2024 Volume 49 Issue 10 Pages 447-457
Caffeine (CFF) is efficiently absorbed after ingestion, and approximately 80% of ingested CFF is metabolized to paraxanthine (PXT). Although PXT has approximately twice the adenosine receptor antagonist activity of CFF, there are few reports measuring the metabolite concentrations during CFF intoxication. Furthermore, no studies have examined the efficacy of hemodialysis (HD) on PXT or the indicators that contribute to treatment strategies for patients with acute CFF intoxication. This study analyzed the association between CFF and PXT blood levels, the blood biochemical data, and the vital signs of 27 cases with information on CFF intake and elapsed time data. It was found that HD was not as effective as CFF against PXT in CFF intoxication; however, HD was effective in cases with relatively high PXT concentrations (>10 μg/mL). Simultaneous analysis of CFF and PXT would make it possible to estimate the time elapsed from CFF intake and the risk of hyperCKemia, which may develop in cases left untreated for a prolonged period after ingestion. Therefore, the measurement of PXT, in addition to CFF, is expected to provide useful information for understanding the pathogenesis of CFF intoxication and the development of treatment strategies of acute CFF intoxication.
Caffeine (1,3,7-trimethylxanthine, CFF) is a plant-derived purine alkaloid found in coffee beans, tea leaves, and kola nuts and is known as a methylxanthine derivative. Because of its central stimulant and excitatory effects, it is added to soft drinks known as energy drinks, over-the-counter drugs, such as anti-drowsiness and cold remedies, and dietary supplements. Since these products are readily available, acute CFF poisoning has increased in recent years, with some cases resulting in death (Cappelletti et al., 2018), and CFF poisoning has become a social problem.
CFF is efficiently absorbed after ingestion and is metabolized into paraxanthine (1,7-dimethylxanthine, PXT) (84%), theobromine (12%), and theophylline (4%) mainly by CYP1A2 in the liver. Among those metabolites, PXT possesses an adenosine receptor antagonistic effect approximately twice as strong as CFF (Fredholm et al., 2001; Kull et al., 1999) and may contribute to the development of CFF intoxication symptoms, such as nausea, vomiting, fatal arrhythmia, circulatory failure, and convulsions. However, few reports, except for one from us (Kaizaki-Mistumoto et al., 2018), have measured the concentrations of the individual metabolites in patients with CFF intoxication. In addition, there have been no studies examining the efficacy of hemodialysis (HD) for CFF metabolites or the indicators that contribute to the treatment strategy for patients with acute CFF intoxication (Ou et al., 2022).
The aim of this study was to determine whether blood CFF and PXT levels can be used to predict prognosis and make treatment decisions in patients with CFF intoxication by analyzing 27 cases where information on the time after CFF ingestion was available. In addition, this study demonstrates the efficacy of HD on PXT during CFF intoxication. These observations should aid in understanding the pathogenesis of CFF intoxication and developing its treatment strategies by considering both the unchanged drug and the metabolite levels in combination.
1,2,3-Benzotriazole was purchased from Tokyo Chemical Industry Co. (Tokyo, Japan). CFF was purchased from Fujifilm Co. (Tokyo, Japan). PXT was purchased from Toronto Research Chemicals, Inc. (Toronto, Canada). Diazepam-d5 was purchased from Sigma-Aldrich Corporation (St. Louis, MO, USA). All other reagents used were of the highest grade and commercially available.
Study design and settingThis was an observational case study of 27 patients who were transported to Showa University Hospital or Showa University Fujigaoka Hospital by emergency vehicles or visited the emergency room for suspected acute CFF poisoning and whose blood levels were measured for CFF and its metabolites in the Department of Toxicology, Showa University School of Pharmacy. Ethics approval was granted by the Showa University Research Ethics Review Board (22-038-B) before the commencement of this study.
