Tramadol is a commonly used analgesic in ICUs. It is metabolized in the liver through cytochrome P450 to 11 demethylation products. The most important ones are O-demethyltramadol and N-demethyltramadol. The CYP2D6 isoenzyme metabolizes tramadol into the active ODT, which is responsible for most of the analgesic effect and has a 200-fold greater affinity for opioid receptors than tramadol (Gillen et al., 2000; Grond and Sablotzki, 2004). CYP2D6 gene is extremely polymorphic, with over 100 allelic variants (Fulton et al., 2019). The alleles vary between fully functional to completely non-functional, producing thus a variety of metabolic phenotypes: ultrafast (UM), extensive (EM), intermediate, and poor metabolizers (Ganoci et al., 2017; Dagostino et al., 2018). In PMs low amounts of ODT are produced and the analgesic effect of tramadol is markedly reduced. Conversely, UMs have a high concentration of active metabolites and could be at increased risk for toxicity (Smith et al., 2018). N-demethylation of tramadol into the inactive NDT is catalyzed by the isoenzyme CYP3A4 and CYP2B6. Although both isoenzymes exhibit gene polymorphism and are susceptible to induction or inhibition by some substrates, no significant differences in tramadol metabolism were observed (Miotto et al., 2017). In addition to the genotype, cytochrome activity is influenced by a number of pathophysiological factors, including proinflammatory cytokines, which reduce cytochrome activity (He et al., 2015). In the work of Tanaka and co-workers, a high level of interleukin (IL)-6 was associated with an increased N-demethylation of tramadol (Tanaka et al., 2018).

Surgery causes a physiological inflammatory response consisting of complex metabolic, hemodynamic, hormonal, and immune changes, which ensure wound healing after surgery (Finnerty et al., 2013). Tissue injury causes an increase in proinflammatory cytokines, tumor necrosis factor-alpha (TNF)-alpha, IL-1, and IL-6, as well as anti-inflammatory cytokines, IL-10 (Hsing and Wang, 2015). Under physiological conditions, the pro and anti-inflammatory systems are in equilibrium. Dysregulation of the immune system with excessive activity of the proinflammatory response leads to the development of systemic inflammation (Paruk and Chausse, 2019). Systemic inflammation occurs in more than 40% of patients during hospitalization, and is particularly common in surgical ICUs, where its prevalence is more than 80% (Brun-Buisson, 2000; Churpek et al., 2015). There are numerous biological markers of systemic inflammation, and in routine clinical practice, the most common is the use of C-reactive protein (CRP) and procalcitonin, whose synthesis is induced by IL-6 (Paruk and Chausse, 2019).

In recent years, the role of the cholinergic nervous system in maintaining homeostasis during the inflammatory response has been extensively studied, and a marker of the cholinergic system readily available in everyday clinical practice is plasma cholinesterase activity (ChE). The cholinergic nervous system plays a central role in inflammatory processes and it is an efferent part of the neuroimmunological reflex. The inflammatory response caused by surgical injury stimulates the activity of the parasympathetic nervous system and activates the anti-inflammatory process as part of the nervous control of innate immunity (Tracey, 2010; Zivkovic et al., 2017). The cholinergic anti-inflammatory process is mediated by acetylcholine and acts by inhibiting the production of TNF-alpha and IL-1 and suppressing the activation of nuclear factor-kappa B (Das, 2007). Cholinesterase hydrolyzes acetylcholine and reduces its plasma concentration. The exact mechanism of reduced plasma ChE activity in states of acute inflammation has not yet been completely clarified. The low plasma ChE activity tends to maintain high levels of acetylcholine and enhance the negative feedback of the cholinergic system to acute inflammation. Plasma ChE activity reflects cholinergic, non-nervous activity and neuroimmune interactions. Low plasma ChE would signal an interruption in acetylcholine hydrolysis and disruption of immune homeostasis and is the earliest predictor of systemic inflammation, which changes in plasma even before standard inflammatory biomarkers. (Zivkovic et al., 2016). The aim of this study was to measure plasma concentrations of tramadol, ODT, and NDT in the patients admitted to the surgical ICU after major abdominal surgery. We hypothesized that patients with preoperative low plasma ChE activity as part of systemic inflammation in the early postoperative period would have higher plasma NDT levels and a different ratio of tramadol and its demethylation metabolites.

