Elevated methemoglobin levels in patients treated with high-dose hydroxocobalamin

Abstract

Objective

The aim of this study was to assess the impact of hydroxocobalamin (OHCbl) infusion on arterial blood gas and oximetry values in patients with vasoplegic syndrome.

Methods

Blood samples collected from 95 patients receiving OHCbl infusion were assayed using the ABL90 FLEX Plus blood gas analyzer for the concentration of methemoglobin (MetHb), total hemoglobin (tHb), carboxyhemoglobin (COHb), arterial oxygen saturation (SaO₂), arterial oxygen partial pressure (PaO₂), and arterial carbon dioxide partial pressure (PaCO₂). Interference of OHCbl on these variables was evaluated using the measured difference between the preinfusion and postinfusion samples.

Results

Blood MetHb (%) measured after the infusion of OHCbl (5g) were significantly higher than the baseline levels, with a median of 4.8 (IQR, 3.0−6.5) versus 1.0 (IQR, 1.0−1.2) (P < .001). Blood COHb (%) increased from a median of 1.3 (IQR, 1.0−1.8) to 1.7 (IQR, 1.3−2.2) (P < .001) following the OHCbl infusion. No differences were seen in median levels of tHb, PaO₂, PaCO₂, and SaO₂ between pre- and post-OHCbl treatment.

Conclusion

The presence of OHCbl in blood clearly interfered with the oximetry measurements of the hemoglobin component fractions by falsely increasing the levels of MetHb and COHb. Blood levels of MetHb and COHb cannot be reliably determined by the co-oximetry when OHCbl is known or suspected.

The use of high-dose hydroxocobalamin (OHCbl) has become more frequent in clinical practice, with no significant adverse events reported, even with doses as high as 30 g within 24 hours.¹ Hydroxocobalamin is currently marketed under the trade name Cyanokit (BTG International) and has become the antidote of choice in cyanide poisoning in the United States, replacing amyl nitrate, sodium nitrate, and sodium thiosulfate.^2,3 A side effect of OHCbl administration is rapid, sustained, and significant increase in blood pressure, which is proportional to dosing of OHCbl, affecting 18% of those receiving a 5 g dose and 28% of those receiving a 10 g dose.^4,5 This side effect of high-dose OHCbl is now being exploited for its utility as a potential therapy for the treatment of vasoplegia and severe refractory hypotension,⁶ particularly in the setting of cardiac surgery following cardiopulmonary bypass.^7–9

Although high-dose OHCbl has been proven to be beneficial to the treatment of vasoplegic syndrome, there are increasing reports that high-dose OHCbl administration may also cause elevated levels of methemoglobin (MetHb), which can interfere with serum and urine laboratory colorimetric assays of certain parameters such as clinical chemistry, hematology, coagulation, and urine parameters.^10–12 There are no comprehensive clinical reviews that evaluate the effects of high-dose OHCbl on arterial blood gas (ABG) and oximetry measurements in patients with vasoplegic syndrome. In this study, we describe the first case series measurements of ABG and oximetry related to high-dose OHCbl treatment for patients with vasoplegic syndrome. Furthermore, we investigated and report on the potential interference effects of high-dose OHCbl on several ABG and oximetry values, with primary interest in MetHb, total hemoglobin (tHb), carboxyhemoglobin (COHb), arterial oxygen saturation (SaO₂), arterial oxygen partial pressure (PaO₂), and arterial carbon dioxide partial pressure (PaCO₂).

Methods

Study Population and Setting

This was a retrospective chart review of all consecutive adult patients aged ≥18 years who were hospitalized in the Memorial Healthcare System, Hollywood, FL, between May 1, 2017, and December 31, 2020, who received OHCbl administration for the treatment of refractory vasoplegic syndrome. Patients were excluded if (1) OHCbl was ordered but never verified as given; (2) dose was given but did not have the requisite pre- or post-ABG with oximetry within 48 hours before and within 48 hours after receiving the OHCbl. Any repeated doses given to patients were not excluded, with any number of subsequent doses acceptable. This study was approved by the institutional review board of the Memorial Healthcare System with a waiver of informed consent (MHS.2021.057).

Data Collection and Definitions

Data involved in this study were extracted from electronic medical records. We collected individual baseline demographic data (age, gender) and ABG and oximetry values, including pH, PaCO₂, PaO₂, COHb, MetHb, and tHb. The OHCbl order time, dispense time, administration time, and doses were also collected.

