Electrochemical Point‐of‐Care Test for Assessing Serum Paracetamol Concentration: Comparison With Traditional Methods and Detection of Concomitant Drugs

ABSTRACT

Rapid serum paracetamol (acetaminophen) concentration measurement is essential in suspected intoxication, but centralized laboratory analyses often delay initiation of antidotal therapy. We studied the feasibility of a novel electrochemical single‐walled carbon nanotube/Nafion‐based point‐of‐care (POC) method in detecting paracetamol in 99 suspected overdose patient serum samples. POC was compared with the standard photoelectric enzymatic method (PEM) and ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector (UHPLC‐DAD‐CAD). We also analysed for 900 concomitant pharmaceuticals, drugs and chemicals in 197 samples with time‐of‐flight mass spectrometry to assess interference with paracetamol concentration measurements. Paracetamol concentrations measured with UHPLC‐DAD‐CAD ranged between 0 and 2100 μmol/L, with 19% above the therapeutic level (≥ 200 μmol/L). Comparing POC with UHPLC‐DAD‐CAD, the false positives and negatives were 10% and 15%, respectively, at concentrations ≥ 30 μmol/L. All POC method false negatives occurred at concentrations < 45 μmol/L. PEM showed 8% false positives and negatives compared with UHPLC‐DAD‐CAD. Other substances detected included caffeine (78%), antidepressants (41%), benzodiazepines (34%) and antipsychotics (28%). They did not interfere with POC concentration measurement. The novel POC method is promising for measuring serum paracetamol at relevant concentrations.

Plain Language Summary

Rapid measurement of serum paracetamol concentration is essential in suspected poisoning. We compared a novel fast electrochemical point‐of‐care (POC) method with two established, more time‐consuming laboratory methods for detecting paracetamol in samples from 99 suspected overdose patients. The sensitivity of POC was 87% and the specificity was 91% compared with the ‘gold standard’ mass spectrometry. All POC false negatives occurred at low concentrations, suggesting a good safety profile. We also analysed for 900 concomitant substances from the same patient group. The most common other substances were caffeine, antidepressants, benzodiazepines and antipsychotics. The POC is promising in the rapid measurement of paracetamol concentration.

1. Introduction

Paracetamol (acetaminophen) toxicity is a common cause of liver failure. Prevention of toxicity by antidote therapy necessitates measurement of paracetamol concentration [1]. The standard methods of paracetamol concentration analysis in venous plasma are laborious and time‐consuming, with the average delay from sampling to results in our tertiary hospital being 60–120 min. To expedite diagnostics, we recently introduced a novel electrochemical single‐walled carbon nanotube/Nafion‐based point‐of‐care (POC) paracetamol concentration analysis tool for rapid concentration analysis from finger‐prick capillary samples [2, 3, 4].

Paracetamol toxicity can be prevented with antidotes, such as an N‐acetylcysteine (NAC) infusion. In addition to NAC, several other interventions such as fomepizole and various antioxidants have been suggested as potential candidates to prevent paracetamol toxicity [5, 6]. Delays in concentration analysis are detrimental because the timing of NAC administration is important in minimizing liver and renal damage [7]. On the other hand, avoiding unnecessary NAC administration is preferable, as it can have adverse effects [8, 9]. Furthermore, the dose of the antidote is under debate. It can be adjusted according to the paracetamol concentration, and precautionary administration without concentration analysis may not be justified [1, 10, 11, 12]. A POC analysis of paracetamol concentrations enables rapid diagnosis and earlier start of treatment, even prehospital admission, reducing delays and enabling patient transfer directly to the appropriate health facility. In addition, rapid exclusion of paracetamol intoxication as a cause of illness can prevent unnecessary antidote treatments and patient transfers to hospitals, thus reducing healthcare costs. There is an obvious need for a fast concentration analysis method that allows rapid analysis of multiple, consecutive samples.

The POC method uses a carbon nanotube‐based electrochemical sensor to detect paracetamol in patient samples. Electrochemical methods have multiple advantages for POC measuring. Detection is fast, does not require additional sample preparation, has portable and relatively simple instrumentation and requires only a small sample size. The sensor is coated with a polymer‐based filter layer to exclude larger molecules, such as blood cells and proteins, and other unwanted interferents. Paracetamol is an electrochemically active molecule and thus readily detectable with direct electrochemical redox reactions on an electrode surface. In the POC, differential pulse voltammetry (DPV) is used as the electrochemical method, where oxidation and/or reduction currents measured by an electrochemical test strip correlate with the paracetamol content in the sample.

