Clinical Toxicology (2014), 52, 506–511 Copyright © 2014 Informa Healthcare USA, Inc. ISSN: 1556-3650 print / 1556-9519 online DOI: 10.3109/15563650.2014.917180
CRITICAL CARE
Predicting acute acetaminophen hepatotoxicity with acetaminophen-aminotransferase multiplication product and the Psi parameter S. CHOMCHAI1 and C. CHOMCHAI2 1Faculty
of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand University International College, Nakhon Pathom, Thailand
2Mahidol
Context. Prediction of potential hepatotoxicity is important for individualizing therapy with N-acetylcysteine (NAC) in patients with acute acetaminophen overdose. Acetaminophen-aminotransferase multiplication product (APAP ⫻ AT) and the Psi Parameter (Psi) have been reported to be the predictors of acetaminophen hepatotoxicity. Objective. To determine the validity of APAP ⫻ AT and Psi in predicting hepatotoxicity secondary to acute acetaminophen overdose. Materials and methods. We retrospectively reviewed acute acetaminophen overdose cases who were treated with NAC at Siriraj Hospital, Thailand during January 2004–June 2012. The patients’ ages were 12 years or more. Initial acetaminophen concentration (mg/L) and aminotransferase (IU/L) were multiplied to obtain APAP ⫻ AT. Psi were derived from initial acetaminophen concentrations (mg/L) and lag time (hours) to NAC therapy. The cut-off values for APAP ⫻ AT and Psi were 1500 mg∙IU/L2 and 5 mM∙h, respectively. Hepatotoxicity (defined as aspartate or alanine aminotransferase (ALT) greater than 1000 IU/L) was the outcome of interest. Results. A total of 255 patients were included, 32 of whom developed hepatotoxicity. APAP ⫻ AT had sensitivity, specificity, and negative likelihood ratio of 90.6%, 62.8%, and 0.2, respectively. The sensitivity of Psi, specificity, and negative likelihood ratio were 96.9%, 91.5%, and 0.0, respectively. The areas under the curve of the receiver operating characteristic (ROC) curve for APAP ⫻ AT and Psi were 0.82 and 0.96, respectively, with a statistically significant difference between the two methods (p ⫽ 0.002). APAP ⫻ AT showed higher specificity (92.5%) in patients who presented 8–24 h after the overdose. Discussion and conclusion. Psi and APAP ⫻ AT are valid clinical tools in predicting hepatotoxicity secondary to acute acetaminophen overdose in adults. APAP ⫻ AT is useful in predicting a low likelihood of hepatotoxicity after standard NAC therapy among late-presenting patients. Keywords
Psi; Paracetamol; Hepatitis; N-acetylcysteine; Half-life
(Psi)5,6 and acetaminophen-aminotransferase multiplication product7 (APAP ⫻ AT) have been devised to assist in assessing the risk of hepatotoxicity from acute acetaminophen overdose. Psi is a mathematical composite of two conventional risk determinants, acetaminophen concentration and the lag-time to initiation of NAC therapy, with the intention to quantify the supratherapeutic exposure to acetaminophen before the treatment with NAC.5 This tool has been studied in a Canadian population to be a significant predictor of hepatotoxicity6 and validated in a Thai population to be superior to using either acetaminophen concentration or lag-time to NAC therapy alone.8 APAP ⫻ AT, on the other hand, is the multiplication product of simultaneous serum acetaminophen and aminotransferase concentration.7 It has been shown to assess the risk of hepatotoxicity from acute acetaminophen overdose.7 However, except for one retrospective study where APAP ⫻ AT was used to predict hepatotoxicity in 216 patients with acute acetaminophen overdose (with sensitivity and specificity of 44% and 92%, respectively), prognostic validity has not been sufficiently evaluated and its utility in the Asian population has never
Introduction Worldwide, acetaminophen overdose is a common cause of drug-induced hepatotoxicity. Timely administration of N-acetylcysteine (NAC) can protect patients against hepatotoxicity, as well as alleviate the severity of the illness once it has developed.1 Incidences of hepatotoxicity as high as 45% have been reported in acute acetaminophen overdose patients, whose NAC therapy started more than 8–10 h postingestion.2 Currently, a timed serum acetaminophen level interpreted by use of the Rumack–Matthew Nomogram is the only accepted parameter for the assessment of hepatotoxicity risk in acute acetaminophen overdose; therefore, it is also a key determinant in the initiation of NAC therapy.3,4 In recent years, two risk assessment tools, Psi parameter Received 30 November 2013; accepted 14 April 2014. This study protocol was approved by Siriraj Institutional Review Board. Address correspondence to Chulathida Chomchai, Mahidol University International College, 999 Phutthamonthon 4 Road, Salaya, Nakhon Pathom 73170, Thailand. E-mail: [email protected]
506
APAP ⫻ AT product and Psi validity been tested.9 And even in this one study, available only in abstract form, the cut-off point used was unclear (10 000 unspecified units).9 In Asian populations, incidences of acute acetaminophen hepatotoxicity ranged from 6 to 10.2%.8,10 Mechanisms that explains increased susceptibility to acetaminophen hepatotoxicity among Asians included a higher proportion of CYP2E1 polymorphisms, than Caucasian populations, which may result in increased production of the hepatotoxic metabolite N-acetyl-para-benzoquinoneimine (NAPQI).11,12 The aim of our study was to evaluate the validity of APAP ⫻ AT in predicting hepatotoxicity secondary to acute acetaminophen overdose in comparison to Psi in a Thai population.