This prospective observational study recruited patients admitted to the surgical ICU after major abdominal surgery. The study was approved by the by the Ethics Committee of the Osijek University Hospital (No. 12272–7/2017), and all patients included in the study signed informed consent. Major abdominal surgery was defined as an open laparotomy with resection of parts of the digestive system. Exclusion criteria for study was a known allergic reaction to tramadol, tramadol therapy in the last 7 days prior surgery, patient age younger than 18 years old, body mass index (BMI) < 18 kg per m² or >40 kg per m², laparoscopic approach to surgery, and liver or renal failure verified before surgery according to Child-Pugh and Kidney disease Improving Global Guidelines (KDIGO) criteria (Durand and Valla, 2005; Khwaja, 2012). Also, patients on chronic therapy with cimetidine, paroxetine, pimozide, metoclopramide, amiodarone, olanzapine, chlorpromazine, fluphenazine, haloperidol, thioridazine, risperidone and clozapine were not included in the study as these drugs inhibit the activity of CYP2D6 enzyme (Bahar et al., 2017).

Pre-operative blood samples were taken in all patients for white blood count (WBC) (Sysmex XN-2000, Sysmex, Kobe, Japan), procalcitonin (PCT) (Roche Cobas E 411, Roche Diagnostics GmbH, Mannheim, Germany), c-reactive protein (CRP), lactate level, urea, creatinine, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyltransferase, total bilirubin, albumin, plasma ChE activity (Beckman Coulter AU680, Beckman Coulter, Brea, CA, USA), and arterial blood gas analysis (Radiometer ABL800 FLEX, Radiometer Medical A/S, Bronshoj, Denmark), according to the manufacturer’s instructions. CYP2D6 gene polymorphism was determined in all patients.

All patients had general anesthesia with sevoflurane in oxygen and air with standard intraoperative monitoring. An arterial catheter was placed in all patients. Induction was with either propofol or etomidate and rocuronium or succinylcholine depending on patients' hemodynamic status and the procedure urgency. After the surgery, patients were admitted to the ICU, where mechanical ventilation was continued until extubation, as well as monitoring of vital parameters. The development of systemic inflammation during the first 24 h was defined as the presence of at least two of four major criteria: tachycardia >90 beats minute⁻¹, fever >38°C or hypothermia <36°C, WBC >12.000/mm³ or <4.000/mm³, and PaCO₂ < 4.3 kPa (Marik and Taeb, 2017). Preoperative CRP >50 mg L⁻¹ and PCT >0.5 μg L⁻¹ were also considered systemic inflammation (Castelli et al., 2004).

Between January 2019 and January 2020, 50 consecutive patients were eligible for the study. Three patients were excluded from the study due to technical errors in the ICU protocol or errors in blood sampling for analysis, and 47 patients were analyzed. The demographic characteristics of the patients are presented in Table 1. According to CYP2D6 genotype, 2 (4%) were PM, 22 (47%) IM, 22 (47%) EM, and 1 (2%) patient was UM. CYP2D6 diplotype and metabolic phenotype are shown in Table 2.

The result of this study showed that concentrations of tramadol, ODT, and NDT are lower in surgical ICU patients than observed earlier in medical patients and healthy volunteers. Data on the concentrations of tramadol and its metabolites available in the literature were obtained mostly on non-surgical patients or healthy volunteers, and the concentrations were measured after a single dose of tramadol. Kirchheiner et al. administered a single oral dose of 100 mg tramadol and measured tramadol C_max and ODT C_max of 208 μg L⁻¹ and 106 μg L⁻¹, respectively (Kirchheiner et al., 2008). Ardakani et al. measured tramadol C_max of 314.4 μg L⁻¹, ODT 88.6 μg L⁻¹, and NDT 33.4 μg L⁻¹ in women after a single dose of 100 mg tramadol (Ardakani and Rouini, 2007). However, the CYP2D6 gene polymorphism was not examined in the latter study. Approximately similar concentrations of tramadol, ODT, and NDT were achieved in our subjects only after repeated IV doses of tramadol. ODT concentrations 4 h after tramadol injection in EM were two times lower than the concentrations achieved after only one oral dose of tramadol in healthy subjects in a study by Kirchheiner et al. (Kirchheiner et al., 2008).