In this study, we compared the paracetamol concentrations measured with the novel POC method [4] with the results of two established methods, the standard photometric enzymatic method (PEM) and ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector (UHPLC‐DAD‐CAD) method for 99 samples. We also conducted an additional analysis of 197 samples including the samples used in the POC analysis to assess which drugs and substances are found in real‐life patients with known or suspected paracetamol intoxication and their possible interactions with paracetamol concentration analyses. Because paracetamol is frequently present in mixed intoxications [13, 14], there is a high demand for concentration analysis methods to control for the effects of confounding drugs. For qualitative analyses of concomitant substances, we used UHPLC‐high resolution quadrupole time‐of‐flight mass spectrometry (UHPLC‐HR‐QTOF‐MS) in 197 anonymized serum samples from patients with suspected paracetamol overdose. These qualitative analyses included screening of about 900 conventional drugs, pharmaceuticals and novel psychoactive substances.

2. Methods

We collected 197 anonymized serum samples sent to the HUS Diagnostic Centre's Department of Clinical Chemistry (Helsinki and Uusimaa Hospital District Laboratory, Helsinki, Finland) for paracetamol concentration analysis with PEM from patients with suspected paracetamol overdose. The study was conducted in accordance with the Basic & Clinical Pharmacology & Toxicology policy for experimental and clinical studies [15], and it received a research permit from Helsinki and Uusimaa Hospital District. A statement from the ethics committee was not needed because this was a comparative study of laboratory methods using anonymized samples. The samples originated from secondary and tertiary hospitals in the Helsinki and Uusimaa Hospital District and were collected, regardless of patient demographics, between August 2020 and February 2021. Systematic collection of consecutive specimens could not be achieved for logistic reasons. After standard analysis with PEM, samples were divided into two aliquots, stored at −80°C for later quantitative analysis of paracetamol with POC and UHPLC‐DAD‐CAD methods and qualitative analysis of concomitant substances with UHPLC‐HR‐QTOF‐MS, at the Finnish Institute for Health and Welfare Forensic Toxicology Laboratory, Helsinki, Finland. The POC method analyses were performed between February and June 2021 and the UHPLC‐DAD‐CAD analyses in November 2021. Thus, the storage time at −80°C varied between 6 and 15 months. There was one freeze–thaw cycle, as the specimens were divided into small aliquots in the first phase. The paracetamol concentration in the specimens was considered stable with deviation ≤ 15% during storage and one freeze–thaw cycle, as it is considered stable during at least three freeze–thaw cycles [16, 17].

The PEM method uses the Atellica CH Acetaminophen (Acet) assay and Siemens Atellica Solution CH 930 analyser (Siemens Healthineers) for performing colorimetric, enzymatic paracetamol analysis. The assay was calibrated with Atellica CH TOX CAL (Siemens Healthineers) and three‐level quality assurance samples (Liquichek Therapeutic Drug Monitoring Control [TDM] Levels 1, 2 and 3 [Bio‐Rad Laboratories, USA]) were used as internal controls. Concentrations of paracetamol in these controls were 128, 339 and 657 μmol/L, as provided by the manufacturer. CV% of these controls was in practice 4.3%, 3.1% and 3.6%, respectively. In the Atellica CH Acetaminophen (Acet) assay, paracetamol is converted to 4‐aminophenol by acyl amidohydrolase. This then reacts with 8‐hydroxyquinoline‐5‐sulfonic acid in the presence of manganese ions, and a coloured complex of 5‐(4‐iminophenol)‐8‐quinoline is formed. The increased absorbance is directly proportional to the concentration of paracetamol in the sample [18]. The limit of detection and limit of quantification for the Atellica CH Acet assay is 13.2 μmol/L.

For comparison of the different methods, the ‘gold standard’ quantitative analysis of paracetamol was based on UHPLC‐DAD‐CAD. In short, serum samples (1 mL) were extracted in basic conditions with an ethyl acetate‐butyl acetate mixture, and the components were separated and quantified with reversed phase UHPLC‐DAD‐CAD. The limit of quantification for paracetamol was 13.2 μmol/L and the limit of detection was 6.6 μmol/L. The external quality control of this method is covered by participating in LGC TOX AXIO PT‐TX‐BLD proficiency testing monthly. The method has been described in detail by Viinamäki et al. [19].