Materials and methods This study was approved by the Siriraj Institutional Review Board. Medical records of patients, aged 12 years or more, who were admitted to Siriraj Hospital, Bangkok, Thailand between January 2004 and June 2012 with the diagnosis of acute acetaminophen poisoning and received NAC therapy, were retrospectively reviewed. The indication for NAC treatment was an acetaminophen level above the Rumack– Matthew Nomogram’s “treatment line”, the line passing the coordinates 150 mg/L and 4 h.2 Standardized data collection forms were used to extract information which included demographic characteristics, time of acetaminophen ingestion, specific data regarding the overdose included types and doses of ingested drugs, lag-time to initiation NAC therapy, time-specific acetaminophen and aminotransferases levels, and final outcome of the case. Cases were excluded if no NAC therapy was initiated, if at least one of the specific overdose data (time and acetaminophen concentration) was missing, if a delayed presentation of more than 24 h after estimated ingestion time was documented, or if the aminotransferase levels were followed for less than 36 h after the ingestion. “Acute acetaminophen overdose” meant ingestion at one single point or within the period of 4 h. For each
individual patient, APAP ⫻ AT and Psi were calculated from the initial laboratory results. For APAP ⫻ AT calculation, simultaneously measured acetaminophen concentration and transaminases were used. Only the higher value of either aspartate aminotransferase (AST) or alanine aminotransferase (ALT) was used in the calculation.7 For each case, Psi was estimated based on a function of supratherapeutic acetaminophen concentration and delay to NAC treatment.5 The title “Psi” was used according to the original inventors, who used the Greek letter 㳗 (Psi). The formula for Psi estimation was 㳗 ⫽ 11.542[APAP]4h (e⫺0.173ti ⫺ e⫺0.173tf) ⫺ [APAP]threshold (tf ⫺ ti) where [APAP]4h estimated serum acetaminophen concentration (mg/L) at 4 h post-acute ingestion when half-life is 4 h. [APAP]threshold represents the serum acetaminophen concentration (mg/L) at which NAPQI formation just matches the rate of detoxification (45 mg/L was used). The tf means time (hour) of initiation of NAC, or the time for acetaminophen concentration to reach [APAP]threshold, whichever occurs first, and ti is the time (hour) required to deplete intrahepatic glutathione (6 h was used).5,6 In our study, Psi was calculated by inputting the lag-time to blood collection, acetaminophen concentration and lag-time to initiation of NAC therapy into a calculator developed and validated according to the original publication in Psi.5 The lag-times to blood collection and to NAC initiation were obtained based on the estimated time of ingestion. A calculator for Psi is downloadable at http://www.si.mahidol.ac.th/th/department/preventive/ eng/dept_news_detail.asp?n_id ⫽ 23&dept_id ⫽ 17. Cutoff points of 1500 mg∙IU/L2(10000 μmol∙IU/L2) and 5 mM∙h (reference in press) were used for APAP ⫻ AT and Psi, respectively.7 Cases with values above the cutoff for either method were independently classified as “test positive”. Hepatotoxicity and prolonged international normalized ratio (INR), which were the primary
Table 1. Comparison of demographic and clinical variables between patients with and without hepatotoxicity. Variable
Hepatotoxicity (n ⫽ 32)
Non-hepatotoxicity (n ⫽ 223)
Female, count (%) 30 (93.8) 185 (83.0) Age, years 22.0 (20.3–26.8) 23.0 (19.0–27.0) Serum [APAP], mg/L 110.3 (67.5–162.0) 76.4 (46.8–103.3) Serum [APAP] plotted on nomogram, count (%) ⱖ 300 mg/L line 31 (96.9) 39 (17.5) ⱖ 200 mg/L line 32 (100) 90 (40.4) Initial AST, IU/L 51.0 (20.0–160.0) 14.0 (11.0–18.0) Initial ALT, IU/L 49.0 (18.0–139.0) 13.0 (11.0–18.0) NAC treatment onset, hours 15.3 (14.2–16.9) 7.3 (5.0–10.0) NAC treatment onset 8–24 h, count (%) 32 (100) 109 (48.9) 3084.1 (1975.6–10081.0) 1396.4 (1183.9–2564.1) APAP⫻ AT, mg∙IU/L2 Psi, mM∙hour 9.96 (7.204–13.033) 0.81 (0.001–1.175)
p value 0.12 0.73 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001
Serum [APAP] ⫽ patient’s first serum acetaminophen concentration drawn 4–24 h post ingestion, [APAP] ⱖ 200 and 300 mg/L lines ⫽ cases with acetaminophen concentration on or above the Modified Rumack–Matthew Nomogram lines that pass 200 and 300 mg/L at 4 h, respectively, NAC ⫽ N-acetylcysteine, AST ⫽ aspartate aminotransferase, ALT ⫽ alanine aminotransferase, APAP⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Hepatotoxicity ⫽ presence of AST or ALT ⱖ 1000 IU/L, Values in the tables are reported as median(interquartile range) or count(percentage). Copyright © Informa Healthcare USA, Inc. 2014
507
508
S. Chomchai & C. Chomchai
Table 2. Validity of Psi and APAP⫻ AT (overall and late-presenting cases) in predicting hepatotoxicity with their ROC’s areas under the curve (AUC) and the comparison of the AUCs.
Overall (255 cases) Late (138 cases)
Psi APAP⫻ AT Psi APAP⫻ AT
Sensitivity (%) (CI)
Specificity (%) (CI)
LR ⫹ (CI)
LR ⫺ (CI)
AUC
p value
96.9 (84.3, 99.4) 90.6 (75.8, 96.8) 96.9 (84.3, 99.4) 90.6 (75.8, 96.8)
91.5 (87.1, 94.5) 62.8 (56.3, 68.9) 82.1 (73.7, 88.2) 92.5 (85.8, 96.1)
11.4 (7.4, 17.6) 2.4 (2.0, 3.0) 5.5 (3.6, 8.2) 12.0 (6.1, 23.6)
0.0 (0.0, 0.2) 0.2 (0.1, 0.4) 0.0 (0.0, 0.3) 0.1 (0.0, 0.3)
0.96 (0.94, 0.98) 0.82 (0.71, 0.89) 0.92 (0.86, 0.96) 0.95 (0.91, 0.99)
0.002 0.31
APAP⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Late ⫽ patients whose NAC therapy started 8–24 h post overdose, Hepatotoxicity ⫽ presence of aspartate or alanine transferase ⱖ 1000 IU/L, CI ⫽ 95% confidence interval, LR⫹ ⫽ positive likelihood ratio, LR⫺ ⫽ negative likelihood ratio.
and secondary outcomes of interest, were defined as an aminotransferase greater than 1000 IU/L or INR more than 2, respectively.6,7,13 Each initial acetaminophen concentration is plotted in the Modified Rumack– Matthew Nomogram with lines that pass the concentrations 200 and 300 mg/L at 4 h in order to aid readers to better visualize the degree of overdose. Statistical analyses Descriptive qualitative and continuous variables were summarized as frequency with percentage and median with interquartile range (IQR), respectively, due to non-normal distributions. Categorical and continuous variables were analyzed with Chi-square and Mann–Whitney-U tests, respectively. Based on the aforementioned cut-off values, diagnostic validities of each test were evaluated as sensitivity, specificity, positive likelihood ratio (LR⫹) and negative likelihood ratio (LR⫺) with 95% confidence intervals (95% CI). A subgroup of the study population, those with delayed presentation (blood test and NAC therapy beginning at 8–24 h post-ingestion) was also analyzed. Receiver operating characteristic (ROC) were built for each test and the area under the curve (AUC) and 95% CI was estimated and compared between APAP ⫻ AT and Psi. Statistical analyses were performed using SPSS 18.0 statistical package and StAR.14 An alpha of 0.05 was used to determine statistical significance.
(11–19) and 14 (12–18) IU/L, respectively. Hepatotoxicity occurred in 32 cases; no fulminant liver failure or deaths were recorded. All hepatotoxicity cases presented to the hospital more than 8 h after the overdose. Overall, hepatotoxic cases had significantly higher initial acetaminophen and aminotransferases levels, longer lag-time to NAC therapy and larger APAP ⫻ AT and Psi, when compared with nonhepatotoxic cases (Table 1). Median (IQR) of APAP ⫻ AT was 1450 (1216–2770) mg∙IU/L2. There were 112 (43.9%) patients whose APAP ⫻ AT products were above the cut-off point, 29 of whom developed hepatotoxicity. The sensitivity and specificity with 95% CI for APAP ⫻ AT were 90.6% (75.8–96.8) and 62.8% (56.3–68.9), respectively. Median (IQR) of Psi was 0.917 mM∙h (0.001–2.213). Among 50 patients with a Psi above 5 mM∙h, 31 cases (62%) experienced hepatotoxicity, yielding sensitivity and specificity with 95% CI for Psi of 96.9% (84.3–99.4) and 91.5% (87.1–94.5), respectively (Table 2). The AUC with 95% CI of Psi and APAP ⫻ AT were, 0.96 (0.94–0.98) and 0.82 (0.71–0.89), respectively. Psi had a significantly larger AUC than APAP ⫻ AT (p ⫽ 0.002; Fig. 1). ROC curves of Psi and APAP ⫻AT departed the 100% sensitivity at 3.0 mM∙h and
Results We identified 417 patients who were admitted for acetaminophen overdose and treated with NAC and of these, 255 cases were enrolled. Among the 162 excluded cases, 49 had no certain time of acetaminophen ingestion, 51 were aged below 12 years, 14 presented after 24 h, and 48 lacked initial or adequate follow-up aminotransferase. The enrolled patients included the patients in the 2011 publication.8 Two hundred and fifteen cases (84.3%) were women with a median age of 23 years (IQR, 19–27). Median lag-time to the initial blood sample was 8.0 h after ingestion (IQR, 5.5–11.0). Median lag-time of initiation of NAC therapy was 8.3 h (5.7–12.3) after ingestion. Using the modified Rumack–Matthew Nomogram,2 122 cases (40%) and 70 cases (23%) had initial acetaminophen concentrations that plotted above the 200 mg/L and 300 mg/L at 4 h lines, respectively. Median (IQR) initial AST and ALT were 14
Fig. 1. Receiver operating characteristic curves of the acetaminophenaminotransferase multiplication product (APAP ⫻ AT) and the Psi Parameter (Psi) in predicting hepatotoxicity secondary to acute acetaminophen overdose for all patients. Clinical Toxicology vol. 52 no. 5 2014
APAP ⫻ AT product and Psi validity
509
Discussion
Fig. 2. Receiver operating characteristic curves of the acetaminophenaminotransferase multiplication product (APAP⫻ AT) and the Psi Parameter (Psi) in predicting hepatotoxicity secondary to acute acetaminophen overdose for patients with late presentation (8–24 h after ingestion).