The Cyanokit (BTG International) for intravenous infusion was standardized to consist of 1 vial containing 5 g lyophilized OHCbl dark red crystalline powder for injection, which is reconstituted with 200 mL of a diluent with normal saline solution for a final concentration of 25 mg/mL. The starting dose of OHCbl for patients was 5 g administered as an intravenous infusion over 15 minutes and could be followed by repeated doses, if desired, depending on the clinical response. We maximized our data collection by accepting baseline ABG with oximetry measurements within 48 hours prior to dose being given, although selected results that were closest to the time of OHCbl administration. Similarly, we included any blood gas and oximetry results within 48 hours after OHCbl administration, although selected the results that were soonest after the dose was given.

All blood samples were run on the ABL90 FLEX PLUS blood gas analyzer (Radiometer Medical) calibrated with internal reference standards. The oximeter analysis provided values of tHb (grams per deciliter) and percentage levels of COHb and MetHb. This instrument is a hospital clinical laboratory grade self-calibrating co-oximeter with accuracy of detection of MetHb and COHb of ± 0.5%-1%, depending on range. If an ABG was run without co-oximetry, the particular dose or sample was excluded from data analysis.

Outcome Measures

The main outcomes of the study were to examine changes in tHb, COHb, and MetHb values in blood after high-dose OHCbl administration and determine whether OHCbl administration interferes with ABG and oximetry values. Interference in measurements was defined as differences in the measured and test results that were caused by OHCbl administration. The magnitude of interference was deemed clinically acceptable if the difference between the pre- and post-samples was within ±2% (absolute) for MetHb and COHb fractions and within ±1.0 g/dL for tHb. The relation of these altered measurement values with the measured MetHb levels were assayed. In addition, a subanalysis of 20 patients with a single dose of OHCbl administration was performed to trace MetHb levels and determine the time of MetHb level returned to the pre-drug baseline level.

Statistical Analysis

Quantitative result was reported as mean with SD or median with IQR, as appropriate. Variables were tested for normality of distribution using a Kolmogorov–Smirnov test. Differences in ABG and oximetry values before and after the OHCbl administration were compared using nonparametric Mann–Whitney U test. The MetHb levels across different dose groups were compared using 2-way analysis of variance with Tukey’s post hoc test. The MetHb levels were classified into 1 of 3 severity grades (<5%, 5%-10%, and ≥10%), and the relationship between the grade of MetHb elevation and ABG/oximetry values were analyzed by a nonparametric Kruskal–Wallis test. The Spearman rank correlation coefficient was used to determine whether a relationship between variables existed. All statistical tests were 2-sided and P < .05 was considered statistically significant. Data were analyzed using SPSS version 28 (SPSS).

Results

A total of 95 patients who received OHCbl were analyzed. Median age of the study patients was 60 years (IQR, 54-70) and 73 patients (77%) were male. Of the 95 patients, 65 patients (68%) received 1 dose, 21 patients (22%) received 2 doses, 5 patients (5%) received 3 doses, 3 patients (3%) received 4 doses, and 2 patients (2%) received 5 doses; and no patient received more than 5 doses.

Effect of OHCbl Administration on Blood MetHb Levels

As shown in FIGURE 1, the baseline MetHb level was approximately 1.1% ± 0.3% of total hemoglobin with a median of 1.0% (IQR, 1%-1.2%). After 1 dose of 5 g OHCbl, MetHb level increased to 5.2% ± 0.5% (P < .001) with a median of 4.8% (IQR, 3.0%-6.5%), except for 2 patients who had no change in their MetHb values. The MetHb increase was as high as 13.1% in 1 patient. We performed a subgroup analysis of the patients who received multiple doses of OHCbl and found that significantly elevated MetHb levels were observed at all doses used in the study (all P < .001; FIGURE 1B), whereas changes in MetHb levels were not significantly different across the doses of OHCbl given (P = .117; FIGURE 1B).

In a subanalysis of 20 patients who had repeated follow-up ABG samples post-OHCbl, the median time of the MetHb value returning to the baseline level was 105 hours (IQR, 59−132 hours). An example of the approximate time course of the MetHb levels following OHCbl infusion as obtained in 1 patient is shown in FIGURE 1C. The concentration of MetHb was found to reach levels above 5.0% within 3 hours following OHCbl infusion then declined gradually and returned to the baseline level by approximately 60 hours post-OHCbl infusion in that particular example.

Effect of OHCbl Administration on the Measurements of Non-MetHb Values

FIGURE 2 shows the measurements of tHb, COHb, PaO₂, PaCO₂, and SaO₂ before and after OHCbl infusion. Prior to OHCbl infusion, the mean COHb level was 1.4% ± 0.5% with a median of 1.3% (IQR, 1.0%-1.8%). After OHCbl (5 g) infusion, COHb level increased to 1.8% ± 0.7% with a median of 1.7% (IQR, 1.3%-2.2%), which was significantly higher than the level of the pretreatment (P < .001) (FIGURE 2B). The tHb concentrations before and after OHCbl infusion were similar in the study subjects (9.9 ± 2.2 g/dL vs 9.9 ± 2.1 g/dL; P = .868) (FIGURE 2A). The PaO₂, PaCO₂, and SaO₂ measurements in the subjects were not different before or after OHCBl treatment (FIGURE 2C-2E).