In the screening of concomitant substances from samples, qualitative analysis was based on UHPLC‐HR‐QTOF‐MS. In short, serum samples (0.5 mL) were extracted using mixed‐mode solid‐phase extraction, the components were separated with reversed‐phase UHPLC and identified with HR‐QTOF‐MS. The in‐house database contained more than 900 therapeutic drugs, drugs of abuse and new psychoactive substances. The identification criteria included mass accuracy (±3 mDa), retention time (±0.3 min), and isotopic pattern and signal matching for both precursor ion and qualifier ion(s). The method has been described in detail by Sundström et al. and Ojanperä et al. [20, 21]. The list of the included substances is available from the corresponding author upon request.

The first 99 samples were analysed with the electrochemical POC sensor, a novel concentration measurement method under development. Because the POC test strip was evolving rapidly at the time of this study, we did not have enough consistent test strips to run all 197 samples with POC. After thawing the patient sample, it was divided into a minimum of nine 20 μL aliquots and diluted with 20 μL 10 mmol/L phosphate‐buffered saline (PBS) solution (pH 7.4), spiked with twice the target concentration of paracetamol. Internal calibration with additions of 0, 50, 250, 500 and 1000 μmol/L was carried out for each sample to form calibration curves (slope CV = 9.9%, n = 99). External calibrations were also carried out with spiked PBS.

The electrochemical measurements were carried out with DPV, using a portable PalmSens4 potentiostat (PalmSens BV, Netherlands) connected to a personal computer. The 40 μL sample was transferred onto the test strip, and the measurement script was run. The script included an automated incubation time of 2.5 min followed by a 30 s DPV scan. Each electrode strip was measured three times with blank PBS to characterize and stabilize the background, after which the spiking series was measured. The electrode was washed with PBS for 2.5 min between each measurement. Each electrode was also measured with 500 μmol/L paracetamol after the concentration series to verify the absence of electrode fouling.

2.1. Statistics

The false positive and false negative rates of POC and PEM methods were reported as compared with UHPLC‐DAD‐CAD, which was considered the ‘gold standard’ method. Bland–Altman plots were used for methods comparison with a user‐written ‘blandaltman’ command in Stata Version 15.1/MP2 (StataCorp Station, TX, USA) [22]. Different drugs detected in the qualitative screening were sorted by Anatomical Therapeutic Chemical (ATC) code for further analysis.

3. Results

To compare different methods of analysis, the first 99 samples were analysed using the POC, PEM and UHPLC‐DAD‐CAD methods. The comparison of methods was conducted with paracetamol concentrations ≥ 30 μmol/L. The risk of paracetamol toxicity is determined using the Rumack–Matthew nomogram for paracetamol plasma concentrations. Concentrations below 30 μmol/L were considered of little clinical relevance because the Rumack–Matthew nomogram limits for potential toxicity are above this within 4–24 h after paracetamol ingestion [23]. Quantitative paracetamol analysis with UHPLC‐DAD‐CAD was not feasible for 23 samples due to insufficient amounts. One sample (ID 3) had a negative result in the PEM analysis, while the concentration of the same sample was 475 μmol/L measured with the POC. The amount of the serum sample was not sufficient for quantitative analysis with UHPLC‐DAD‐CAD. From the remaining POC sample, we were able to measure paracetamol concentration with high‐performance liquid chromatography tandem mass spectrometry (HPLC‐MS/MS) by dilution of the sample, resulting in a concentration of 488 μmol/L. Because of the different analysis methods used, ID 3 is not included in the comparison of the PEM with UHPLC‐DAD‐CAD. This exclusion causes an error in the false negative rate reporting. Individual paracetamol concentrations measured with different methods are presented in Table S1.

All 197 samples were analysed for concomitant pharmaceuticals, drugs and chemicals. Of the 197 samples, 37% (n = 73) had paracetamol concentrations of ≥ 30 μmol/L. Concentrations ranged from 0 to 2100 μmol/L, with 19% (n = 38) of samples having concentrations ≥ 200 μmol/L, indicating obvious paracetamol overdoses.

A flowchart of different analyses is presented in Figure 1.

FIGURE 1.

A flowchart presenting the different methods of analysis. POC (point‐of‐care method under development), PEM (photometric enzymatic method), UHPLC‐DAD‐CAD (ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector); n = number of results.