1230.1 mg∙IU/L2, respectively, and reached the greatest distances from reference line at 2.9 mM∙h and 1566.8 mg∙IU/L2, respectively. The LR⫹ and LR⫺ with 95% CI of Psi were 11.4 (7.4–17.6) and 0.0 (0.0–0.2), respectively. The APAP ⫻ AT had a LR⫹ and LR⫺ of 2.4 (2.0–3.0) and 0.2 (0.1–0.4), respectively. When only the subgroup of late presenters was analyzed, APAP ⫻ AT’s specificity increased to 92.5% (95% CI, 85.8– 96.1) and AUC increased to 0.95 (95% CI, 0.91–0.99), which are comparable to Psi’s specificity 82.1% (73.7–88.2) and AUC 0.92 (0.86–0.96; Table 2 and Fig. 2). In this subgroup the APAP ⫻ AT had a sensitivity of 90.6% (75.8–96.8), LR⫹ of 12.0 (6.1–23.6) and LR⫺ of 0.1 (0.0–0.3), while sensitivity of Psi, LR⫹, and LR⫺ were 96.9% (84.3–99.4), 5.5 (3.6–8.2), and 0.0 (0.0–0.3), respectively. INRs were available for analysis in 169 overall and 91 latepresenting cases. Analysis of prolonged INR yielded similar results to hepatotoxicity. APAP ⫻ AT had high sensitivities of 100% (64.6–100.0) and 100% (61.0–100.0) in overall and late-presenting patients, respectively. APAP ⫻ AT had low specificity 57.0% (49.7–64.8) and gained higher specificity 81.2% (71.6–88.1) in late presenters (Table 3).
Acetaminophen causes hepatotoxicity through its transformation into the reactive intermediate NAPQI via the cytochrome P 450 2E1 pathway and depletion of protective glutathione.15 Increased acetaminophen concentration and delayed NAC therapy are risk factors for hepatotoxicity due to a higher NAPQI load and more prolonged scarcity of glutathione.15 Psi takes into account these well-recognized physiologic impacts and converts them into quantifiable risk.5 Psi increases in concordance with both risk factors; with the exception that Psi remains low when NAC therapy is started within 6 h, regardless of the acetaminophen concentration.6 APAP ⫻ AT operates on a different pathophysiologic basis from Psi. It recognizes that as serum acetaminophen decreases through drug elimination over time, patients who develops acetaminophen-induced hepatitis will experience an increase in hepatic aminotransferases.7 This increase occurs with a faster onset and achieves a higher magnitude in patients destined to have hepatotoxicity.7,16 The fact that the elimination half-life for acetaminophen is longer in hepatotoxic cases makes it a useful prognostic indicator for acute acetaminophen-induced hepatotoxicity, so that it is used to guide patient-tailored therapy for acute acetaminophen overdose in some institutions.17 In the original APAP ⫻ AT study, Sivilotti et al. used retrospective data of 164 acute acetaminophen overdose cases. The multiplication product was performed using acetaminophen concentrations, which were either obtained (132 cases) or derived (32 cases), together with the measured aminotransferase. All patients with hepatotoxicity had multiplication products that were 10 000 μM∙IU/L (1500 mg∙IU/L2) or more. Moreover, larger multiplication products were associated with a faster onset of hepatotoxicity.7 This is in concordance with our findings of high sensitivity of APAP ⫻ AT. The fact that APAP ⫻ AT has high sensitivity and relatively high specificity for prolonged INRs, especially among late presenters, verifies validity of the tests as a predictor of acute acetaminophen hepatotoxicity. The population in our study was predominately young female and the rate of hepatotoxicity was similar to those reported in previous studies.1,18–21 Our follow-up time for aminotransferase of at least 36 h was expected to be long enough to detect hepatotoxicity if one developed. Both Psi and APAP ⫻ AT are shown to have high sensitivities as initial tests. The superiority of Psi’s ROC comparison for hepatotoxicity can be explained by APAP ⫻ AT’s low specificity in patients who
Table 3. Validity of Psi and APAP⫻ AT (overall and late-presenting cases) in predicting prolonged INR (ⱖ 2) with their ROC’s areas under the curve (AUC) and the comparison of the AUCs.
Overall (169 cases) Late (91 cases)
Psi APAP⫻ AT Psi APAP⫻ AT
Sensitivity (%) (CI)
Specificity (%) (CI)
71.4 (35.9, 91.8) 100.0 (64.6, 100.0) 83.3 (43.6, 97.0) 100.0 (61.0, 100.0)
85.2 (78.9, 89.8) 57.0 (49.7, 64.8) 72.9 (62.7, 81.2) 81.2 (71.6, 88.1)
LR⫹ (CI)
LR⫺ (CI)
AUC
p value
4.8 (2.7, 8.8) 2.4 (2.0, 2.8) 0.4 (0.1, 1.1) 5.3 (3.4, 8.3)
0.3 (0.1, 1.1) 0.0 (0.0, 3.3) 0.2 (0.0, 1.4) 0.0 (0.0, 2.7)
78.7 (58.6, 98.9) 80.2 (67.2, 93.1) 80.2 (70.5, 87.8) 91.4 (83.6, 96.2)
0.91 0.38
APAP ⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Late ⫽ patients whose NAC therapy started 8–24 h post overdose, CI ⫽ 95% confidence interval, INR ⫽ international normalized ratio, LR ⫹ ⫽ positive likelihood ratio, LR⫺ ⫽ negative likelihood ratio. Copyright © Informa Healthcare USA, Inc. 2014
510
S. Chomchai & C. Chomchai
present early after overdose. The high false positive rate (low specificity) is likely the result of early initiation of NAC, which has high protective property against hepatotoxicity. These are demonstrated by low rates of hepatotoxicity among patients whose initial acetaminophen levels were high who had early NAC therapy or low initial aminotransferases (Figs. 3 and 4). We propose APAP ⫻ AT as a risk assessment tool in patients who present with an unknown time of ingestion (in which case the Rumack–Matthew Nomogram and Psi are not applicable). In such cases, the criteria used to determine initiation of NAC therapy is any detectable acetaminophen concentration or an elevated aminotransferase level at the time of presentation. Once treatment is completed, close follow-up of liver function must continue for all such cases because it is uncertain whether the treated patient will develop hepatotoxicity. The high sensitivity of APAP ⫻ AT and low negative likelihood ratio found in this study suggest that APAP ⫻ AT is useful in evaluation of patients with acute acetaminophen overdose regardless of presentation time. Cases with a presenting APAP ⫻ AT lower than 1500 mg∙IU/L2 have low probability to develop hepatotoxicity after a standard course of NAC therapy and may not require further follow-up or prolongation of NAC administration. In our study population, APAP ⫻ AT can exclude at least 56% (143 out of 255) of the patients from further follow-up and treatment. When only late-presenting cases were considered, APAP ⫻ AT’s specificity, positive likelihood ratio and AUC increased markedly, especially in patients presenting at 8–24 h post ingestion. APAP ⫻ AT’s specificity and positive likelihood ratio became larger than that of Psi, with
Fig. 3. Scatter plots showing initial aminotransferases (the higher one among aspartate and alanine aminotransferases) on measured initial acetaminophen concentrations for 255 patients on semi-logarithmic scales. Ranges of aminotransferases and acetaminophen concentrations were 10.0–446.0 IU/L and 11.0–335.1 mg/L, respectively. Circles denote hepatotoxicity cases and small dots indicate non-hepatotoxic cases.
statistical significance (p ⫽ 0.04), while APAP ⫻ AT AUC was comparable with Psi (p value, 0.38; Table 2). These might be the results of earlier rise in transaminases and higher acetaminophen levels from more prolonged halflife. Late presenters whose initial APAP ⫻ AT is above 1500 mg∙IU/L2 have a high positive likelihood ratio and a high probability of developing hepatotoxicity despite NAC therapy and further follow-up and prolonged NAC administration can be expected.
