Osmol gap method for the detection of diethylene glycol in human serum

Abstract

BACKGROUND:

Measurement of the osmol gap (OG) is a technique that is used frequently in toxic alcohol poisonings (ethylene glycol (EG) and methanol) as a rapid means to estimate exposure, and can be performed in virtually all hospital laboratories. The value of the OG has not been previously evaluated for diethylene glycol (DEG) exposures. The primary objective of this study was to evaluate the utility of the OG in estimating DEG serum concentrations using the most common formula that is currently used for estimating methanol, ethanol, and ethylene glycol concentrations.

METHODS:

This was a controlled laboratory investigation using serum samples individually spiked with a known quantity of toxic alcohol compared to no toxic alcohol. Test samples were spiked with ethanol, DEG, EG, and methanol. Serum chemistries and osmolality and osmolarity were determined, and the OG was determined for each specimen.

RESULTS:

The percent error of estimating DEG concentrations of 26.3% was similar to the mean percent error for estimating other alcohol concentrations, 30.5%±5.6% (P>0.05, 95% confidence interval 16.7%-44.3%).

CONCLUSIONS:

The severity of metabolic effects associated with DEG and the need to appropriately determine rescue treatments mandate early detection of significant exposures for effective triage and patient management. Our results indicate that the percent error of the osmol gap method for estimating DEG concentration is similar to that of other toxic alcohols; this simple technique could be a valuable clinical tool, since quantitative DEG analysis is rarely available.

INTRODUCTION

Diethylene glycol (DEG) is a commonly used solvent found in many commercial products including fog solution,[1] brake fluid,[2] canned fuel,[3] and wallpaper stripper.[4] DEG exposure has resulted in serious epidemic poisonings and deaths.[5-8] The two most recent tragedies were the poisoning deaths of over 100 children in Nigeria due to a teething solution that had been contaminated with DEG in the period of 2008-2009[9] and the tragedy in Panama which was responsible for the deaths and disabilities of hundreds of adult patients due to contaminated cough syrup.[10] Because of the severity of metabolic effects and the need to rapidly institute appropriate treatments, early detection of significant exposures is imperative for effective triage and patient management.

DEG consists of two ethylene glycol molecules joined by an ether bond, with a molecular weight of 106 daltons. Currently, there is no readily available clinical laboratory test to measure DEG blood levels. Although DEG can be measured via gas chromatography/ mass spectrometry (GC/MS), or gas chromatography/ flame ionization detector (GC/FID), the test is almost uniformly available only as a send-out to reference laboratories, with results available days later. The virtual non-existent capabilities for DEG detection at the majority of healthcare facilities results in a delay in quantitation and is not helpful for providing real-time patient care. Since DEG is an osmotically active compound, if it is present in sufficient quantities an elevated osmol gap (OG; the difference between the measured osmolality and calculated serum osmolarity) should be detectable.

Measurement of the OG is a technique that is used frequently in ethylene glycol (EG) and methanol poisonings as a rapid means to estimate exposure, and can be performed in virtually all hospital laboratories. While the osmol gap is not a definitive diagnostic test by any means, particularly when toxic alcohol concentrations are low, it has been shown to be a useful adjunct to estimate toxic alcohol concentrations in select clinical settings.[11] To estimate toxic alcohol serum concentrations, the OG value is multiplied by one tenth of the MW of the compound in question, thereby providing an estimate of the serum toxic alcohol concentration in appropriate units (mg/dl). The primary objective of this pilot study was to evaluate whether DEG causes a detectable OG similar to ethanol, methanol, and ethylene glycol. If this study shows that DEG does cause an OG, the next phase will be used to determine whether the OG method for estimating DEG has clinical utility in estimating DEG levels, similar to that for the other toxic alcohols.

METHODS

This study was a laboratory investigation using five tubes of blood provided by the primary investigator. The blood was collected in five tiger top Vacutainer® serum separator tubes (Becton Dickson, #366513) and four tubes were immediately spiked with known quantities of EG, DEG, ethanol, and methanol, respectively, using a microliter pipette. The specific gravity (SG) of each of the alcohols was used to determine the aliquot of alcohol delivered in each microliter using the formula: weight in mg × SG = mg of alcohol per microliter of solution in an attempt to achieve a target serum alcohol concentration above 200 mg/dl (calculations in table 1). One tube was unadulterated and served as a control to ensure that this tube-spiking method did not over-dilute or adversely affect serum chemistry values. After gentle inversion mixing of the tubes, each tube was then centrifuged and serum separated as per standard laboratory protocol. All alcohols were reagent grade, >99% alcohols (Aldrich Chemical Co, Inc., Milwaukee, WI).

Table 1.

Calculation of toxic alcohols added to serum separator tubes, attempting to achieve a target concentration of 200 mg/dl of each toxic alcohol, or 10 mg/ 5 ml

Laboratory analysis was performed in a tertiary-care hospital using standard clinical laboratory instruments and protocols for quality control (QC). Serum osmolality was determined for all samples by freezing point depression using a single sample micro-osmometer (Model 3300, Advanced Instruments, Inc.,Norwood, MA). Ethanol and methanol concentrations were determined using a packed column GC/FID, and DEG and EG concentrations were measured using capillary GC/FID (Model Autosystem, Perkin Elmer, Waltham, MA). The limit of quantitation and coefficient of variation (CV) for each method were 5 mg/dl (CV, 3%-8%) and 15 mg/dl (CV, 5%-7%), respectively. Sodium, blood urea nitrogen (BUN), and glucose were measured on all samples (Model: LX20, Beckman Coulter, Fullerton, CA). Osmolarity was calculated using the following formula:[12]

Calculated osmolarity=(2×Na)+(BUN/2.8)+ (glucose/18)

The OG was obtained by subtracting the calculated serum osmolarity from the measured serum osmolality. It was then multiplied by 1/10 of the MW of each alcohol (10.6, 6.2, 4.6, and 3.2 for DEG, EG, ethanol, and methanol, respectively) to achieve the estimated serum concentration of the toxic alcohol.