The patient data included physical examination (sex, age, and weight), patient background (diagnosis; date of admission; date of transfer; date of discharge; length of stay; length of hospitalization; outcome; medical history; medications ingested; estimated CFF intake (Dose; g); estimated time elapsed between intake and first blood draw after admission (Time; hr); and treatment and biochemistry data from admission to discharge, such as potassium (K+; mM), aspartate aminotransferase (AST; IU/L), serum alanine aminotransferase (ALT; IU/L), creatinine (Scr; mg/dL), creatine kinase (CK; IU/L), glucose concentration (GLU; mg/ dL), inorganic phosphorus (IP; mg/dL), and lactic acid (LAC; mM). Vital data included the level of consciousness measured using the Glasgow Coma Scale (GCS), diastolic blood pressure (dBP; mmHg), systolic blood pressure (sBP; mmHg), pulse rate (HR; times/min), respiratory rate (RR; times/min), oxygen saturation (SpO2; %), body temperature (BT; °C), and corrected QT interval (QTc; msec). The blood concentrations of CFF (µg/mL) and its metabolites were also measured.
Data analysisBlood data were defined as follows: hypokalemia (K < 3.4 mM), hypophosphatemia (IP < 2.5 mg/dL), hyperphosphatemia (IP > 4.5 mg/dL), hyperCKemia (CK > 200 IU/L), liver damage (ALT > 80 IU/L [twice the upper limit of normal]), hyperglycemia (GLU > 180 mg/dL), acute renal failure (Cr > 1.5 times or > 0.3 mg/dL at admission), hyperlactatemia (LAC > 2 mM), suspected lactic acidosis (LAC >5 mM) (Doi et al., 2018; Kamijo et al., 2018). Vital signs were defined as follows: decreased consciousness (GCS 3–14 points), hypertension (sBP > 140 mmHg or dBP > 90 mmHg), hypotension (sBP < 100 mmHg or dBP < 60 mmHg), tachypnea (RR > 25 beats/min), tachycardia (HR > 100 beats/min), hyperthermia (BT > 37.5°C) (Doi et al., 2018). For the electrocardiogram, QT prolongation was defined as a QTc greater than 460 msec (Schwartz and Crotti, 2011). Data not documented in the medical record were designated as n.d.
The correlation between Dose, Time, vital signs, and blood biochemical data and blood concentrations of CFF (CCFF), PXT (CPXT), and xanthine (XAN) were evaluated in 27 patients. The blood XAN concentration (CXAN) was calculated as [blood CFF concentration] + 2 × [blood PXT concentration], considering the affinity for adenosine receptors, at the first blood draw after admission. The correlation of PXT/CFF or CFF at the first blood draw after admission to Time was also evaluated. The correlation was assessed using Spearman's rank correlation coefficient (ρ). Of the 27 cases, 11 cases whose Time was within 4.5 hr, approximately equal to or shorter than the mean half-life of CFF (half-life: 3–6 hr (Mort and Kruse, 2008)), were also evaluated similarly. The odds of having hyperCKemia upon admission were compared at the different Times. The presence or absence of HD and the presence or absence of a re-elevation of CPXT after HD were analyzed using the Wilcoxon rank sum or chi-square test. A probability of significance (p) of less than 5% was considered statistically significant. Statistical analysis was performed using JMP Pro (version 16).
Measurement of CFF and its metabolitesBlood concentrations of CFF and PXT were measured either with UV-HPLC or LC-MS/MS analyses.
HPLC analysis: The serum samples (200 μL) were vigorously mixed with 250 mg of ammonium sulfate, 200 μL of 0.1% acetic acid, 2 mL of chloroform, isopropanol (85:15), and 500 ng of 1,2,3-benzotriazole as the internal standard. The lower layer was collected via centrifugation at 5,000 × g for 10 min. The solvent was evaporated under a nitrogen stream, and the residue was dissolved in 200 μL of 10 mM phosphate buffer (pH 6.8). The solution was then filtered and subjected to HPLC analysis. Calibration curves using authentic standards of CFF and PXT were prepared at ranges of 1–100 μg/mL and 0.5–50 μg/mL, respectively. When blood concentrations exceeded the calibration range, samples were diluted and extracted as appropriate.
The samples (20 μL) were analyzed using an HPLC-SPD-M20A at 273 nm and a Zorbax SB-Aq narrow bore RR column (2.1 x 100 mm, 3.5 μm; Agilent, Santa Clara, CA, USA). The column temperature was set at 40°C. The gradient elution was performed using (A) 10 mM phosphate buffer (pH 6.8) and (B) acetonitrile at a flow rate of 1 mL/min. The initial elution condition was set at 2% B for 9.5 min, increased to 6% over 3.5 min, held for 1 min, immediately returned to 2%, and then held for 7 min.
