4. Discussions

4.1. Is MetHb Increased in COVID-19 Patients?
According to the case reports of Kuipers et al. [17], Al-Aamri et al. [23], Palmer et al. [22], Faisal et al. [20], and Naymagon et al. [18], MetHb values for COVID-19 patients were above the reference range of 0.67 ± 0.33% for healthy non-smokers (Borland et al. 1985) with the highest MetHb value of >30% for a patient reported by Naymagon et al. [18].
In the case reports of Palmer et al. [22], Faisal et al. [20] and Naymagon et al. [18], MetHb increased in the COVID-19 patients during the course of the disease. Data of multiple MetHb measurements during the disease course were published by Palmer et al. [22], showing a rise of MetHb to a peak after a few days and a decline after treatment (Figure 3d).
COVID-19 patients (n = 25) were shown to have a higher MetHb compared to healthy individuals (n = 25) as demonstrated in a cross-sectional study by Alamdari et al. [21], supporting the findings reported. However, the cross-sectional study of Soltan et al. [19] with a large cohort (534 COVID-19 patients and 114,957 pre-pandemic controls) showed no statistically significant differences in the MetHb values despite the fact that MetHb was an important parameter for the prediction of COVID-19 based on the algorithm the group developed. This apparent discrepancy between the result of Alamdari et al. and Soltan et al. seems to be due to the following reason: the MetHb data used by Soltan et al. stem from the time of emergency presentations and admission to hospital, whereas the MetHb data from Alamdari et al. were collected from the whole time-course of the hospital stay. Since MetHb has been reported to increase during the development of the disease [18,20,22], the results of Soltan et al. are understandable since, during the MetHb sampling time at the beginning of the disease, MetHb is not necessarily increased (at least at the group level).
In conclusion, MetHb seems to be elevated in COVID-19 patients, with a dynamic following the disease progression.

4.2. Is COHb Increased in COVID-19 Patients?
A significantly higher COHb value in COVID-19 patients compared to reference value of healthy non-smokers (0.5–1.5% for healthy non-smokers [16]) was found in two of the three available reports on COHb and COVID-19 published so far. In the case report of Faisal et al. [20], COHb was 3.2% during the disease course. According to the study of Paccaud et al. [24], COHb rose above the reference range on about the 10th day after hospitalization. Interestingly, this study also clearly demonstrated that at admission, COHb was in the normal range for the COVID-19 patients, while COHb was statistically significantly more elevated on the 31st day of the hospital stay in COVID-19 patients compared to patients suffering from other illnesses. These results are in line with the results of Pawloski et al. [25], who reported that their COVID-19 cohort did not show COHb values above the reference range during the time-span investigated (admission until day seven to nine of the stay). Since Paccaud et al. found that it needs about 10 days until COHb is above the reference range, the too short time-span used by Pawloski et al. to compare COHb values from COVID-19 patients and controls might explain the apparent discrepancy.
In conclusion, COHb can be elevated in COVID-19 patients, especially from about two weeks after onset of the disease. The magnitude of COHb elevation seems to be correlated with the survival probability of the COVID-19 patients.