3.1. Comparison of POC and UHPLC‐DAD‐CAD

Comparison of POC with UHPLC‐DAD‐CAD in 76 samples revealed a POC false positive rate of 10% and a false negative rate of 15%, giving POC sensitivity of 87% and specificity of 91%, and positive predictive value of 65% and negative predictive value of 97%. All false negative POC samples were detected at concentrations below 45 μmol/L as measured by the UHPLC‐DAD‐CAD. The results are shown in Table 1.

TABLE 1.

Comparison of POC (electrochemical point‐of‐care method) with UHPLC‐DAD‐CAD (ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector) (control) for detecting paracetamol with a threshold of ≥ 30 μmol/L. n = number of results.

	UHPLC‐DAD‐CAD negative n (%)	UHPLC‐DAD‐CAD positive n (%)	Total n (%)
POC negative	57 (97)	2 (3) a	59 (100)
POC positive	6 (35)	11 (65)	17 (100)
Total	63 (83)	13 (17)	76 (100)

^a Concentrations measured with POC were 22 and 28 μM, and those measured with UHPLC‐DAD‐CAD were 31 and 43 μM, respectively.

For visualization of the comparison between POC and UHPLC‐DAD‐CAD methods, we performed a Bland–Altman plot (see Figure 2).

FIGURE 2.

A Bland–Altman plot showing good agreement between POC (electrochemical point‐of‐care method) and UHPLC‐DAD‐CAD (ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector) for paracetamol concentrations ≥ 30 μmol/L in UHPLC‐DAD‐CAD (n = 13).

3.2. Comparison of PEM and UHPLC‐DAD‐CAD

At paracetamol concentrations > 30 μmol/L, comparison of PEM with UHPLC‐DAD‐CAD showed a false positive rate of 8% and a false negative rate of 0%, with a positive predictive value of 72% and a negative predictive value of 100%. Sample ID 3 was analysed with HPLC‐MS/MS, as described above. When it was included, the percentage of false negative results for PEM compared with the reference method was 8%, resulting in PEM having both sensitivity and specificity of 93%. Comparison of PEM with the UHPLC‐DAD‐CAD is presented in Table 2.

TABLE 2.

Comparison of PEM (photometric enzymatic method) with UHPLC‐DAD‐CAD (ultra‐high performance liquid chromatography–photodiode array and corona‐charged aerosol detector) (control) for detecting paracetamol with a threshold of ≥ 30 μmol/L; n = number of results.

	UHPLC‐DAD‐CAD negative n (%)	UHPLC‐DAD‐CAD positive n (%)	Total n (%)
PEM negative	58 (100)	0 (0)	58 (100)
PEM positive	5 (28)	13 (72)	18 (100)
Total	63 (83)	13 (17)	76 (100)

3.3. Comparison of POC and PEM

Finally, we compared the novel POC method with PEM in 99 samples. The POC method gave false positives in 11% and false negatives in 22% of measured samples. All discrepancies occurred at paracetamol concentrations below 45 μmol/L, except for sample ID 3 (475 and 488 μmol/L in POC and HPLC‐MS/MS analysis, respectively, and a negative result with PEM). The results of the comparisons are shown in Table 3.

TABLE 3.

Comparison of POC with PEM (control) for detecting paracetamol with a threshold of > 30 μmol/L; n = number of results.

	PEM negative n (%)	PEM positive n (%)	Total n (%)
POC negative	56 (88)	8 (12)	64 (100)
POC positive	7 (20)	28 (80)	35 (100)
Total	63 (64)	36 (36)	99 (100)

3.4. Concomitant Substances

The most common substances other than paracetamol were caffeine in 78% (n = 153), antidepressants 41% (n = 81), benzodiazepines 34% (n = 66) and antipsychotics in 28% (n = 56) of samples. Cotinine, a metabolite of nicotine, was found in 34% (n = 66) of samples; illicit drugs such as amphetamine or its derivatives were infrequent (n = 2). Gabapentinoids, frequently present in mixed intoxications, were detected in only one sample. The numbers of drugs and substances detected in the samples are shown in Figure 3 and Table 4. The list of the substances found sorted by Anatomical Therapeutic Chemical (ATC) code is shown in Table S4.

FIGURE 3.

Number of drugs (excluding caffeine and tobacco) in 197 serum samples.