Measurable DEG concentrations were detected simultaneously with the ability to detect the presence of other alcohols using the OG method. Descriptive statistical analyses were performed and the percent error for this determination of DEG was compared with the average of the percent error for EG, methanol, and ethanol.

RESULTS

Table 2 summarizes collected clinical data. The results indicate that the percent error of estimating DEG concentrations is 26.3% similar to the mean percent error for estimating other alcohol concentrations, 30.5% ± 5.6%. There was no statistical difference, and 95% confidence interval was 16.7%-44.3%.

Table 2.

Collected clinical data

DISCUSSION

Excluding ethanol, most hospital emergency departments do not have access to laboratories that can readily and rapidly test toxic alcohol concentrations. Because of this, toxicologists and emergency physicians estimate toxic alcohol concentrations by calculating the OG, and it has been found to be useful when it is significantly elevated. Diethylene glycol (DEG) is an osmotically active toxic di-alcohol (glycol). We postulated that if we could demonstrate that the OG method was a reasonable method for detecting significant DEG exposure, similar to its utility for the other toxic alcohols, clinicians would have improved diagnostic capabilities using this readily available test as a screening tool when considering DEG exposure.

This pilot study demonstrates the utility of using an elevated OG to detect DEG, and that it appears to be as useful as it is for EG, ethanol, and methanol (ie, in this trial, it has a similar percent error). Rapid assessment of potentially consequential exposures to DEG is imperative for improving patient outcome. The ability to use the OG in this capacity has some clinical value, although it must be interpreted in light of the patient’s history, including the presence of other osmotically active agents such as ethanol. While the potential for severe morbidity and mortality due to DEG poisoning is high, the minimum threshold for toxicity following an exposure to DEG is currently unknown. When toxic alcohols are present in high concentrations, significantly elevated OGs (e.g. when>50) are almost invariably confirmatory for toxic alcohol exposure, and proper antidotal therapy and/or dialysis should be initiated.[13] Recently, Lynd et al[11] have shown the value of the OG in estimating EG and methanol concentrations, and the predictive value for needing dialysis, suggesting that the OG provides at least some discriminatory diagnostic information. Conversely, as has been well-described in the toxicology literature, the OG method has limited utility when it is in the normal range, since toxic alcohols could still be present in low, but clinically significant quantities.[14-17] Therefore, it should be emphasized that this study does not intend to imply that a normal OG rules out a toxic alcohol exposure.

The severity of DEG poisoning was most recently demonstrated in February 2009, when 84 Nigerian children were reported dead after being given “My Pikin”, a teething syrup contaminated with diethylene glycol. More widespread DEG poisoning occurred in Panama in the fall of 2006,[18,19] when the National Health Service of Panama prepared liquid sugar-free cough medicines with a glycerin diluent that had been unknowingly contaminated with DEG (23%-24%). Approximately 30 000 Panamanians were potentially exposed to the contaminated product and at least 118 patients were reported to have clinical toxicity, including permanent renal and neurologic sequelae or death.[20]

Another DEG-contaminated pharmaceutical diluent led to poisoning of an intravenous silimarin preparation in China in 2006 as well, possibly from the same supply chain problems. There have been numerous past poisonings when the same tragic mistake had been made, and DEG was an unrecognized ingredient/contaminant in liquid pharmaceutical preparations.[21] Perhaps most notable was the Massengill Elixir Sulfanilamide disaster of 1937, which resulted in the fatal poisoning of 105 patients and the subsequent passage of the U.S. Federal Food Drug and Cosmetic Act.[7] DEG continues to be used as a solvent in a variety of commercially available consumer products, and the potential for both intentional and unintentional exposures exists.

The limitations of this study include the large variability in alcohol concentrations achieved via manual administration and the lack of sample replication. Administration of the proper volume of alcohol was difficult when aliquots were extremely small, particularly of the higher molecular weight alcohols (EG, DEG). Since this pilot study shows that DEG does indeed cause a detectable osmol gap in human serum, and the feasibility of using this technique to study the gap, in a future study we will use higher volume pooled serum specimens, which will allow larger volumes (and hence more accurately-measured quantities) of toxic alcohols to be added, and hence will likely yield concentrations closer to the target values (the current plan is to use a range of concentrations to demonstrate its clinical utility).

Despite these limitations, our findings were consistent with what is known regarding other alcohols and the observation that an elevated OG may serve as both a marker of exposure, and perhaps as a crude estimation of DEG serum concentrations when the alcohol is present in significant quantities. This small study found that the percent error of the OG method for estimating DEG concentration was similar to that for other toxic alcohols in our current investigation. As the majority of institutions are not able to perform DEG analysis via GC/FID methodology, the OG is a clinical tool that appears to be as useful for assessing potentially toxic exposures of DEG as it is for the other toxic alcohols.

This pilot study shows that the OG method would be feasibly used as a screening tool for the detection of DEG exposure, similar to its current use for other toxic alcohols. With all toxic alcohols, the utility of this screening tool diminishes when alcohol concentrations are low. The planned future study with larger serum samples and varying levels of toxic alcohol will be able to determine whether the OG method is valid for lower DEG serum concentrations.