4.3. Possible Reasons for Methemoglobinemia in COVID-19 Patients
There are several factors relevant to explain why methemoglobinemia and carboxyhemoglobinemia can be present in COVID-19 patients. The SARS-CoV-2 infection, the individual constitution and the medical treatment seem to be the major ones.
It is well known that several medical drugs can increase MetHb concentration in the blood as a side-effect [26,27]. A recent review reports the early recognition, pathophysiology, and management of methemoglobinemia in the intensive care unit [28]. Chloroquine is such a drug for which reports were published regarding induced methemoglobinemia due to its intake [29,30,31]. Both chloroquine and hydroxychloroquine (a derivative of chloroquine) are currently used to treat COVID-19, while debate is ongoing about their effectiveness and safety [32,33,34,35,36,37,38,39]. Moderately certain evidence suggests that hydroxychloroquine lacks efficacy in reducing short-term mortality in patients hospitalized with COVID-19 or at risk of hospitalization in outpatients with COVID-19.
A G6PD deficiency can enhance the probability of methemoglobinemia induced by oxidizing drugs, such as hydroxychloroquine [40]. However, a very recent experimental animal study suggests that short-course high doses of hydroxychloroquine do not induce methemoglobinemia or clinically significant hemolytic anemia or organ damage in a murine model of G6PD deficiency [41]. Clinically, chronic hemolytic anemia associated with G6PD deficiency is rare [42]. So far, no evidence of hemolysis was observed in patients with G6PD deficiency when exposed to low doses of hydroxychloroquine [43]. The first case of severe hemolytic crisis was found in a seriously ill COVID-19 patient with G6PD deficiency following treatment with high doses of hydroxychloroquine [44]. Several other cases have subsequently been reported by others [45,46,47]. Nevertheless, Mastroianni et al. [47] and Afra [48], concerning hydroxychloroquine use in G6PD-deficient patients, have indicated that it is difficult to assess the relationship between hydroxychloroquine and hemolysis in COVID-19 patients. With the global spread of COVID-19, especially in regions with a high prevalence of G6PD deficiency, these cases should alert physicians to the possible correlation between G6PD-deficiency and hydroxychloroquine treatment.
In the reports on MetHb and COVID-19 summarized in Section 3.2, the use of chloroquine or hydroxychloroquine is described. In the case reported by Kuipers et al. [17], the subject received chloroquine and had a G6PD deficiency; hydroxychloroquine was administered to the subject reported by Faisal et al. [20] (the G6PD status was not reported); and in the three cases reported by Naymagon et al. [18] all received hydroxychloroquine, while one subject was tested for G6PD deficiency and was positive. In the cases reported by Al-Aamri et al. [23] and Palmer et al. [22], the patients did not receive chloroquine and hydroxychloroquine, but the patient described by Al-Aamri et al. had a G6PD deficiency. A G6PD deficiency is not only relevant with respect to the reaction to antiviral oxidizing drugs, but also for the effects of the drug methylene blue administered to treat the methemoglobinemia. Methylene blue may be ineffective in patients with a G6PD deficiency since they lack sufficient reduced nicotinamide adenine dinucleotide phosphate (NADPH) to reduce methylene blue to leuko-methylene [49,50].
Despite the obvious effect of oxidizing drugs on the formation of MetHb, it can also be formed as a byproduct of a physiological reaction in the form of an adaptive increased nitric oxide (NO) signaling due to an acute anemia [51]. Anemia can be associated with an infection and/or a systemic inflammatory reaction, termed “anemia of inflammation”, as part of the physiological reaction to the disease [52,53]. According to a study by Bellmann-Weiler et al. [54] on 259 hospitalized patients with COVID-19, 24.7% were anemic on admission, with the majority suffering from anemia of inflammation (68.8%). During hospitalization, the percentage of patients with anemia increased (around 68.8% at day 7). A significantly higher mortality during hospitalization was also found in those with anemia upon admission. Anemia is associated with an increased NO expression, leading to vasodilation and thus preventing tissue hypoxia, but also causing increased NO-based oxidation of Hb to MetHb [51]. Interestingly, anemia, and drops in total Hb (tHb), have been reported in COVID-19 patients [22,55,56].
An anticorrelated development of tHb levels and MetHb levels in a COVID-19 patient has been reported [22], in line with the pathway described by Hare et al. [51] of an NO-induced MetHb increase caused by anemia. The role of MetHb as a marker of anemic stress has been also been validated in a study investigating MetHb changes in patients undergoing heart surgery [57]. While “anemia of inflammation” is associated with methemoglobinemia, iron-deficient anemia seems to be a further risk factor for acquired methemoglobinemia by enhancing red blood cell oxidative stress [58]. The occurrence of methemoglobinemia due to a viral infection has been reported for several types of infection [59,60,61]. For example, the activity of MetHb reductase has been shown to be negatively affected by infections with Plasmodium yoelii nigeriensis [62,63] or Plasmodium knowlesi [61].
The fact that the severity of the methemoglobinemia observed in COVID-19 patients was dependent on the subject firstly reflects the different disease severities of the patients but is also most probably due to the physiological constitution of the subjects before the disease, related to their medical history, their overall fitness and age. The age factor seems to be of particular interest since, for example, erythrocytes from elderly humans are more easily affected by oxidative stress, facilitating the formation of MetHb [64].
MetHb can have proinflammatory properties. For example, it activates the NF-κB pathway in endothelial cells associated with chemokine (IL-8) and cytokine (IL-6) production [65]. The activation of the NF-κB and MAPK pathways with subsequent release of the chemokines IL-8 and the chemokine monocyte chemoattractant protein-1 (MCP-1) has also been observed in endothelial cells exposed to MetHb [66]. This underlines that an elevation of MetHb in the blood has an effect on cytokine/chemokine production—a fact that might be of particular relevance for COVID-19 since a “cytokine storm” has been observed in severe courses of the disease [67,68,69,70]. At the same time, it must also be borne in mind that hypoxia also causes the production of cytokines and cytokines, like IL-8 and IL-6 [71,72].

