When hyperglycemic patients present to the emergency department, it is common practice to evaluate them for diabetic ketoacidosis (DKA) using both venous blood gas (VBG) and serum chemistry results. Within the past several years, however, many labs have implemented blood gas analyzers that report not only standard VBG measurements but also electrolyte results typically provided through serum chemistry panels. This prompted researchers at the University of Southern California (USC) to evaluate the performance of VBG electrolytes in comparison to standard serum chemistry results. Their findings are the subject of this issue of Strategies.

Amidst the rising tide of diabetes and diabetic complications, emergency physicians typically assess hyperglycemic patients for DKA when their triage blood glucose is ≥250 mg/dL, regardless of the purpose for their visit. American Diabetes Association (ADA) criteria for DKA include serum glucose ≥250 mg/dL, serum anion gap >10 mEq/L, bicarbonate ≤18 mEq/L, serum pH ≤7.30, and presence of ketosis. In the past, measuring these parameters required both venous blood gas and serum chemistry results. Today, many labs have implemented blood gas analyzers that report not only standard VBG measurements such as serum pH and bicarbonate, but also analytes like sodium, potassium, and glucose. Having observed that these parameters were available via VBG, USC researchers sought to evaluate their performance in comparison to standard serum chemistry results with the idea that if the methods were comparable they might be able to use only the VBG results (Acad Emerg Med 2011;18:1105-8).

“The emergency environment is very challenging due to crowding, so we’re constantly looking for processes that improve our throughput. One doesn’t think of traditional serum chemistry panels as taking a long time, but they can take a full hour longer than venous blood gas analysis. In institutions like mine that are chronically severely crowded, those hours add up,” explained Michael Menchine, MD, MPH. “On any given day we easily see dozens of patients presenting with hyperglycemia, among other complaints, so saving time for tests results across all those patients results in a pretty substantial number of wait hours that we could theoretically decrease.” He is an associate professor of emergency medicine at the USC Keck School of Medicine in Los Angeles.

Menchine and his co-authors conducted a prospective study of 342 patients who presented to the Los Angeles County-USC Hospital emergency department with a triage capillary blood glucose ≥250 mg/dL and who had VBG and serum chemistry samples collected within 30 minutes of each other. In all, 46, or 13.5% met ADA criteria for DKA. The investigators found that VBG had a sensitivity of 97.8% and specificity of 100% for diagnosing DKA. VBG results missed one DKA case because VBG bicarbonate results for this patient were just above the ADA threshold criterion for that one analyte. The researchers reported correlation coefficients between VBG and serum chemistry to be 0.90, 0.73, 0.94, and 0.81 for sodium, chloride, bicarbonate, and anion gap, respectively.

As a result of these findings, procedures at LAC-USC Hospital have changed, according to Menchine. “It’s not a firm policy to only order venous blood gases when the consideration is diabetic ketoacidosis. However, practically speaking, when there is concern that a patient might have diabetic ketoacidosis, more often than not we’re now running venous blood gases to answer that question,” he explained. “As long as we don’t need serum chemistry panels to evaluate other complaints the patient is having, we often don’t order any other lab tests on them.”

Menchine and his colleagues are conducting further research to see whether using only VBG electrolytes actually results in any time savings in treating suspected DKA patients. “It’s possible that there are so many other delays and considerations that a one-hour savings might not matter much. However, we’re hopeful that it will,” he said. “Ultimately, a one-hour delay probably doesn’t result in clinically important side-effects, but many hours spent waiting for chemistry results could have a cumulative effect. Patients may have to wait in the waiting room longer, get frustrated and leave, and not receive care at all. That’s where we think the impact might be.”

David Koch, PhD, agreed that VBG electrolyte results could supplant serum chemistry electrolytes measurements in this patient population, and in fact, his lab at Grady Memorial Hospital in Atlanta is following a similar protocol. “We haven’t measured whether it has cut throughput times, so I don’t know whether we have seen decreases. That may be in part, as Dr. Menchine mentioned, that there may be additional tests required that can only be obtained from the serum chemistry panel. If that’s the case, then you’re going to be stuck with duplication between the venous blood gas and serum chemistry panel tests.” Koch is director of clinical chemistry, toxicology, and point-of-care testing (POCT) at Grady and associate professor of pathology and laboratory medicine at Emory University.

Koch, who was not involved in the study, stressed that analyses of this nature are important as healthcare systems aim to boost efficiency. “Every hospital should be doing what’s clinically justified and not going beyond that. The need to try and corral this over-utilization of the lab is important,” he noted. Koch argued, however, that this particular method comparison would have benefitted from further analysis. “When you’re comparing one instrument with another for measuring the same parameter, you should report the regression statistics, and a calculation of the systematic error from that. You would then compare the actual observed error with the allowable error,” he explained.

Koch also noted that the authors assessed the presence of ketosis using urine dipstick, and suggested that a better method would be measurement of serum beta-hydroxybutyrate, which the National Academy of Clinical Biochemistry strongly endorsed instead of ketone measurements in its guidelines on laboratory analysis in the diagnosis and management of diabetes. “It’s better clinically to report beta-hydroxybutyrate than ketones, which amount to acetone and acetoacetate. The reason is, in ketoacidosis, beta-hydroxybutyrate rises first and acetoacetate and acetone actually drop a little. So you could provide misleading information from serum or urine ketone measurements using acid tablets, sticks, or some other semi-quantitative method,” he said.

Koch noted, however, that some labs continue to use the outdated practice, possibly due to their choice of analyzers. “None of the major vendors offer beta-hydroxybutyrate testing on their large chemistry analyzers, so a lab making a choice to buy one of these instruments is faced with having to find an alternative to measure this analyte,” he explained. “Several beta-hydroxybutyrate kits are available, but you have to have an analyzer that’s at least a partially open system, allowing you to place the kit on your analyzer.”

Menchine agreed that beta-hydroxybutyrate is a more robust method, but explained that at the time of the study, LAC+USC Hospital was still using urine dipstick to measure ketones. He and his co-authors recently published two studies, one reporting on POCT beta-hydroxybutyrate for assessing the severity of diabetic ketoacidosis, and the other comparing POCT beta-hydroxybutyrate with urine dipstick analysis.

Both Menchine and Koch agreed that this study should encourage labs to meet with physicians and discuss any redundancies that might exist in their testing practices and how labs might better meet clinicians’ needs. As Koch observed, “it’s always a good idea to have a conversation about what’s needed clinically and what labs can offer.”