OHCbl Interference with MetHb, COHb, and tHb Measurements

FIGURE 3 shows the absolute changes of MetHb, COHb, and tHB in patients who received one dose (5 g) OHCbl. Of the 95 patients studied, 27 patients (28%) presented with the magnitude of interference within ±2% and 68 patients (72%) with interference of >2% for MetHb, whereas the magnitude of interference for COHb was within ±2% in all patients (FIGURE 3A). Of the 95 patients studied, 70 (74%) presented with interference of ±1g/dL for tHb, and 26% of patients (n = 25) had a magnitude of interference of >1 or <1 g/dL for tHb (FIGURE 3B).

Relationship Between MetHb Level and Blood Gas and Oximetry Values

Increases in MetHb values were classified into 3 severity levels: mild increase being <5%, moderate increase of 5%-10%, and severe increase of >10%. We performed a subgroup analysis to determine the difference of blood gas (PaCO₂, SaO₂) and oximetry values (tHb, COHb) among the different concentration increases in MetHb levels (FIGURE 4). There was a significant difference in the levels of tHb and COHb across the different levels of MetHb (Kruskal–Wallis tests, P = .004 and P < .0001, respectively) (FIGURE 4A and 4B). There was a weak negative relationship between MetHb concentration and tHb level (Spearman coefficient = -0.293, P = .0041), whereas there was a significant positive relationship between MetHb and COHb (Spearman coefficient = 0.565, P < .0001). There was no difference in the levels of PaCO₂ across the different concentration of MetHb groups (Kruskal–Wallis test, P = .385), whereas SaO₂ level appeared to be different across the different concentration of MetHb groups (Kruskal–Wallis test, P = .022) (FIGURE 4C and 4D), which was significantly different in the values of SaO₂ between the MetHb > 10% and the MetHb < 5% groups (P = .0278). There was a weak positive relationship between the MetHb concentration and SaO₂ level (Spearman coefficient = 0.324, P = .0014), with no significant relationship between MetHb and PaO₂ level (Spearman coefficient = -0.135, P = .193).

Discussion

The main finding of this study is that OHCbl administration significantly affects the arterial blood oximetry values in patients with vasoplegic syndrome. This phenomenon has been observed in several case reports with poisoning,^13–15 but this study is the first comprehensive investigation of a relatively large number of patients with hemodynamic failure since the first report using OHCbl as a rescue therapy for the treatment of refractory vasodilatory shock in 2014.⁷ Results from our study demonstrate a baseline MetHb value of 1.1% in these patients, with an average rise to 5.2% after 5 g OHCbl administration accompanied by a significant increase in COHb level with an average rise to 1.7%. By subgroup analysis, we found a strong positive correlation between the concentrations of COHb and MetHb as measured by the ABL90 FLEX PLUS blood gas analyzer. The SaO₂ values were found to positively correlate with the level of MetHb, and no significant correlation was observed between the PaCO₂ values and the level of MetHb. This is the first study to identify such correlations. Thus, OHCbl interferes with the determination of MetHb and COHb by the ABL90 FLEX PLUS blood gas analyzer.

The question of whether MetHb elevation after OHCbl was artifactual versus true methemoglobinemia was not part of this study but is interesting to consider.

From a strictly clinical standpoint, despite showing MetHb level elevation after OHCbl administration, the level of MetHb in our study did not reach any level >13%, well below critical or symptomatic methemoglobinemia levels. This would suggest that even with single or multiple doses of OHCbl, the MetHb levels do not rise to levels that would be defined as more than “mild.” Therefore, even if the MetHb level elevation observed in our study and others was real and not an artifact, it would be of limited clinical significance. Additionally, true methemoglobinemia in nonsevere or mild forms is often difficult to establish and would require investigation of more subtle clinical features, which was not part of this study. However, we did find that the SaO₂ remained in the normal range in our patients despite significant MetHb elevation, which is suggestive that the rise in MetHb in our study was an artifact and not real.