LC-MS/MS analysis: Sample extraction was performed using a Micro Volume QuEChERS Kit (Shimadzu, Kyoto, Japan). QuEChERS salt (100 mg) was mixed with 300 μL of acetonitrile, 200 μL of water, 100 μL of serum, and 2 μL of 10 ng/mL diazepam-d5 as the internal standard. The upper layer was collected via centrifugation at 15,000 x g for 10 min and analyzed using LC-MS/MS. Calibration curves using authentic standards of CFF and PXT were prepared at ranges of 0.5–75 μg/mL and 0.5–100 μg/mL, respectively. When blood concentrations exceeded the calibration range, samples were diluted and extracted as appropriate.
An LC-MS/MS system consisting of an LC-40ADXR and LCMS-8045 (Shimadzu) was used for the measurements. Chromatographic separation was achieved using a Phenomenex Kinetex XB-C18 column (2.1 mm ID × 100 mm, 2.6 μm; Shimadzu) with an equivalent Phenomenex Security Ultra C18 guard column (2.1 mm ID; Shimadzu). The column temperature was set at 40°C. The mobile phases were composed of (A) 10 mM ammonium formate and 0.1% formic acid and (B) methanol containing 10 mM ammonium formate and 0.1% formic acid. The gradient started at 5% B, increased to 35% over 15 min, increased to 95% for 7.5 min, held for 2.5 min, immediately returned to 5%, and then held for 5 min. The flow rate was set at 0.3 mL/min, and the injection volume was 1 or 5 μL. After ionization with electrospray ionization in positive mode, the samples were analyzed in multiple reaction monitoring mode. The flow rates of the nebulizer gas, the drying gas, and the heating gas were set at 3 L/min, 10 L/min, and 10 L/min, respectively. The interface temperatures, the desolvation line, and the heat block were set to 300°C, 250°C, and 400°C, respectively. The monitored ions were precursor ions of m/z 195.05 to product ions of m/z 138.05 for CFF, m/z 181.50 to 124.10 for PXT, and m/z 290.15 to 198.20 for diazepam-d5.
Of the 27 patients enrolled in this study, 81.5% (22/27) were in their 10s and 20s, 55.6% (15/27) were female, and the median weight at admission was 53.1 kg (range: 33.5–98.6). The proportion of patients with a medical history was 59.3% (16/27), of which 75% (12/16) had a psychiatric-related illness. Regarding drug intake, 70.4% (19/27) of the patients were taking CFF alone (16 over-the-counter medications and 3 supplements). The other 8 patients ingested intoxicating doses of CFF from over-the-counter drugs, which included acetaminophen (3 patients), ibuprofen (3 patients, one of whom also had over-dosed on a psychiatric prescription drug), pseudoephedrine (1 patient), and dihydrocodeine (1 patient). The median Dose was 8 g (range: 0.56–21.6) and the percentage of patients who ingested a lethal dose of CFF (> 5 g) was 63.0% (17/27). The median CCFF, CPXT, and CXAN at the first blood draw were 64.0 µg/mL (range: 14.9–385), 6.5 µg/mL (range: 2.1–30.6), and 78.4 µg/mL (range: 23.2– 397), respectively. The median Time was 4.8 hr (range: 1.7–35), and the proportion within 4.5 hr was 40.7% (11/27). The median hospital stay length was three days (range: 1–18), and all the patients were discharged or transferred after hospitalization.
Regarding the blood data, hyperCKemia was demonstrated in 40.7% (11/27) of the patients upon admission and 63.0% (17/27) during their hospitalization. In 48.1% (13/27) of the patients, hyperglycemia was shown upon admission, with no new cases during their hospitalization. Hypokalemia was demonstrated in 70.4% (19/27) of patients upon admission and 77.8% (21/27) during their hospitalization. Hypophosphatemia and hyperphosphatemia were shown in 31.8% (7/22) and 4.5% (1/22) of the patients, respectively, upon admission. During their hospitalization, hypophosphatemia was observed in 60.0% (15/25) of patients, and there were no new cases of hyperphosphatemia. Hyperlactatemia was demonstrated in 84.0% (21/25) of the patients upon admission, and 52.4% (11/21) were suspected of lactic acidosis. No new cases of hyperlactatemia or suspected lactic acidosis occurred during their hospitalization. Liver injury was demonstrated in 7.4% (2/27) of the patients upon admission, and increasing to 18.5% (5/27) during their hospitalization. No patient demonstrated acute kidney injury upon admission or developed it during hospitalization. QT prolongation was shown in 7.4% (2/27) of the patients upon admission, and increasing to 18.5% (5/27) during their hospitalization.