4.4. Possible Reasons for Carboxyhemoglobinemia in COVID-19 Patients
Since the blood COHb concentration reflects the balance between endogenous CO production and CO elimination, carboxyhemoglobinemia in COVID-19 patients could indicate an increased endogenous CO production and/or a decreased CO elimination ability.
Endogenous CO production is mainly due to the inducible heme oxygenase (HO-1) enzyme, which catalyzes the heme moiety of Hb to biliverdin and liberates CO during this process. CO can then react with Hb, leading to the formation of COHb. HO-1 is upregulated in case of oxidative stress and inflammation which leads to increased COHb production [73,74]. Hemolytic anemia facilitates the production process of COHb so that an increased COHb blood level can be seen as a manifestation of hemolytic anemia [75]. Since anemia and hemolysis possibly occur during the course of disease in COVID-19 patients [17,18,22,54], hemolytic anemia may also be responsible for their COHb elevation. Because intracellular NADPH depletion and consecutive oxidative stress with damaged erythrocytes (hemolysis) is typical for G6PD deficiency, it is not surprising that G6PD deficiency in COVID-19 patients seems to be associated with elevated MetHb and COHb levels [20].
A decreased CO elimination occurs when respiration is impaired. As COVID-19 patients are characterized by respiratory impairment, increased COHb levels can be explained by reduced CO elimination and thus a higher probability of COHb formation. Mechanical ventilation may also be relevant since, for example, an increase in the inspired O2 fraction leads to an increase in exhaled CO concentration [76], possibly leading to a reduced COHb production.
Interestingly, while elevated COHb levels seem to be correlated with COVID-19 severity, intensive care mortality from other causes was found to be associated with too low [12,77] and both too low or too high COHb values [11], indicating the existence of an optimal COHb level for optimal physiological functioning [12]. HO-1 upregulation, associated with elevations of COHb, has immunomodulatory and anti-inflammatory effects [78]. Inflammation changes COHb levels in the blood in a complex time-dependent manner as demonstrated by experimental endotoxemia in humans [79], highlighting the non-linear relationship between inflammation, disease severity and COHb levels.

4.5. Methemoglobinemia and Carboxyhemoglobinemia in COVID-19 Patients: Consequences for Patient Monitoring and Treatment
Methemoglobinemia and carboxyhemoglobinemia seem to play a role in the pathophysiology of COVID-19, especially in more severe cases of the disease.
While the ability to determine MetHb and COHb levels is normally routinely available in clinical settings via blood gas analysis, there are only a few commercial monitoring devices that enable continuous non-invasive measurement of MetHb and COHb values, mainly the fingertip pulse CO-oximeters by Masimo and Nonin. These devices do not normally feature in standard clinical equipment. This is unfortunate since continuous monitoring of tHb, MetHb and COHb levels could be helpful in guiding the treatment and monitoring of COVID-19 disease progression.
From the reports discussed in Section 3.2 and Section 3.3 it is clear that the determination of MetHb and COHb is especially warranted when patients are treated with oxidizing drugs such as chloroquine and hydroxychloroquine. This is even more important when the patients have a confirmed G6PD deficiency.
For the interpretation of COHb values, the smoking status of the patient needs to be considered as smokers have a higher COHb concentration in the blood than non-smokers (2.7 ± 2.6% [80], 3.26 ± 2.2% [15], 5.12 ± 2.25% [81], 2.1 ± 1.02% [82]).
Knowledge of MetHb and COHb levels in the blood of COVID-19 patients is also relevant to prevent misinterpretations of arterial oxygen saturation values measured with fingertip pulse oximetry (SpO2). This is because MetHb and COHb interfere with the measurement of SpO2. An overestimation of the true arterial oxygenation (SaO2) can occur. In case of a decrease in SaO2 and an increase in MetHb or COHb, SpO2 will diverge more from SaO2 the higher the MetHb and COHb concentration (see Figure 4). For example, assuming a MetHb concentration of around 25% (corresponding to the upper end of the confidence interval of MetHb values in COVID-19 patients reported by Alamdari et al. [21]) and an assumed decrease of SaO2 to 75%, the SpO2 measurements would indicate a falsely too high SpO2 of about 88%. Measurement with pulse CO-oximetry instead of pulse oximetry would circumvent this problem since pulse CO-oximetry is able to non-invasively measure MetHb, COHb, tHb, and the correct SpO2. [83,84,85].
Methemoglobinemia and carboxyhemoglobinemia cause a shift in the Hb dissociation curve to the left, leading to a reduced ability of O2 to be released from Hb, which can result in hypoxia. Methemoglobinemia and carboxyhemoglobinemia need to be monitored and treated, therefore. Therapies of methemoglobinemia in COVID-19 patients have been performed with methylene blue and, in some cases, combined with blood transfusions [18,20,21].
The fact that COVID-19 patients could show hypoxemia without having dyspnea, i.e., silent hypoxia or so-called “happy” hypoxia [88,89,90], seems to be primarily due to (i) a blunted response of the respiratory control system to hypoxia which is prevalent in older subject and those with diabetes, (ii) changes in arterial CO2 levels, (iii) temperature-induced shifts in the O2 dissociation curve, and (iv) the inaccuracy of pulse oximeters at low SpO2 values, as highlighted by Tobin et al. [89,91].
For hospitals it might be also relevant to re-evaluate their water disinfection procedure since the use of a hydrogen peroxide/silver ion preparation for treating the water supplied in the hospital caused elevated levels of MetHb in the severely ill patients (treated with daily hemodialysis/hemodiafiltration) drinking this water [92]. When blood transfusions are given to COVID-19 patients, it should be also considered that the MetHb content of the banked blood increases over time [93] and that banked blood from smoking donors can have a relatively high COHb concentration [94], representing a possible risk for critically ill patients. It makes sense therefore to test the banked blood for MetHb and COHb concentration levels before administering it to patients, especially in case of COVID-19.