Unintended Consequence of High-Dose Vitamin C Therapy for an Oncology Patient: Evaluation of Ascorbic Acid Interference With Three Hospital-Use Glucose Meters

Abstract

The use of high-dose vitamin C in cancer care has offered promising results for some patients. However, the intravenous (IV) doses used for these patients can reach concentrations that interfere with some strip-based glucose meters. We characterized the impact of vitamin C interference, from standard to the very high doses used for some cancer protocols, using three different hospital-use glucose meters. For two of the three devices tested, increasing concentrations of ascorbic acid caused false elevations in the glucose measurements. The third glucose meter did not provide inaccurate results, regardless of the vitamin C concentration present. Rather, above a certain threshold, the device generated error messages and no results could be obtained.

Introduction

Positron emission tomography (PET)/computed tomography is an imaging tool used broadly for the diagnosis, staging, and assessment of therapeutic response in oncology. The technique utilizes a radioactive glucose analogue tracer, ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG). Shortly after clinical use of PET became established in the 1990s, reports of hyperglycemia and the role endogenous glucose played in reducing tumor uptake of ¹⁸F-FDG due to competition began to surface in the literature. As a result, various professional societies in Europe and the United States proposed recommendations on patient preparations (eg, fasting requirements) and the need to measure glucose concentrations prior to PET.¹

To facilitate rapid assessment of glucose concentration prior to PET and other procedures, point-of-care glucose meters are often used. However, these devices are subject to preanalytic and analytic interferences not observed with laboratory glucose measurement.^2-5 In particular, inaccurate or even invalid glucose measurements can result in unnecessarily rescheduled procedures that may further delay diagnosis or treatment.

Ascorbic acid or vitamin C is an antioxidant that has been shown to produce both falsely decreased and elevated glucose meter results depending on the technology employed in the meter.⁴ All three strip-based glucose meters included in this study use an electrochemical detection method with either a maltose-insensitive variant of glucose dehydrogenase (Roche Accu-Chek Inform II, Roche Diagnostics, Indianapolis, IN, USA and Abbott Precision Xceed Pro, Abbott Laboratories, Abbott Park, IL, USA) or a glucose oxidase (Nova StatStrip Glucose Hospital Meter, Nova Biomedical, Waltham, MA, USA) enzyme system. In devices with electrochemical strips the false increase in glucose results is due to the oxidation of ascorbic acid at the electrode and subsequent current generation. The Nova StatStrip device has previously been shown to correct for ascorbic acid interference up to 10 mg/dL.² We recently observed a case of repeat invalid results in a patient who had received high doses of ascorbic acid for the treatment of recurrent lymphoma. The patient received intravenous ascorbic acid (1 g/kg body weight) on nine occasions over 19 days. Four hours following the 12th ascorbic acid infusion, the patient presented for PET scan and had invalid results obtained by glucose meter using 12 strips on four different Nova StatStrip devices using both capillary and venous (peripheral intravenous) blood. The manufacturer has a claim of no ascorbic acid interference up to 12.5 mg/dL. We designed a series of experiments to (1) evaluate the impact of ascorbic acid interference on the glucose meter technology used in our institution and (2) assess the impact of increasing concentrations of ascorbic acid on glucose meter measurements using all three Food and Drug Administration–approved hospital-use glucose meter technologies available in the United States.

Methods

The Mayo Clinic Institutional Review Board (IRB) determined that the study met institutional criteria for a quality assurance/improvement initiative and did not require IRB review.

Impact of Ascorbic Acid on Patient’s Glucose Meter Measurements

Using the patient’s residual plasma and waste lithium heparin plasma with a normal glucose concentration as “diluent,” a recovery experiment was performed to determine if the interfering substance could be sufficiently diluted. Four mixtures of patient:plasma “diluent” (75:25, 50:50, 25:75, and 10:90) were prepared and tested on the Nova StatStrip Glucose Hospital Meter System (Nova Biomedical). The plasma sample was then submitted for ascorbic acid measurement by liquid chromatography-tandem mass spectrometry.