Using patient-controlled analgesia, Lehmann et al. concluded that the minimum effective plasma concentrations of tramadol and ODT were 287 μg L⁻¹ and 36.2 μg L⁻¹, respectively (Lehmann et al., 1990). This level of tramadol was reached in most of our subjects, but this level of ODT was achieved in the first measurement only in EM, while IM reached this ODT level only in the last measurement. Nevertheless, tramadol analgesia was effective in both groups of patients. The metabolic ratio of ODT/tramadol in most studies was about 0.3 (Nobilis et al., 2002; Grond and Sablotzki, 2004; Siepsiak-Połom et al., 2019), while in our IM and EM patients after 2 h it was only 0.07 and 0.16, respectively. The ratio of 0.2 was reached after the 2^nd and 3^rd doses of tramadol only in EM patients.

The diversity of previously conducted studies is the reason that the data obtained in our study can only be partially compared with them. The comparison with previous studies is difficult because many researchers have not determined the polymorphism of the CYP2D6 gene. However, the lower concentrations measured in surgical patients in our study differed significantly from the concentrations in medical patients and healthy volunteers (Ardakani and Rouini, 2007; Stamer et al., 2007; Kirchheiner et al., 2008; Lu et al., 2020).

Following the major surgical procedure and high-volume resuscitation in the first 24 postoperative hours, the lower plasma concentrations of tramadol, ODT, and NDT measured in our patients are probably associated with increased volume of distribution. The higher volume of distribution in postoperative patients is associated with hyperpermeability of vascular endothelium which occurs early during inflammation and in the postoperative period due to glycocalyx damage (Uchimido et al., 2019). Endothelial damage and impaired vascular permeability combined with volume resuscitation can significantly increase the volume of drug distribution with a consequent decrease in their plasma concentration (Charlton and Thompson, 2019).

The differences in the concentration of tramadol and its metabolites, as observed in or study, may differ from those reported in other studies; this could be due to different patient populations, as well as due to potentially different approaches to the analytical method used. Our HPLC method for the determination of tramadol concentration is standardized, but small differences in the method and thus in the results may exist among different researchers. Further research (with a control group of non-surgical patients, analyzed with the same analytical method) is needed for a more definite conclusion.

The enzyme cholinesterase in our study significantly correlated with the development of systemic inflammation, and with changes in the concentrations of NDT. Zivkovic et al. have observed that the plasma activity of ChE after trauma-induced systemic inflammation decreases before standard proinflammatory biomarkers began to increase, and confirmed its importance in the early diagnosis of systemic inflammation, and in predicting morbidity and mortality (Zivkovic et al., 2016; Zivkovic et al., 2019). Li et al. demonstrated that low ChE activity at the time of hospitalization was an independent risk factor for death in patients with ischemic stroke. Patients with low ChE activity had higher CRP values, lower albumin, were more often anemic, and had increased mortality and longer hospital stays, as observed in our patients (Li et al., 2020).

Studies in healthy individuals have confirmed that in PMs, due to genetically reduced CYP2D6 activity, tramadol metabolism is predominantly redirected (metabolic switch) to N-demethylation (Quetglas et al., 2007; Haage et al., 2018). Our study suggests that a similar metabolic switch may also occur during systemic inflammation with low ChE activity.

To date, no study of tramadol metabolism has been conducted in surgical patients with signs of acute systemic inflammation, and this is the first study to include both elective and emergency patients with inflammatory conditions. Like previous studies, this one confirmed that urgent surgical patients with clinical signs of systemic inflammation were more likely to have low ChE activity (Becher et al., 2012; Smajic et al., 2018).