Regarding the vital signs on admission, 40.7% (11/27) were unconscious, 37.0% (10/27) were tachypnea, 63.0% (17/27) were tachycardia, 12.5% (3/24) were hyperthermia, 22.2% (6/27) were hypertension, and 14.8% (4/27) were hypotension. Regarding the medical treatments, HD or CHDF was performed in 63.0% (17/27, of which 15 were HD and 2 were CHDF), 11.1% (3/27) were intubated, 40.7% (11/27) had a gastric lavage, 66.7% (18/27) were on activated charcoal and laxative medications, 63.0% (17/27) were on anti-nausea medication, 51.9% (14/27) were administered a PPI, 18.5% (5/27) were on antimicrobials for aspiration pneumonia, 51.9% (14/27) had K correction, 11.1% (3/27) had Mg2+ correction, and 11.1% (3/27) had IP correction. In addition, two patients received benzodiazepines for restlessness, one patient received N-acetylcysteine to detoxify a high dose of an acetaminophen-containing cold remedy (5 g acetaminophen), one patient received acetaminophen for fevers, and two patients received a vasopressor. In addition, electrical defibrillation and assisted circulatory devices were used in one case.
Interrelationships among the data obtainedThe correlations between Dose, Time, blood drug concentrations, blood data, and vital signs in the 27 cases were analyzed (Table 1). Dose was positively correlated with CCFF (ρ = 0.60, p < 0.01; Fig. 1A and Table 1), CPXT (ρ = 0.44, p < 0.05; Fig. 1B and Table 1), CXAN (ρ = 0.72, p < 0.01; Table 1), and blood LAC (ρ = 0.44, p < 0.05; Table 1) and negatively correlate with BT (ρ = -0.41, p < 0.05; Table 1). Time was positively correlated with CPXT (ρ = 0.61, p < 0.01; Table 1), AST (ρ = 0.50, p < 0.01; Table 1), CK (ρ = 0.50, p < 0.01; Table 1), and Cr (ρ = 0.52, p < 0.01; Table 1) and negatively correlated with GLU (ρ = -0.42, p < 0.05; Table 1).
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/49_447-t1.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
The correlations between Dose and CCFF or CPXT. Correlations between Dose and CCFF of all patients (A; n = 27), Dose and CPXT of all patients (B; n = 27), Dose and CCFF of patients whose Time was less than 4.5 hr (C; n = 11), Dose and CPXT of patients whose Time was less than 4.5 hr (D; n = 11) and Dose and CPXT of patients whose Time was 4.5 hr or more (E; n = 16) are illustrated. A dot indicates CCFF or CPXT for Dose of each patient, and the straight line shows the correlation in the case where a significant correlation was observed.
When the relationship with blood drug concentrations was examined, CCFF was positively correlated with CXAN (ρ = 0.98, p < 0.01), GLU (ρ= 0.50, p < 0.01), IP (ρ = 0.53, p < 0.05), and LAC (ρ = 0.75, p < 0.01) and negatively correlated with K (ρ = -0.68, p < 0.01) and GCS (ρ = -0.53, p < 0.01). CPXT was positively correlated with AST (ρ = 0.46, p < 0.05), CK (ρ = 0.68, p < 0.01), Cr (ρ = 0.49, p < 0.05), and IP (ρ = 0.57, p < 0.01), and none showed a negative correlation. CXAN were positively correlated with GLU (ρ = 0.41, p < 0.05), IP (ρ = 0.54, p < 0.05), and LAC (ρ = 0.73, p < 0.01)-and negatively correlated with K (ρ = -0.62, p < 0.01), sBP (ρ = -0.40, p < 0.05), and GCS (ρ = -0.41, p < 0.01). CXAN had a significant correlation with sBP, which was not seen with CCFF and had smaller p-values for the other blood biochemical data compared with CCFF (Table 1).