TABLE 4.

Substances, other than paracetamol, sorted by ATC codes, in 197 serum samples, from most to least frequent.

Substance	n (%)
Caffeine	154 (78)
Antidepressants	81 (41)
Benzodiazepines or other sedatives	66 (33)
Cotinine (metabolite of nicotine)	66 (33)
Antipsychotics	56 (28)
Antiarrhythmics	44 (22)
Non‐steroidal anti‐inflammatory drugs	42 (21)
Local anaesthetics	36 (18)
Opioids	32 (16)
Antiemetics	18 (9)
Antiepileptics	13 (7)
Anaesthetics	13 (7)
Muscle relaxants	11 (6)
Antibiotics	6 (3)
Melatonin	5 (3)
Oral anticoagulants	5 (3)
Cannabinoids	4 (2)
Oral diabetes medications	3 (2)
Quinine	2 (1)
Amphetamine and its derivatives	2 (1)
Antihypertensives	1 (0.5)
Antidementia drugs	1 (0.5)
Gabapentin	1 (0.5)
Anticholinergics	1 (0.5)
Mucolytics	1 (0.5)

In the qualitative analysis with UHPLC‐HR‐QTOF‐MS, 96% (n = 189) of the samples included concomitant drugs, cotinine and/or caffeine. Excluding cotinine and caffeine, 85% (n = 168) of samples presented with additional drugs. Substances present in the samples where POC showed a false positive or a false negative concentration value compared with measurements with UHPLC‐DAD‐CAD are presented in Tables S2 and S3.

4. Discussion

We compared paracetamol concentrations measured with a recently introduced electrochemical POC method with the standard photometric enzymatic method and the ‘gold standard’ liquid chromatography method in 76 samples. The electrochemical POC method had a sensitivity of 87% and a specificity of 91% compared with the ‘gold standard’ UHPLC‐DAD‐CAD when used to screen for paracetamol at a cut‐off of 30 μmol/L. In comparison, the sensitivity and specificity of the standard photometric method were 93%, indicating that both methods are subject to an acceptable level of error. In addition, we used time‐of‐flight mass spectrometry to identify concomitant substances from 197 anonymized samples from patients with known or suspected paracetamol intoxication to find out what substances were present and their possible interactions with paracetamol measurements.

In the qualitative screening analysis, the most common co‐detected drugs were antidepressants, benzodiazepines and/or antipsychotics, in 63% (n = 125) of the analysed samples, indicating frequent use of psychoactive drugs and possibly also high psychiatric comorbidity in this group of patients. This is in line with the study by Piotrowska et al., who reported psychiatric comorbidities in 78% of the paracetamol intoxication population [24]. In that study, other substances were taken together with paracetamol in 50% of cases, with psychotropic medications and analgesics being the most common. The study by Piotrowska et al. analysed samples from patients ≥ 16 years of age with actual paracetamol intoxication, while in our study, samples were taken from patients with suspected intoxication, regardless of age. In a toxicological postmortem analysis in a paediatric population by Swatek et al. [14], the most prevalent drug detected was diphenhydramine (35%), followed by acetaminophen (34%) and dextro/levomethorphan (17%). Only one drug class was reported in 40% of paediatric cases, including pain relievers in 16%, antihistamines in 16% and symptomatic cold/flu medications in 6%. Thus, in 60% of the cases, a combination of two or more drug classes was reported. In our study, paracetamol was the sole detected agent in only 15% (n = 29) of the samples, and 40% (n = 79) of the samples involved five drugs or more. Caffeine was involved in most of the samples, as Finns are avid coffee consumers. The high frequency of other drugs in the serum samples of suspected paracetamol intoxication cases suggests that taking into account interactions with other drugs is highly important in POC method development.

The method of concentration analysis must withstand interference from various drugs and patient‐derived features. Each method has its sources of bias that are often recognized, but others may still be unknown. For example, bilirubin is known to cause positive and lipaemia negative bias in PEM [18]. Interference of bilirubin is remarkable, as the possible liver injury caused by paracetamol intoxication raises bilirubin to levels high enough to interfere with the PEM method. This upward bias may complicate clinical judgement. Electrochemical methods, such as that used in the POC, are also susceptible to interference from various sources. However, these interferences can be overcome by adjusting the properties of the sensors. For example, caffeine causes no interference with the POC method used in this study [3]. Instead, patient‐derived endogenous interferents caused a marked upward bias at low concentrations (< 20 μmol/L) in our previous study with the same POC in healthy volunteers [4]. The POC method was found more accurate at higher therapeutic and toxic concentrations where the significance of these interferents is minor [4]. At concentrations > 30 μmol/L, the POC method had an upward bias of 7% compared with HPLC‐MS/MS. For an intoxication screening test, upward bias is preferred to downward, as false negative results may lead to unjustified exclusion of paracetamol intoxication and detrimental withdrawal of antidote treatment.