Limitations Our study enrolled only acute APAP overdose patients who were treated within 24 h of overdose, thus we essentially excluded cases with very late treatment whose outcomes are potentially the most serious. Such exclusion can cause an ascertainment bias and, consequently reduced sensitivity. The use of aminotransferase to calculate APAP ⫻ AT and to evaluate the hepatotoxicity can result in incorporation bias, which overestimates the sensitivity. Nevertheless, high sensitivity of APAP ⫻ AT for prolonged INR verifies the test’s validity for hepatotoxicity and disproves the effect of incorporation bias. The small number of hepatotoxicity cases may make the calculated sensitivity and specificity susceptible, at least in theory, to misclassification bias. However, the actual chance for such error may be low because most of the parameters of interest are objectively measured laboratory values. The retrospective nature of the study presents a challenge in terms of estimation of times of ingestion and blood collection, which could lead to imprecise calculation of Psi. Despite the high sensitivity, because our study enrolled only patients who received NAC, it is premature to conclude that NAC therapy
Fig. 4. Scatter plots showing lag-periods to initiation of N-acetylcysteine therapy (NAC onset) on measured initial acetaminophen concentration for 255 patients on semi-logarithmic scales. Ranges of NAC onsets and acetaminophen concentrations were 4.0–24.0 h and 11.0–335.1 mg/L, respectively. Circles denote hepatotoxicity cases and small dots indicate non-hepatotoxic cases. Clinical Toxicology vol. 52 no. 5 2014
APAP ⫻ AT product and Psi validity can be withheld in patients with a low APAP ⫻ AT. Such limitations can be remedied by designing a larger prospective study that includes cases with and without NAC therapy. Since we currently do not have a good risk assessment tool for time unknown ingestions, APAP ⫻ AT might prove to be useful in these situations. We suggest that APAP ⫻ AT’s prediction of hepatotoxicity is prospectively investigated in such scenarios. Moreover, correlation of acetaminophen half-life and APAP ⫻ AT should be assessed to conceptualize the APAP ⫻ AT better.
Conclusions APAP ⫻ AT and Psi are useful clinical tools in assessing acute acetaminophen-induced hepatotoxicity in adults. Overall, they perform with high sensitivity. In late presenting cases, the APAP ⫻ AT yields a higher specificity for hepatotoxicity
Declaration of interest The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.
References 1. Brok J, Buckley N, Gluud C. Interventions for paracetamol (acetaminophen) overdose. Cochrane Database Syst Rev 2006:CD003328. 2. Rumack BH. Acetaminophen hepatotoxicity: the first 35 years. J Toxicol Clin Toxicol 2002; 40:3–20. 3. Dart RC, Rumack BH. Patient-tailored acetylcysteine administration. Ann Emerg Med 2007; 50:280–281. 4. Tsai CL, Fang CC, Walson PD. Key features of the patient-tailored N-acetylcysteine protocol. Ann Emerg Med 2008; 51:451–452. 5. Sivilotti ML, Good AM, Yarema MC, Juurlink DN, Johnson DW. A new predictor of toxicity following acetaminophen overdose based on pretreatment exposure. Clin Toxicol (Phila) 2005; 43:229–234. 6. Sivilotti ML, Yarema MC, Juurlink DN, Good AM, Johnson DW. A risk quantification instrument for acute acetaminophen overdose patients treated with N-acetylcysteine. Ann Emerg Med 2005; 46:263–271.
Copyright © Informa Healthcare USA, Inc. 2014
511
7. Sivilotti ML, Green TJ, Langmann C, Yarema M, Juurlink D, Johnson D. Multiplying the serum aminotransferase by the acetaminophen concentration to predict toxicity following overdose. Clin Toxicol (Phila) 2010; 48:793–799. 8. Chomchai S, Chomchai C, Anusornsuwan T. Acetaminophen psi parameter: a useful tool to quantify hepatotoxicity risk in acute acetaminophen overdose. Clin Toxicol (Phila) 2011; 49:664–667. 9. Offerman S, Young M. Use of the initial acetaminophen concentration times serum aminotransferaseproduct to predict significant liver enzyme elevation after acetaminophen overdose. Clin Toxicol (Phila) 2011; 49:591. 10. Marzilawati AR, Ngau YY, Mahadeva S. Low rates of hepatotoxicity among Asian patients with paracetamol overdose: a review of 1024 cases. BMC Pharmacol Toxicol 2012; 13:8. 11. Sangrajrang S, Jedpiyawongse A, Srivatanakul P. Genetic polymorphisms of CYP2E1 and GSTM1 in a Thai population. Asian Pac J Cancer Prev 2006; 7:415–419. 12. Ueshima Y, Tsutsumi M, Takase S, Matsuda Y, Kawahara H. Acetaminophen metabolism in patients with different cytochrome P-4502E1 genotypes. Alcohol Clin Exp Res 1996; 20:25A–28A. 13. Yarema MC, Johnson DW, Berlin RJ, Sivilotti ML, Nettel-Aguirre A, Brant RF, et al. Comparison of the 20-hour intravenous and 72-hour oral acetylcysteine protocols for the treatment of acute acetaminophen poisoning. Ann Emerg Med 2009; 54:606–614. 14. Vergara IA, Norambuena T, Ferrada E, Slater AW, Melo F. StAR: a simple tool for the statistical comparison of ROC curves. BMC Bioinformatics 2008; 9:265. 15. Hodgman MJ, Garrard AR. A review of acetaminophen poisoning. Crit Care Clin 2012; 28:499–516. 16. Green TJ, Sivilotti ML, Langmann C, Yarema M, Juurlink D, Burns MJ, et al. When do the aminotransferases rise after acute acetaminophen overdose? Clin Toxicol (Phila) 2010; 48:787–792. 17. Schiodt FV, Ott P, Christensen E, Bondesen S. The value of plasma acetaminophen half-life in antidote-treated acetaminophen overdosage. Clin Pharmacol Ther 2002; 71:221–225. 18. Tsai CL, Chang WT, Weng TI, Fang CC, Chen WJ. Acute acetaminophen intoxication in Taiwan: outcomes and risk factors. J Formos Med Assoc 2004; 103:830–835. 19. Senarathna SM, Ranganathan SS, Dawson AH, Buckley N, Fernandopulle BM. Management of acute paracetamol poisoning in a tertiary care hospital. Ceylon Med J 2008; 53:89–92. 20. Johnson MT, McCammon CA, Mullins ME, Halcomb SE. Evaluation of a simplified N-acetylcysteine dosing regimen for the treatment of acetaminophen toxicity. Ann Pharmacother 2011; 45:713–720. 21. Offerman SR. The clinical management of acetaminophen poisoning in a community hospital system: factors associated with hospital length of stay. J Med Toxicol 2011; 7:4–11.