SubanalysisNext, an analysis was conducted on the 11 cases where the Time was less than 4.5 hr (Table 2). The correlation coefficient between Dose and CCFF (ρ = 0.74, p < 0.01; Fig. 1C) was found to be higher than that for all 27 cases (ρ = 0.60, p < 0.01; Table 1), despite the smaller number of cases. Thus, it is shown that Dose and CCFF have a strong correlation when the Time is shorter than the CFF half-life (about 4.5 hr). Meanwhile, no correlation was observed between Dose and CPXT when the Time was less than 4.5 hr (Fig. 1D), and a significant correlation was obtained in 16 cases where the Time was longer than 4.5 hr (ρ = 0.58, p < 0.05; Fig. 1E). Thus, these results suggest that Dose and CPXT correlate when the Time is longer than the CFF half-life. The correlations between Dose, Time, and blood drug concentration with blood biochemistry data, vital signs, SpO2, and QTc in the 11 cases where the Time was less than 4.5 hr showed that Dose was positively correlated with IP (ρ = 0.69, p < 0.05) and negatively correlated with K (ρ = -0.77, p < 0.01), sBP (ρ = -0.90, p < 0.01), dBP (ρ = -0.79, p < 0.01), and GCS (ρ = -0.61, p < 0.05), all of which were not significant in the 27 cases (Table 2). The blood CCFF and CXAN showed similar trends in the correlation between the blood biochemical data and vital signs, correlating positively with GLU (ρ = 0.72, p < 0.05 and ρ = 0.69, p < 0.05) and LAC (ρ = 0.90, p < 0.01 and ρ = 0.88, p < 0.01) and negatively with K (ρ = -0.82, p < 0.01 and ρ = -0.78, p < 0.01), sBP (ρ = -0.72, p < 0.05 and ρ = 0.80, p < 0.01), dBP (ρ = -0.66, p < 0.05 and ρ = -0.66, p < 0.03), and GCS (ρ = -0.84, p < 0.01 and ρ = -0.84, p < 0.01). CPXT was positively correlated with IP (ρ = 0.68, p < 0.05). There was no correlation with Time in this analysis (Table 2).
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/49_447-t2.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
Of the 11 patients with hyperCKemia (≥ 200 IU/L) at admission, 8 had a Time of 10 hr or more, and of the 16 patients with a normal CK at admission, only 1 had a Time of 10 hr or more. The odds ratio of hyperCKemia after 10 hr was 40 (95% CI: 3.56–450; Table 3). Therefore, these results demonstrated that patients with a Time of 10 hr or more are at risk for hyperCKemia.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/49_447-t3.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
The ratio of CPXT and CCFF at the time of the first blood draw after admission was calculated and examined for a correlation with Time. The analysis revealed a positive correlation (ρ = 0.58, p < 0.01; Fig. 2A). Meanwhile, there was no correlation between CCFF and Time (Fig. 2B). Therefore, these results suggest that Time can be estimated by the ratio of CPXT and CCFF at the time of the first blood draw after admission.
The correlations between Time and the CPXT/CCFF ratio or CCFF (n = 27). Correlation between Time and the CPXT/CCFF ratio (A) and Time and CCFF (B) are illustrated. A dot indicates the CPXT/CCFF ratio or CCFF for Dose of each patient, and the straight line shows the correlation in the case where a significant correlation was observed.
The effect of HD was analyzed using 23 cases where 2 CHDF cases and 2 cases where the drug concentrations before and after HD were missing were excluded. Dose (p < 0.01), blood CFF (p < 0.05), blood XAN (p < 0.01), and PPI administration (p < 0.01) were significantly higher in the HD group (13 patients) than the without HD group (10 patients). Meanwhile, the HD group had significantly fewer patients (p < 0.05) where medicinal ingredients other than CFF were detected (Table 4). In 13 patients who underwent HD, the CCFF immediately before HD exceeded the toxic range (CCFF ≥ 25 µg/mL) and exceeded the lethal range (CCFF ≥ 80 µg/mL) in 6 cases. The decreased rate in CCFF due to HD ranged from 33.2% to 77.6% (median 58.7%; Figs. 3A and 3B), and the reduced rate in CPXT ranged from -15.9% to 58.8% (median 32.1%; Fig. 3C and 3D). In 6 patients whose CPXT immediately before HD was higher than 10 µg/mL, the decrease in PXT ranged from 32.1% to 56.9% (median 49.7%; white circles in Fig. 3D).