From the standpoint of determining whether MetHb and COHb elevation is a result of colorimetric interference on measurements, this was not determined by our study, as only the ABL machines were used. The ABL90 FLEX PLUS machines that were used in our study have been previously investigated by Pamidi et al¹¹ and were shown to exhibit significant variability of MetHb measurements after OHCbl when compared with the GEM Premier 4000 and Siemen’s Rapidpoint 405 analyzers. The highest amount of discrepancy in measurements was observed on the ABL line of machines in the Pamidi study. This discrepancy between different manufacturer’s machines is curious, as the basic physical properties of each measured compound to absorb or reflect light within the 475 nm-655 nm wavelength spectrum should be consistent, even if affected by the red pigment of OHCbl. As only the ABL line of machines was used in our study, we cannot comment further on variation between different manufacturers of these machines. Nevertheless, hardware, software, or chemical components related to the photometry light source, varying software that drives the light source, and varying algorithms that interpret the measurements are involved. Other contributing factors may include but are not limited to (1) blood lysing solutions and their propensity for “foaming,” (2) process control solution, solution pack, or cassette, (3) reagent type and reagent shelf life, (4) “intelligent quality management system” in some devices, and (5) machine calibration.

Regarding tHb, data presented in FIGURE 3B do not suggest an interference pattern for tHb in the presence of OHCob; rather, there is an inverse correlation between tHb values and the level of MetHb. In FIGURE 4A, the value of tHb is shown to be significantly lower in patients with a MetHb level of >10% than those with a MetHb level of <5% and those with MetHb level of 5%-10%. As percentage of MetHb was calculated by dividing the concentration of MetHb by the concentration of tHb, the presence of the same concentration of MetHb after OHCbl infusion in an anemic patient would represent a higher percentage of MetHb. Thus, the underlying medical conditions in these patients (eg, anemia) might contribute to this observation.

In regard to those patients who had repeat doses of OHCbl, the change in the concentration of MetHb was not detected in a dose-dependent manner. We observed that despite repeated dosing of OHCbl, elevated concentration of MetHb did not correlate. It appears that the patients who received repeated doses of OHCbl had a less pronounced increase in MetHb level if they received their dose during the time when the previous dose was still circulating. This decremental rise in MetHb values is likely reflective of the early repeat dosing of OHCbl within 24 to 48 hours following the initial dose. Essentially, the repeat OHCbl dose was given before the MetHb value had returned to baseline, thereby not allowing enough “wash-out time” before the repeat dose. This indicates that the repeat doses were given in the period of time in which the MetHb values were high and still had not decreased significantly toward their baseline. As this was a retrospective study, we did not have standardized times that the pre- and post-samples were measured in relation to the time the OHCbl was given, limiting our data somewhat. This finding suggests that the MetHb level acts as a surrogate qualitative marker for presence of circulating OHCbl but not necessarily as a quantitative marker of OHCbl levels.

In the subanalysis of the 20 patients who received 1 dose OHCbl and had repeated follow-up samples post-OHCbl, the median time required to return to prior baseline value of MetHb was 105 hours (IQR, 59−132 hours). Anecdotally, after OHCbl administration, the urine of patients will stay a “red wine” color for up to 1 week after drug administration. One could imply from this urine discoloration that the serum retains a similar pigmentation change, thereby effecting arterial blood oximetry measurements. As a surrogate, the clearing of urine may coincide with the clearing of MetHb in co-oximetry measurements; however, urine color or color intensity was not accounted for in this study.

There are several limitations to our study. First, this study was limited by its retrospective nature, and some patients receiving OHCbl infusion were not included due to the lack of data analyzed, most notably the lack of oximetry in the blood gas analysis. Second, sampling intervals for ABG/oximetry were not consistent, as results obtained in patients were not in the same time windows following the administration of OHCbl. Therefore, differences in the time of sampling may have affected our results, although we aimed to correct for these differences by selecting the blood gas results that were soonest before or after the OHCbl was given. Also, we only used 1 type of blood gas machine in our study, limiting the ability to compare results between different machine manufacturers. Finally, the concentration of OHCbl presented in the blood was not estimated because it is not possible to measure OHCbl concentration in conventional laboratories.

Conclusion

Our findings clearly demonstrate that OHCbl administration interferes with the determination of the hemoglobin component fractions in whole blood by the ABL90 FLEX PLUS blood gas analyzer by increasing the measured MetHb and COHb fractions. These errors may potentially influence clinical decision making and thus affect patient outcomes. Our findings highlight that when OHCbl is known to be present, blood MetHb and CoHb cannot be reliably determined by the ABL90 FLEX PLUS blood gas analyzer.

Abbreviations

Abbreviations
 
• OHCbl

  hydroxocobalamin

 
• MetHb

  methemoglobin

 
• tHb

  total hemoglobin

 
• COHb

  carboxyhemoglobin

 
• HbO₂

  oxyhemoglobin

 
• PaO₂

  arterial oxygen partial pressure

 
• PaCO₂

  arterial CO₂ partial pressure

 
• SaO₂

  arterial oxygen saturation

 
• ABG

  arterial blood gas

Availability of Data and Materials

The dataset is available from the corresponding author on reasonable request

Conflict of Interest Disclosure

The authors have nothing to disclose.

REFERENCES