Impact of Ascorbic Acid on Glucose Measurement Using Three Different Meters

Mayo Clinic Pharmacy provided a stock solution of 500 mg/mL Ascorbic Acid Injection, USP (Mylan, Canonsburg, PA, USA). A working solution of ascorbic acid (25 mg/mL) was prepared using 0.9% saline. Three pools of residual waste lithium heparinized whole blood were collected with low, normal, and high glucose concentrations measured using a centrifuged aliquot of plasma on the Roche c 501 (Roche Diagnostics). Aliquots from each pool were spiked with varying ascorbic acid concentrations (0, 5, 10, 25, 50, 100, and 200 mg/dL). Control samples were spiked with an equal volume of saline to serve as a reference and account for dilution effects. Diluent constituted <10% sample volume to maintain sample matrix integrity. Each spiked sample from the three glucose concentration pools (n = 21 total samples) were measured in duplicate on the following strip-based glucose meters: meter 1—Nova StatStrip Glucose Hospital Meter System (Nova Biomedical), meter 2—Roche Accu-Chek Inform II System (Roche Diagnostics), and meter 3—Abbott Precision Xceed Pro (Abbott Laboratories). Duplicate measurements were averaged and the glucose difference was calculated Mean_(spike) – Mean_(Control). The glucose difference vs ascorbic acid concentration was plotted for each glucose pool.

Results

The recovery experiment performed using residual plasma from the patient demonstrated that the ascorbic acid interference could not be overcome until the sample was diluted 10-fold (one part patient sample to 9 parts plasma “diluent”). This mixture resulted in a 106% recovery. All other mixtures resulted in error messages on the StatStrip glucose meter. An aliquot of the residual plasma sent for ascorbic acid measurement was found to contain 104 mg/dL (reference interval: 0.4-2.0 mg/dL).

Using lithium heparin plasma pools at low (30 mg/dL), normal (74 mg/dL), and high (249 mg/dL) glucose concentrations, the ascorbic acid interference was assessed on all three commercially available hospital-use glucose meters at ascorbic acid concentrations ranging from 0 to 200 mg/dL (Figure 1). Across all glucose concentrations tested, as the ascorbic acid concentration increased, meters 2 and 3 demonstrated a positive bias. The extent of interference was greatest with the low glucose concentration pool (>13-fold increase for meter 2 and >15-fold increase for meter 3). However, interfering concentrations on meter 1 resulted in instrument “errors” rather than inaccurate meter results. For both meters 2 and 3, at the highest concentration of vitamin C tested (200 mg/dL), the glucose result obtained was >analytical measuring range (AMR) of the device, >600 and >451 mg/dL, respectively.

Figure 1.

Graphs of mean glucose difference vs ascorbic acid concentration for the (a) low, (b) normal, and (c) high glucose concentration pools tested on three meters. Missing points for meter A reflect instrument “errors.” When results >AMR were obtained, 601 and 452 mg/dL were used for meter 2 and meter 3, respectively.

A change in glucose of at least 12.5 mg/dL (glucose <100 mg/dL) or 12.5% (glucose >100 mg/dL) was observed when the ascorbic acid reached the following concentrations for each meter: meter 1: N/A (instrument had “bad strip” errors at 50 mg/dL for all three glucose concentration pools); meter 2: 10 mg/dL (low glucose), 5 mg/dL (normal glucose), and 25 mg/dL (high glucose); and meter 3: 5 mg/dL (low and normal glucose) and 10 mg/dL (high glucose). Based on the patient’s measured vitamin C concentration (104 mg/dL) and the normal plasma glucose (89 mg/dL) obtained from the laboratory-based instrument, if meter 2 or meter 3 had been used for her glucose measurements, values of 393and 325 mg/dL, respectively, would have been reported—the highest value nearly 4.5 times the accurate result.

Discussion

The use of glucose meters in critically ill patients remains controversial. A number of variables including hematocrit, PO₂, PCO₂, pH, and medications may interfere with or impact the accuracy of glucose meters.^2-4,6 Many of the studies demonstrating limitations to glucose meters were performed with older glucose meter technologies that did not correct for hematocrit or other interferences. Newer meter technologies may provide more accurate glucose measurement^7-9 by reducing systematic bias and resolving medication and hematocrit interferences. However, not all glucose meters are equally effective in resolving or eliminating medication interferences.

In this study, we found that interference with some devices may result in falsely elevated results, while interfering concentrations on another device resulted in instrument “errors” rather than inaccurate meter results. With clinical studies and literature data highlighting the potential of high-dose vitamin C therapy in patients with cancer,¹⁰ current use for severe burn patients,^5,11 and more recently in the treatment of patients with coronavirus disease 2019,^12,13 its use may expand more broadly. Therefore, manufacturers and users of glucose meters should be aware of this potential interference and the challenge it presents with some devices.

Conclusions

We present a case of glucose meter interference caused by extremely high therapeutic doses of vitamin C. While the Nova StatStrip glucose meter effectively detected the presence of ascorbic acid interferant and suppressed glucose results, the Roche Inform II and Abbott Precision Xceed Pro demonstrated falsely increased results that could have impacted patient care (delayed PET scan) or possibly led to inappropriate patient treatment.