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/49_447-t4.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
Effect of HD on the elimination of CFF and PXT (n = 13). Changes in CCFF (A) and CPXT (C) before and after HD of each patient are illustrated. The gray dashed lines in (A) and (C) indicate concentration trends for the same patient. The correlations between the decreased rate of CCFF by HD and CCFF immediately before HD (B) and between the decreased rate of CPXT by HD and CPXT immediately before HD (D) are illustrated. The white and black dots in (D) indicate patients with CPXT immediately before HD over 10 μg/mL and 10 μg/mL or less, respectively.
The metabolism and excretion of CFF are prolonged during CFF intoxication; therefore, the CCFF and the CPXT may become higher than expected, producing a strong pharmacological effect. However, there are no reports that describe CFF metabolites during CFF intoxication other than the one from our laboratory (Kaizaki-Mistumoto et al., 2018). In addition, to the best of our knowledge, there are no reports on the efficacy of HD for CFF metabolites, although CFF has been reported to be effectively removed by HD. In recent years, energy drinks and supplements containing CFF have become widely available at Internet shops and community pharmacies. According to a report by Kamijo et al. (Kamijo et al., 2018), the number of CFF poisonings among young people has been increasing since 2013, and this trend was confirmed by the current study in which 82.1% of the cases were in their 10s and 20s. Despite such an increasing trend in acute CFF poisoning, there is no indicator that contributes to the treatment policy for patients with acute CFF poisoning.
The present study found that when the Time was less than 4.5 hr, the average half-life of CFF, Dose strongly correlated with CCFF but not with CPXT. The reason why Dose and PXT do not correlate with each other may be due to the individual differences in factors related to CFF metabolism, such as body weight, age, sex, and renal and hepatic functions (Tian et al., 2019), as well as the differences in the rate of gastric emptying due to eating conditions at the time of CFF intake. Other factors that may affect the rate of CFF metabolism include daily oral medications, alcohol consumption, smoking, and other habits and environments (Cornelis et al., 2007; Rodenburg et al., 2012). Another possibility is that excessive CFF intake may prolong absorption and metabolism time. In addition, since the Time is based on patient and family reports, there may be some deviation from the actual time, which may be relevant.
The ratios of CPXT and CCFF at the time of the first blood draw after admission were strongly correlated with Time, suggesting that it is possible to estimate Time, which is important for the treatment and is sometimes uncertain at the time of transport, using these concentrations. In addition, CXAN correlates more strongly with symptoms and the vitals at the time of admission than CFF, suggesting that CXAN is a useful marker for understanding the intoxication symptoms. Since Dose, in addition to Time, is often uncertain at the time of transporting patients with drug poisoning and there are no known markers, the treatment plan depends on the facility and judgment of the physician in charge. The significantly higher PPI administration among patients with HD shown in this study could be due to institutional criteria. If CCFF and CPXT could be rapidly measured, it would be possible to estimate Time, which would be useful in determining the treatment strategy.