Copyright of Clinical Toxicology (15563650) is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
CRITICAL CARE
Predicting acute acetaminophen hepatotoxicity with acetaminophen-aminotransferase multiplication product and the Psi parameter S. CHOMCHAI1 and C. CHOMCHAI2 1Faculty
of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand University International College, Nakhon Pathom, Thailand
2Mahidol
Context. Prediction of potential hepatotoxicity is important for individualizing therapy with N-acetylcysteine (NAC) in patients with acute acetaminophen overdose. Acetaminophen-aminotransferase multiplication product (APAP ⫻ AT) and the Psi Parameter (Psi) have been reported to be the predictors of acetaminophen hepatotoxicity. Objective. To determine the validity of APAP ⫻ AT and Psi in predicting hepatotoxicity secondary to acute acetaminophen overdose. Materials and methods. We retrospectively reviewed acute acetaminophen overdose cases who were treated with NAC at Siriraj Hospital, Thailand during January 2004–June 2012. The patients’ ages were 12 years or more. Initial acetaminophen concentration (mg/L) and aminotransferase (IU/L) were multiplied to obtain APAP ⫻ AT. Psi were derived from initial acetaminophen concentrations (mg/L) and lag time (hours) to NAC therapy. The cut-off values for APAP ⫻ AT and Psi were 1500 mg∙IU/L2 and 5 mM∙h, respectively. Hepatotoxicity (defined as aspartate or alanine aminotransferase (ALT) greater than 1000 IU/L) was the outcome of interest. Results. A total of 255 patients were included, 32 of whom developed hepatotoxicity. APAP ⫻ AT had sensitivity, specificity, and negative likelihood ratio of 90.6%, 62.8%, and 0.2, respectively. The sensitivity of Psi, specificity, and negative likelihood ratio were 96.9%, 91.5%, and 0.0, respectively. The areas under the curve of the receiver operating characteristic (ROC) curve for APAP ⫻ AT and Psi were 0.82 and 0.96, respectively, with a statistically significant difference between the two methods (p ⫽ 0.002). APAP ⫻ AT showed higher specificity (92.5%) in patients who presented 8–24 h after the overdose. Discussion and conclusion. Psi and APAP ⫻ AT are valid clinical tools in predicting hepatotoxicity secondary to acute acetaminophen overdose in adults. APAP ⫻ AT is useful in predicting a low likelihood of hepatotoxicity after standard NAC therapy among late-presenting patients. Keywords
Psi; Paracetamol; Hepatitis; N-acetylcysteine; Half-life
(Psi)5,6 and acetaminophen-aminotransferase multiplication product7 (APAP ⫻ AT) have been devised to assist in assessing the risk of hepatotoxicity from acute acetaminophen overdose. Psi is a mathematical composite of two conventional risk determinants, acetaminophen concentration and the lag-time to initiation of NAC therapy, with the intention to quantify the supratherapeutic exposure to acetaminophen before the treatment with NAC.5 This tool has been studied in a Canadian population to be a significant predictor of hepatotoxicity6 and validated in a Thai population to be superior to using either acetaminophen concentration or lag-time to NAC therapy alone.8 APAP ⫻ AT, on the other hand, is the multiplication product of simultaneous serum acetaminophen and aminotransferase concentration.7 It has been shown to assess the risk of hepatotoxicity from acute acetaminophen overdose.7 However, except for one retrospective study where APAP ⫻ AT was used to predict hepatotoxicity in 216 patients with acute acetaminophen overdose (with sensitivity and specificity of 44% and 92%, respectively), prognostic validity has not been sufficiently evaluated and its utility in the Asian population has never
Introduction Worldwide, acetaminophen overdose is a common cause of drug-induced hepatotoxicity. Timely administration of N-acetylcysteine (NAC) can protect patients against hepatotoxicity, as well as alleviate the severity of the illness once it has developed.1 Incidences of hepatotoxicity as high as 45% have been reported in acute acetaminophen overdose patients, whose NAC therapy started more than 8–10 h postingestion.2 Currently, a timed serum acetaminophen level interpreted by use of the Rumack–Matthew Nomogram is the only accepted parameter for the assessment of hepatotoxicity risk in acute acetaminophen overdose; therefore, it is also a key determinant in the initiation of NAC therapy.3,4 In recent years, two risk assessment tools, Psi parameter Received 30 November 2013; accepted 14 April 2014. This study protocol was approved by Siriraj Institutional Review Board. Address correspondence to Chulathida Chomchai, Mahidol University International College, 999 Phutthamonthon 4 Road, Salaya, Nakhon Pathom 73170, Thailand. E-mail: [email protected]
506
APAP ⫻ AT product and Psi validity been tested.9 And even in this one study, available only in abstract form, the cut-off point used was unclear (10 000 unspecified units).9 In Asian populations, incidences of acute acetaminophen hepatotoxicity ranged from 6 to 10.2%.8,10 Mechanisms that explains increased susceptibility to acetaminophen hepatotoxicity among Asians included a higher proportion of CYP2E1 polymorphisms, than Caucasian populations, which may result in increased production of the hepatotoxic metabolite N-acetyl-para-benzoquinoneimine (NAPQI).11,12 The aim of our study was to evaluate the validity of APAP ⫻ AT in predicting hepatotoxicity secondary to acute acetaminophen overdose in comparison to Psi in a Thai population.
Materials and methods This study was approved by the Siriraj Institutional Review Board. Medical records of patients, aged 12 years or more, who were admitted to Siriraj Hospital, Bangkok, Thailand between January 2004 and June 2012 with the diagnosis of acute acetaminophen poisoning and received NAC therapy, were retrospectively reviewed. The indication for NAC treatment was an acetaminophen level above the Rumack– Matthew Nomogram’s “treatment line”, the line passing the coordinates 150 mg/L and 4 h.2 Standardized data collection forms were used to extract information which included demographic characteristics, time of acetaminophen ingestion, specific data regarding the overdose included types and doses of ingested drugs, lag-time to initiation NAC therapy, time-specific acetaminophen and aminotransferases levels, and final outcome of the case. Cases were excluded if no NAC therapy was initiated, if at least one of the specific overdose data (time and acetaminophen concentration) was missing, if a delayed presentation of more than 24 h after estimated ingestion time was documented, or if the aminotransferase levels were followed for less than 36 h after the ingestion. “Acute acetaminophen overdose” meant ingestion at one single point or within the period of 4 h. For each
