas should be performed. A review of reported toxic alcohol cases attempted to identify risk factors for mortality in adults with methanol or ethylene glycol poisoning. For methanol poisoning, no patient with an anion gap less than 30 mEq/L or an osmolar gap less than 49 osmols died. A pH less than 7.22 was an even better predictor of mortality, as no patient with a pH greater than 7.22 died. For ethylene glycol, the tests were less useful. One patient with an osmolar gap of only 25 osmols died, no patient with an anion gap less than 20 mEq/L died, and pH did not predict mortality with statistical significance.54 This study has been criticized for missing a substantial number of patients,207 and it still needs to be validated in another population. Another retrospective study of risk factors for poor outcomes in methanol poisoning only found that pH less than 7.00 (as well as coma or a >24-hour delay to presentation) was associated with death.103 In a large series from several epidemics of methanol poisoning, a pH less than 7.00 and coma were again identified as risk factors associated with death. In patients with a pH less than 7.00, PCO2 greater than or equal to 23.3 mm Hg (3.1 kPa) was also a risk factor.185 In methanol poisoned patients unlikely to die, the pH may still be useful for predicting retinal toxicity. Another retrospective study examined markers for poor visual outcome after methanol poisoning and again found pH to be the best predictor, with a pH greater than 7.20 associated with a high likelihood of only transient visual sequelae.61