The therapeutic paracetamol concentration ranges up to 200 μmol/L [25], while concentrations < 30 μmol/L can be considered of less significance when evaluating acute toxicity. The POC method shows linear performance within the range of paracetamol concentrations from 0.82 to 2000 μmol/L in vitro [3]. In this study, the concentrations in samples analysed with POC ranged from 0 to 1900 μmol/L in venous serum. All false negative POC samples had concentrations < 45 μmol/L, which indicates good performance of the sensor at clinically relevant concentrations. Supratherapeutic paracetamol concentrations above 200 μmol/L were analysed in 38 samples, out of which 15 were analysed with the POC. The results are promising, but the number of samples is insufficient to confirm the performance of the POC method at toxic concentrations.

Based on these results, the performance of the POC test strip is promising. The POC sensor used in this study was an early prototype, and the measurement protocol included multiple phases. The protocol described above does not meet the criteria for an easy‐to‐use POC method, but the sensor has since been developed towards a fully automated analytical tool. A weakness of this study is the limited number of samples and their random collection. The sample size is inadequate to demonstrate confounding agents that would disrupt the different methods of concentration analysis. However, this study provides a unique insight into the pharmacological profile of paracetamol intoxication patients analysed from blood samples.

Sample 3 presented with a false negative result in PEM with a supratherapeutic concentration in the other two analyses. Unfortunately, the reason for this error cannot be identified retrospectively, as the samples were completely anonymized. Sample 3 contained caffeine, vortioxetine and codeine, in addition to paracetamol. Vortioxetine was also found also in Sample 9, which did not show paracetamol when measured by any of the methods used, so the effect of vortioxetine on the analysis of paracetamol could not be assessed. Codeine was found in 10 samples, and it did not cause systematic error in the other samples. Human error in the handling of the samples is possible.

When the electrochemical POC method was compared with the standard PEM and the ‘gold standard’ UHPLC‐DAD‐CAD, both POC and PEM methods were found to have acceptable sensitivity and specificity. All false negative POC results appeared at low concentrations (< 45 μmol/L). This indicates a favourable safety profile of the POC sensor, as the false negatives were at the low concentration levels. At therapeutic and toxic concentrations, the performance of the POC sensor is promising, though further studies are needed to confirm its sensitivity and specificity.

Ethics Statement

This study was approved by the HUS Helsinki University Hospital (HUS/200/2020). An ethics committee statement was not needed because this was a comparative study of laboratory methods using anonymized samples.

Conflicts of Interest

Eija Kalso has received fees from Orion Pharma for consulting and Haleon for one lecture. Tuomas Lilius has received fees from Biogen for one lecture. All are unrelated to this work.

After the work done in this study, authors N.W. and E.V. were employed by a company which is commercializing the POC method studied in this paper.

Supporting information

Table S1: Concentrations measured with electrochemical POC, photoelectric (PEM) and liquid chromatography method (UHPLC‐DAD‐CAD) sorted by sample ID. Concentrations presented in μmol/L. The lowest value reported by the PEM is 14 μmol/L. NA: not available.

Table S2: Substances in the sample when POC was false negative compared with UHPLC‐DAD‐CAD (control) with threshold of 30 μmol/L.

Table S3: Substances in the sample when POC was false positive for paracetamol compared with UHPLC‐DAD‐CAD (control) with threshold of ≥ 30 μmol/L.

Table S4: Substances found in samples, sorted by Anatomical Therapeutic Chemical (ATC) code.

Kujala J., Lohela T., Wester N., et al., “Electrochemical Point‐of‐Care Test for Assessing Serum Paracetamol Concentration: Comparison With Traditional Methods and Detection of Concomitant Drugs,” Basic & Clinical Pharmacology & Toxicology 137, no. 5 (2025): e70127, 10.1111/bcpt.70127.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.