individual patient, APAP ⫻ AT and Psi were calculated from the initial laboratory results. For APAP ⫻ AT calculation, simultaneously measured acetaminophen concentration and transaminases were used. Only the higher value of either aspartate aminotransferase (AST) or alanine aminotransferase (ALT) was used in the calculation.7 For each case, Psi was estimated based on a function of supratherapeutic acetaminophen concentration and delay to NAC treatment.5 The title “Psi” was used according to the original inventors, who used the Greek letter 㳗 (Psi). The formula for Psi estimation was 㳗 ⫽ 11.542[APAP]4h (e⫺0.173ti ⫺ e⫺0.173tf) ⫺ [APAP]threshold (tf ⫺ ti) where [APAP]4h estimated serum acetaminophen concentration (mg/L) at 4 h post-acute ingestion when half-life is 4 h. [APAP]threshold represents the serum acetaminophen concentration (mg/L) at which NAPQI formation just matches the rate of detoxification (45 mg/L was used). The tf means time (hour) of initiation of NAC, or the time for acetaminophen concentration to reach [APAP]threshold, whichever occurs first, and ti is the time (hour) required to deplete intrahepatic glutathione (6 h was used).5,6 In our study, Psi was calculated by inputting the lag-time to blood collection, acetaminophen concentration and lag-time to initiation of NAC therapy into a calculator developed and validated according to the original publication in Psi.5 The lag-times to blood collection and to NAC initiation were obtained based on the estimated time of ingestion. A calculator for Psi is downloadable at http://www.si.mahidol.ac.th/th/department/preventive/ eng/dept_news_detail.asp?n_id ⫽ 23&dept_id ⫽ 17. Cutoff points of 1500 mg∙IU/L2(10000 μmol∙IU/L2) and 5 mM∙h (reference in press) were used for APAP ⫻ AT and Psi, respectively.7 Cases with values above the cutoff for either method were independently classified as “test positive”. Hepatotoxicity and prolonged international normalized ratio (INR), which were the primary
Table 1. Comparison of demographic and clinical variables between patients with and without hepatotoxicity. Variable
Hepatotoxicity (n ⫽ 32)
Non-hepatotoxicity (n ⫽ 223)
Female, count (%) 30 (93.8) 185 (83.0) Age, years 22.0 (20.3–26.8) 23.0 (19.0–27.0) Serum [APAP], mg/L 110.3 (67.5–162.0) 76.4 (46.8–103.3) Serum [APAP] plotted on nomogram, count (%) ⱖ 300 mg/L line 31 (96.9) 39 (17.5) ⱖ 200 mg/L line 32 (100) 90 (40.4) Initial AST, IU/L 51.0 (20.0–160.0) 14.0 (11.0–18.0) Initial ALT, IU/L 49.0 (18.0–139.0) 13.0 (11.0–18.0) NAC treatment onset, hours 15.3 (14.2–16.9) 7.3 (5.0–10.0) NAC treatment onset 8–24 h, count (%) 32 (100) 109 (48.9) 3084.1 (1975.6–10081.0) 1396.4 (1183.9–2564.1) APAP⫻ AT, mg∙IU/L2 Psi, mM∙hour 9.96 (7.204–13.033) 0.81 (0.001–1.175)
p value 0.12 0.73 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001
Serum [APAP] ⫽ patient’s first serum acetaminophen concentration drawn 4–24 h post ingestion, [APAP] ⱖ 200 and 300 mg/L lines ⫽ cases with acetaminophen concentration on or above the Modified Rumack–Matthew Nomogram lines that pass 200 and 300 mg/L at 4 h, respectively, NAC ⫽ N-acetylcysteine, AST ⫽ aspartate aminotransferase, ALT ⫽ alanine aminotransferase, APAP⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Hepatotoxicity ⫽ presence of AST or ALT ⱖ 1000 IU/L, Values in the tables are reported as median(interquartile range) or count(percentage). Copyright © Informa Healthcare USA, Inc. 2014
507
508
S. Chomchai & C. Chomchai
Table 2. Validity of Psi and APAP⫻ AT (overall and late-presenting cases) in predicting hepatotoxicity with their ROC’s areas under the curve (AUC) and the comparison of the AUCs.
Overall (255 cases) Late (138 cases)
Psi APAP⫻ AT Psi APAP⫻ AT
Sensitivity (%) (CI)
Specificity (%) (CI)
LR ⫹ (CI)
LR ⫺ (CI)
AUC
p value
96.9 (84.3, 99.4) 90.6 (75.8, 96.8) 96.9 (84.3, 99.4) 90.6 (75.8, 96.8)
91.5 (87.1, 94.5) 62.8 (56.3, 68.9) 82.1 (73.7, 88.2) 92.5 (85.8, 96.1)
11.4 (7.4, 17.6) 2.4 (2.0, 3.0) 5.5 (3.6, 8.2) 12.0 (6.1, 23.6)
0.0 (0.0, 0.2) 0.2 (0.1, 0.4) 0.0 (0.0, 0.3) 0.1 (0.0, 0.3)
0.96 (0.94, 0.98) 0.82 (0.71, 0.89) 0.92 (0.86, 0.96) 0.95 (0.91, 0.99)
0.002 0.31
APAP⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Late ⫽ patients whose NAC therapy started 8–24 h post overdose, Hepatotoxicity ⫽ presence of aspartate or alanine transferase ⱖ 1000 IU/L, CI ⫽ 95% confidence interval, LR⫹ ⫽ positive likelihood ratio, LR⫺ ⫽ negative likelihood ratio.
and secondary outcomes of interest, were defined as an aminotransferase greater than 1000 IU/L or INR more than 2, respectively.6,7,13 Each initial acetaminophen concentration is plotted in the Modified Rumack– Matthew Nomogram with lines that pass the concentrations 200 and 300 mg/L at 4 h in order to aid readers to better visualize the degree of overdose. Statistical analyses Descriptive qualitative and continuous variables were summarized as frequency with percentage and median with interquartile range (IQR), respectively, due to non-normal distributions. Categorical and continuous variables were analyzed with Chi-square and Mann–Whitney-U tests, respectively. Based on the aforementioned cut-off values, diagnostic validities of each test were evaluated as sensitivity, specificity, positive likelihood ratio (LR⫹) and negative likelihood ratio (LR⫺) with 95% confidence intervals (95% CI). A subgroup of the study population, those with delayed presentation (blood test and NAC therapy beginning at 8–24 h post-ingestion) was also analyzed. Receiver operating characteristic (ROC) were built for each test and the area under the curve (AUC) and 95% CI was estimated and compared between APAP ⫻ AT and Psi. Statistical analyses were performed using SPSS 18.0 statistical package and StAR.14 An alpha of 0.05 was used to determine statistical significance.
(11–19) and 14 (12–18) IU/L, respectively. Hepatotoxicity occurred in 32 cases; no fulminant liver failure or deaths were recorded. All hepatotoxicity cases presented to the hospital more than 8 h after the overdose. Overall, hepatotoxic cases had significantly higher initial acetaminophen and aminotransferases levels, longer lag-time to NAC therapy and larger APAP ⫻ AT and Psi, when compared with nonhepatotoxic cases (Table 1). Median (IQR) of APAP ⫻ AT was 1450 (1216–2770) mg∙IU/L2. There were 112 (43.9%) patients whose APAP ⫻ AT products were above the cut-off point, 29 of whom developed hepatotoxicity. The sensitivity and specificity with 95% CI for APAP ⫻ AT were 90.6% (75.8–96.8) and 62.8% (56.3–68.9), respectively. Median (IQR) of Psi was 0.917 mM∙h (0.001–2.213). Among 50 patients with a Psi above 5 mM∙h, 31 cases (62%) experienced hepatotoxicity, yielding sensitivity and specificity with 95% CI for Psi of 96.9% (84.3–99.4) and 91.5% (87.1–94.5), respectively (Table 2). The AUC with 95% CI of Psi and APAP ⫻ AT were, 0.96 (0.94–0.98) and 0.82 (0.71–0.89), respectively. Psi had a significantly larger AUC than APAP ⫻ AT (p ⫽ 0.002; Fig. 1). ROC curves of Psi and APAP ⫻AT departed the 100% sensitivity at 3.0 mM∙h and
Results We identified 417 patients who were admitted for acetaminophen overdose and treated with NAC and of these, 255 cases were enrolled. Among the 162 excluded cases, 49 had no certain time of acetaminophen ingestion, 51 were aged below 12 years, 14 presented after 24 h, and 48 lacked initial or adequate follow-up aminotransferase. The enrolled patients included the patients in the 2011 publication.8 Two hundred and fifteen cases (84.3%) were women with a median age of 23 years (IQR, 19–27). Median lag-time to the initial blood sample was 8.0 h after ingestion (IQR, 5.5–11.0). Median lag-time of initiation of NAC therapy was 8.3 h (5.7–12.3) after ingestion. Using the modified Rumack–Matthew Nomogram,2 122 cases (40%) and 70 cases (23%) had initial acetaminophen concentrations that plotted above the 200 mg/L and 300 mg/L at 4 h lines, respectively. Median (IQR) initial AST and ALT were 14
Fig. 1. Receiver operating characteristic curves of the acetaminophenaminotransferase multiplication product (APAP ⫻ AT) and the Psi Parameter (Psi) in predicting hepatotoxicity secondary to acute acetaminophen overdose for all patients. Clinical Toxicology vol. 52 no. 5 2014
APAP ⫻ AT product and Psi validity
509
Discussion
Fig. 2. Receiver operating characteristic curves of the acetaminophenaminotransferase multiplication product (APAP⫻ AT) and the Psi Parameter (Psi) in predicting hepatotoxicity secondary to acute acetaminophen overdose for patients with late presentation (8–24 h after ingestion).