This study showed that the odds ratio of developing hyperCKemia was very high if the Time was 10 hr or more compared with less than 10 hr. This finding indicates that the risk of developing hyperCKemia increases the longer the patient is untreated after ingestion. Therefore, when a patient with acute CFF intoxication with a Time of 10 hr or more is transported to the hospital, the possibility of developing hyperCKemia should be taken into consideration before starting the treatment. In addition, the risk of hyperCKemia due to a delayed start of treatment in CFF-intoxicated patients should be determined. Although the detailed mechanism of the development of hyperCKemia due to acute CFF intoxication is not known to date, it could be due to the effects of CFF and its metabolites on rhabdomyocytes (Chiang et al., 2014). Excessive doses of CFF may enhance nonselective phosphodiesterase (PDE) inhibition and ryanodine receptor (RyR) sensitivity, as well as A2A receptor antagonism (Pesta et al., 2013). In rhabdomyocytes, RyR on the sarcoplasmic reticulum membrane functions as the main Ca2+ release channel into the cytoplasm, and high concentrations of CFF could cause RyR to remain open for a long time, leading to an increase in cytosolic Ca2+ concentration. This may cause excessive muscle contractions and convulsions, leading to hyperCKemia. Nonselective PDE inhibition may also increase intracellular cAMP concentration, which contributes to the increase in intracellular Ca2+ concentration. Although this study did not examine blood Ca2+ concentrations, it will be necessary to accumulate data on Ca2+ in future cases of acute CFF poisoning. In the five patients who developed QT prolongation syndrome at the time of admission and during their hospitalization, the only common feature among the patients was hyperCKemia, with some cases having high CK levels exceeding 10,000 IU/L. A marked elevation of CK is a predictor of fatal arrhythmias, such as ventricular tachycardia and ventricular fibrillation (Oikawa et al., 2019; Yang et al., 2003), and patients with hyperCKemia at admission should be cautioned against the occurrence of arrhythmias. Although there were no cases in which acute kidney injury was suspected due to the short duration of hospitalization in this study, ongoing attention is needed due to the possibility of acute kidney injury secondary to hyperCKemia (Młynarska et al., 2022; Mosquera et al., 2013).
This study, as previously reported (Willson, 2018; Gahona et al., 2022), confirmed that HD during CFF intoxication is effective in reducing blood CFF levels and found that 3–4 hr of HD reduces blood CFF by 33.2%–77.6% (median 58.7%). Meanwhile, the reduction rate of PXT by HD ranged from -15.9%–58.8% (median 32.1%), which varied greatly depending on the CPXT in the blood immediately before HD. This may be because HD is based on the principle of passive diffusion, and the lower the concentration, the lower the rate of drug reduction. In one patient with a relatively low CPXT, the CPXT immediately after HD was higher (5.8 µg/mL) than before HD (5.0 µg/mL), possibly due to the metabolism of the remaining CFF. When the CPXT immediately before HD was relatively high (≥ 10 µg/mL), approximately 50% of the PXT was removed from the blood, about the same efficiency as CFF. These findings suggest that the efficacy of HD for PXT, which has not been previously reported, is as effective as that of CFF in cases with CPXT above a certain level. PXT has similar characteristics to CFF in terms of low molecular weight (180 g/mol), low protein binding rate, and low distribution volume; therefore, it is expected to have similar pharmacokinetics as CFF (Lelo et al., 1986), which may make HD effective (Kaizaki-Mistumoto et al., 2018), although no direct evidence is available (Lelo et al., 1986). The present study demonstrated that HD was less effective than CFF for PXT, which has a lower blood concentration than CFF. It also showed that PXT was as effective as CFF at high concentrations.
It should be noted that in the seven patients whose CPXT increased 6 hr after HD, the CCFF immediately after HD was significantly higher than in the six patients whose CPXT did not increase after HD (CCFF: 49.5 ± 6.0 μg/mL vs. 14.9 ± 6.5 μg/mL, p < 0.05 [Wilcoxon rank sum]). In addition, in the seven patients whose PXT levels increased, hyperCKemia was observed in 85.7% (6/7) of the patients, which persisted for at least 6 hours after HD. In acute CFF intoxication, even if HD reduces the CCFF to a lower level of the intoxication range, CXAN may not decrease as much due to the subsequent increase in CPXT from residual CFF. Therefore, patients should be carefully monitored for possible prolongation of CFF intoxication symptoms after HD.
In summary, this study shows that in acute CFF poisoning, Time can be estimated by measuring CCFF and CPXT, and CXAN calculated from these concentrations can be a useful indicator for understanding intoxication symptoms. Currently, there are no clear clinical markers that contribute to the treatment guidelines or treatment strategies for acute CFF poisoning, and treatment strategies vary. Therefore, the measurement of PXT, in addition to CFF, will provide useful information for treatment and is necessary to establish a rapid and simple measurement system for CFF and PXT in the future. This study also reveals that hyperCKemia may develop in cases left untreated for a prolonged period after ingestion. The results of this study are expected to lead to the prompt treatment of CFF intoxication when the patient arrives at the hospital. Furthermore, even if CCFF is reduced by HD, the production of PXT, which is a stronger A2A receptor antagonist than CFF, may prolong the symptoms of CFF intoxication.
Conflict of interestThe authors declare that there is no conflict of interest.