Surgery and acute inflammation reduced CYP3A4 enzyme activity in the study conducted by Haas et al. in postoperative patients (Haas et al., 2003). This inhibition can lead to changes in the metabolism of drugs that are catalyzed by this enzyme. Proinflammatory cytokines are thought to act indirectly on cytochromes, reducing the expression of transcription factors such as HNF4- α (hepatocyte nuclear factor 4—α) which is responsible for the transcription of the CYP2D6 gene (He et al., 2015). However, CYP2B6, important in N-demethylation of tramadol in NDT, in contrast to CYP2D6 and CYP3A4 enzymes, shows a different response to proinflammatory cytokines. Thus, for example, IL-1, which is a potent inhibitor of the CYP3A4 enzyme, has no influence on CYP2B6. Transforming growth factor-beta (TGF-β), which is secreted from hepatocytes during stress and inflammation, induces CYP2B6 enzyme activity almost two and half times, which was not observed with CYP3A4 and CYP2D6 (Aitken and Morgan, 2007; Babeu and Boudreau, 2014). Unhindered CYP2B6 pathway may be a reason for higher NDT concentration in emergency patients and patients with low ChE activity.

A complex combination of genetic factors, the effect of cytokines on cytochromes, ODT transporters on hepatocytes (Vee et al., 2009; Tzvetkov et al., 2011), and altered volume of distribution may be the reason for the altered ODT concentrations in postsurgical patients. Also, the same factors may have been the cause of the loss of genetically determined differences in ODT concentrations in the group of patients with systemic inflammation. Although the sample of patients in this analysis was small, this observation requires attention and study in a larger number of subjects. All of these factors may influence the level of pain reported, and the effectiveness of drugs used.

Patients with acute inflammation and low pre-operative ChE with altered concentrations of metabolites had a lower expression of pain at all measurement points compared with those having normal ChE. This can be partly explained by brain dysfunction that may co-exist with systemic inflammation, as observed by McGrane and co-workers. The authors found that high values of proinflammatory biomarkers are a good predictor of delirium and coma during critical illness. They have observed that systemic inflammation is important in the pathophysiology of acute brain dysfunction in critically ill patients (McGrane et al., 2011). It should be emphasized here that higher concentrations of inactive NDT metabolite measured in the low ChE patients are not the cause of decreased pain reported. Such higher concentrations are merely a consequence of systemic inflammation that alters both the pain perception and the tramadol metabolism (McGrane et al., 2011; Uchimido et al., 2019).

With the introduction of fast-tracking within the Enhanced Recovery After Surgery protocols (Grant et al., 2017), it is expected that an increasing number of elective patients will bypass the ICUs. At the same time, an increasing proportion of ICU patients will likely be elderly, with numerous comorbidities and emergencies. In these patients, inflammation will be more common, as well as changes in drug metabolism.

The inability to compare pain and PONV in all study patients is a drawback of this study. Predominantly emergency patients who were longer drowsy and were extubated a few hours after ICU admission were not included in the pain analysis at initial pain assessments. It is possible that the altered consciousness in these patients was influenced by systemic inflammation, as mentioned earlier. Furthermore, we have not determined the tramadol enantiomers and their metabolites. The pharmacokinetics of tramadol and its metabolites are known to be stereoselective (Quetglas et al., 2007) (+) ODT is a more potent analgesic than the (−) ODT enantiomer (Grond et al., 1999). It is not yet known how systemic inflammation affects the synthesis of different enantiomers whose ratios can be altered in postoperative patients.

Personalized medicine strives to individualize therapy and determine treatment of patients based on genotype to increase treatment effectiveness with the lowest risk of adverse reactions (Caudle et al., 2014). Our and numerous other studies have confirmed that cytochrome activity is strongly influenced by genetic and pathophysiological factors, one of which is systemic inflammation. This study showed that ChE correlates with systemic inflammation, in addition to affecting tramadol metabolism, also alters pain perception. Consequently, it seems reasonable to conclude that inflammation is indeed a covert threat to effective genotype-based therapy (Shah, 2017).

In conclusion, our study in postoperative surgical patients has confirmed that the O-demethylation of tramadol is predominantly influenced by the CYP2D6 polymorphism, whereas N-demethylation is under the strong influence of systemic inflammation. Systemic inflammation also changes the perception of pain, and future studies should confirm whether and how the dose of tramadol should be changed in patients with systemic inflammation.