1230.1 mg∙IU/L2, respectively, and reached the greatest distances from reference line at 2.9 mM∙h and 1566.8 mg∙IU/L2, respectively. The LR⫹ and LR⫺ with 95% CI of Psi were 11.4 (7.4–17.6) and 0.0 (0.0–0.2), respectively. The APAP ⫻ AT had a LR⫹ and LR⫺ of 2.4 (2.0–3.0) and 0.2 (0.1–0.4), respectively. When only the subgroup of late presenters was analyzed, APAP ⫻ AT’s specificity increased to 92.5% (95% CI, 85.8– 96.1) and AUC increased to 0.95 (95% CI, 0.91–0.99), which are comparable to Psi’s specificity 82.1% (73.7–88.2) and AUC 0.92 (0.86–0.96; Table 2 and Fig. 2). In this subgroup the APAP ⫻ AT had a sensitivity of 90.6% (75.8–96.8), LR⫹ of 12.0 (6.1–23.6) and LR⫺ of 0.1 (0.0–0.3), while sensitivity of Psi, LR⫹, and LR⫺ were 96.9% (84.3–99.4), 5.5 (3.6–8.2), and 0.0 (0.0–0.3), respectively. INRs were available for analysis in 169 overall and 91 latepresenting cases. Analysis of prolonged INR yielded similar results to hepatotoxicity. APAP ⫻ AT had high sensitivities of 100% (64.6–100.0) and 100% (61.0–100.0) in overall and late-presenting patients, respectively. APAP ⫻ AT had low specificity 57.0% (49.7–64.8) and gained higher specificity 81.2% (71.6–88.1) in late presenters (Table 3).
Acetaminophen causes hepatotoxicity through its transformation into the reactive intermediate NAPQI via the cytochrome P 450 2E1 pathway and depletion of protective glutathione.15 Increased acetaminophen concentration and delayed NAC therapy are risk factors for hepatotoxicity due to a higher NAPQI load and more prolonged scarcity of glutathione.15 Psi takes into account these well-recognized physiologic impacts and converts them into quantifiable risk.5 Psi increases in concordance with both risk factors; with the exception that Psi remains low when NAC therapy is started within 6 h, regardless of the acetaminophen concentration.6 APAP ⫻ AT operates on a different pathophysiologic basis from Psi. It recognizes that as serum acetaminophen decreases through drug elimination over time, patients who develops acetaminophen-induced hepatitis will experience an increase in hepatic aminotransferases.7 This increase occurs with a faster onset and achieves a higher magnitude in patients destined to have hepatotoxicity.7,16 The fact that the elimination half-life for acetaminophen is longer in hepatotoxic cases makes it a useful prognostic indicator for acute acetaminophen-induced hepatotoxicity, so that it is used to guide patient-tailored therapy for acute acetaminophen overdose in some institutions.17 In the original APAP ⫻ AT study, Sivilotti et al. used retrospective data of 164 acute acetaminophen overdose cases. The multiplication product was performed using acetaminophen concentrations, which were either obtained (132 cases) or derived (32 cases), together with the measured aminotransferase. All patients with hepatotoxicity had multiplication products that were 10 000 μM∙IU/L (1500 mg∙IU/L2) or more. Moreover, larger multiplication products were associated with a faster onset of hepatotoxicity.7 This is in concordance with our findings of high sensitivity of APAP ⫻ AT. The fact that APAP ⫻ AT has high sensitivity and relatively high specificity for prolonged INRs, especially among late presenters, verifies validity of the tests as a predictor of acute acetaminophen hepatotoxicity. The population in our study was predominately young female and the rate of hepatotoxicity was similar to those reported in previous studies.1,18–21 Our follow-up time for aminotransferase of at least 36 h was expected to be long enough to detect hepatotoxicity if one developed. Both Psi and APAP ⫻ AT are shown to have high sensitivities as initial tests. The superiority of Psi’s ROC comparison for hepatotoxicity can be explained by APAP ⫻ AT’s low specificity in patients who
Table 3. Validity of Psi and APAP⫻ AT (overall and late-presenting cases) in predicting prolonged INR (ⱖ 2) with their ROC’s areas under the curve (AUC) and the comparison of the AUCs.
Overall (169 cases) Late (91 cases)
Psi APAP⫻ AT Psi APAP⫻ AT
Sensitivity (%) (CI)
Specificity (%) (CI)
71.4 (35.9, 91.8) 100.0 (64.6, 100.0) 83.3 (43.6, 97.0) 100.0 (61.0, 100.0)
85.2 (78.9, 89.8) 57.0 (49.7, 64.8) 72.9 (62.7, 81.2) 81.2 (71.6, 88.1)
LR⫹ (CI)
LR⫺ (CI)
AUC
p value
4.8 (2.7, 8.8) 2.4 (2.0, 2.8) 0.4 (0.1, 1.1) 5.3 (3.4, 8.3)
0.3 (0.1, 1.1) 0.0 (0.0, 3.3) 0.2 (0.0, 1.4) 0.0 (0.0, 2.7)
78.7 (58.6, 98.9) 80.2 (67.2, 93.1) 80.2 (70.5, 87.8) 91.4 (83.6, 96.2)
0.91 0.38
APAP ⫻ AT ⫽ acetaminophen-aminotransferase multiplication product, Late ⫽ patients whose NAC therapy started 8–24 h post overdose, CI ⫽ 95% confidence interval, INR ⫽ international normalized ratio, LR ⫹ ⫽ positive likelihood ratio, LR⫺ ⫽ negative likelihood ratio. Copyright © Informa Healthcare USA, Inc. 2014
510
S. Chomchai & C. Chomchai
present early after overdose. The high false positive rate (low specificity) is likely the result of early initiation of NAC, which has high protective property against hepatotoxicity. These are demonstrated by low rates of hepatotoxicity among patients whose initial acetaminophen levels were high who had early NAC therapy or low initial aminotransferases (Figs. 3 and 4). We propose APAP ⫻ AT as a risk assessment tool in patients who present with an unknown time of ingestion (in which case the Rumack–Matthew Nomogram and Psi are not applicable). In such cases, the criteria used to determine initiation of NAC therapy is any detectable acetaminophen concentration or an elevated aminotransferase level at the time of presentation. Once treatment is completed, close follow-up of liver function must continue for all such cases because it is uncertain whether the treated patient will develop hepatotoxicity. The high sensitivity of APAP ⫻ AT and low negative likelihood ratio found in this study suggest that APAP ⫻ AT is useful in evaluation of patients with acute acetaminophen overdose regardless of presentation time. Cases with a presenting APAP ⫻ AT lower than 1500 mg∙IU/L2 have low probability to develop hepatotoxicity after a standard course of NAC therapy and may not require further follow-up or prolongation of NAC administration. In our study population, APAP ⫻ AT can exclude at least 56% (143 out of 255) of the patients from further follow-up and treatment. When only late-presenting cases were considered, APAP ⫻ AT’s specificity, positive likelihood ratio and AUC increased markedly, especially in patients presenting at 8–24 h post ingestion. APAP ⫻ AT’s specificity and positive likelihood ratio became larger than that of Psi, with
Fig. 3. Scatter plots showing initial aminotransferases (the higher one among aspartate and alanine aminotransferases) on measured initial acetaminophen concentrations for 255 patients on semi-logarithmic scales. Ranges of aminotransferases and acetaminophen concentrations were 10.0–446.0 IU/L and 11.0–335.1 mg/L, respectively. Circles denote hepatotoxicity cases and small dots indicate non-hepatotoxic cases.
statistical significance (p ⫽ 0.04), while APAP ⫻ AT AUC was comparable with Psi (p value, 0.38; Table 2). These might be the results of earlier rise in transaminases and higher acetaminophen levels from more prolonged halflife. Late presenters whose initial APAP ⫻ AT is above 1500 mg∙IU/L2 have a high positive likelihood ratio and a high probability of developing hepatotoxicity despite NAC therapy and further follow-up and prolonged NAC administration can be expected.
Limitations Our study enrolled only acute APAP overdose patients who were treated within 24 h of overdose, thus we essentially excluded cases with very late treatment whose outcomes are potentially the most serious. Such exclusion can cause an ascertainment bias and, consequently reduced sensitivity. The use of aminotransferase to calculate APAP ⫻ AT and to evaluate the hepatotoxicity can result in incorporation bias, which overestimates the sensitivity. Nevertheless, high sensitivity of APAP ⫻ AT for prolonged INR verifies the test’s validity for hepatotoxicity and disproves the effect of incorporation bias. The small number of hepatotoxicity cases may make the calculated sensitivity and specificity susceptible, at least in theory, to misclassification bias. However, the actual chance for such error may be low because most of the parameters of interest are objectively measured laboratory values. The retrospective nature of the study presents a challenge in terms of estimation of times of ingestion and blood collection, which could lead to imprecise calculation of Psi. Despite the high sensitivity, because our study enrolled only patients who received NAC, it is premature to conclude that NAC therapy
Fig. 4. Scatter plots showing lag-periods to initiation of N-acetylcysteine therapy (NAC onset) on measured initial acetaminophen concentration for 255 patients on semi-logarithmic scales. Ranges of NAC onsets and acetaminophen concentrations were 4.0–24.0 h and 11.0–335.1 mg/L, respectively. Circles denote hepatotoxicity cases and small dots indicate non-hepatotoxic cases. Clinical Toxicology vol. 52 no. 5 2014
APAP ⫻ AT product and Psi validity can be withheld in patients with a low APAP ⫻ AT. Such limitations can be remedied by designing a larger prospective study that includes cases with and without NAC therapy. Since we currently do not have a good risk assessment tool for time unknown ingestions, APAP ⫻ AT might prove to be useful in these situations. We suggest that APAP ⫻ AT’s prediction of hepatotoxicity is prospectively investigated in such scenarios. Moreover, correlation of acetaminophen half-life and APAP ⫻ AT should be assessed to conceptualize the APAP ⫻ AT better.
Conclusions APAP ⫻ AT and Psi are useful clinical tools in assessing acute acetaminophen-induced hepatotoxicity in adults. Overall, they perform with high sensitivity. In late presenting cases, the APAP ⫻ AT yields a higher specificity for hepatotoxicity
Declaration of interest The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.
References 1. Brok J, Buckley N, Gluud C. Interventions for paracetamol (acetaminophen) overdose. Cochrane Database Syst Rev 2006:CD003328. 2. Rumack BH. Acetaminophen hepatotoxicity: the first 35 years. J Toxicol Clin Toxicol 2002; 40:3–20. 3. Dart RC, Rumack BH. Patient-tailored acetylcysteine administration. Ann Emerg Med 2007; 50:280–281. 4. Tsai CL, Fang CC, Walson PD. Key features of the patient-tailored N-acetylcysteine protocol. Ann Emerg Med 2008; 51:451–452. 5. Sivilotti ML, Good AM, Yarema MC, Juurlink DN, Johnson DW. A new predictor of toxicity following acetaminophen overdose based on pretreatment exposure. Clin Toxicol (Phila) 2005; 43:229–234. 6. Sivilotti ML, Yarema MC, Juurlink DN, Good AM, Johnson DW. A risk quantification instrument for acute acetaminophen overdose patients treated with N-acetylcysteine. Ann Emerg Med 2005; 46:263–271.
Copyright © Informa Healthcare USA, Inc. 2014
511
7. Sivilotti ML, Green TJ, Langmann C, Yarema M, Juurlink D, Johnson D. Multiplying the serum aminotransferase by the acetaminophen concentration to predict toxicity following overdose. Clin Toxicol (Phila) 2010; 48:793–799. 8. Chomchai S, Chomchai C, Anusornsuwan T. Acetaminophen psi parameter: a useful tool to quantify hepatotoxicity risk in acute acetaminophen overdose. Clin Toxicol (Phila) 2011; 49:664–667. 9. Offerman S, Young M. Use of the initial acetaminophen concentration times serum aminotransferaseproduct to predict significant liver enzyme elevation after acetaminophen overdose. Clin Toxicol (Phila) 2011; 49:591. 10. Marzilawati AR, Ngau YY, Mahadeva S. Low rates of hepatotoxicity among Asian patients with paracetamol overdose: a review of 1024 cases. BMC Pharmacol Toxicol 2012; 13:8. 11. Sangrajrang S, Jedpiyawongse A, Srivatanakul P. Genetic polymorphisms of CYP2E1 and GSTM1 in a Thai population. Asian Pac J Cancer Prev 2006; 7:415–419. 12. Ueshima Y, Tsutsumi M, Takase S, Matsuda Y, Kawahara H. Acetaminophen metabolism in patients with different cytochrome P-4502E1 genotypes. Alcohol Clin Exp Res 1996; 20:25A–28A. 13. Yarema MC, Johnson DW, Berlin RJ, Sivilotti ML, Nettel-Aguirre A, Brant RF, et al. Comparison of the 20-hour intravenous and 72-hour oral acetylcysteine protocols for the treatment of acute acetaminophen poisoning. Ann Emerg Med 2009; 54:606–614. 14. Vergara IA, Norambuena T, Ferrada E, Slater AW, Melo F. StAR: a simple tool for the statistical comparison of ROC curves. BMC Bioinformatics 2008; 9:265. 15. Hodgman MJ, Garrard AR. A review of acetaminophen poisoning. Crit Care Clin 2012; 28:499–516. 16. Green TJ, Sivilotti ML, Langmann C, Yarema M, Juurlink D, Burns MJ, et al. When do the aminotransferases rise after acute acetaminophen overdose? Clin Toxicol (Phila) 2010; 48:787–792. 17. Schiodt FV, Ott P, Christensen E, Bondesen S. The value of plasma acetaminophen half-life in antidote-treated acetaminophen overdosage. Clin Pharmacol Ther 2002; 71:221–225. 18. Tsai CL, Chang WT, Weng TI, Fang CC, Chen WJ. Acute acetaminophen intoxication in Taiwan: outcomes and risk factors. J Formos Med Assoc 2004; 103:830–835. 19. Senarathna SM, Ranganathan SS, Dawson AH, Buckley N, Fernandopulle BM. Management of acute paracetamol poisoning in a tertiary care hospital. Ceylon Med J 2008; 53:89–92. 20. Johnson MT, McCammon CA, Mullins ME, Halcomb SE. Evaluation of a simplified N-acetylcysteine dosing regimen for the treatment of acetaminophen toxicity. Ann Pharmacother 2011; 45:713–720. 21. Offerman SR. The clinical management of acetaminophen poisoning in a community hospital system: factors associated with hospital length of stay. J Med Toxicol 2011; 7:4–11.
Copyright of Clinical Toxicology (15563650) is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
