Background: Serum alanine and aspartate aminotransferases (ALT/AST) have been the gold standard for detection and quantification of liver injury for over 6 decades, but have relatively long half-lives (T½) (literature estimates approximately 17 and 47 h, respectively) and thus do not reflect immediate changes in liver injury or recovery. A new point-of-care immunoassay for α-glutathione S-transferase (α-GST) measures this cytosolic liver enzyme with a predicted T½ of 60–90 min based on preliminary studies and might enable earlier detection of improving or worsening liver injury than conventional enzyme testing.
Methods: Serial serum samples collected daily from 31 patients enrolled in the Acute Liver Failure Study Group, with acetaminophen (APAP) toxicity, drug-induced liver injury, ischemic hepatopathy (IH), or autoimmune hepatitis were analyzed to determine α-GST using the Qualigen FastPack® α-GST Assay (Carlsbad), a chemiluminescent immunoassay using a paramagnetic particle matrix with an upper limit of normal of 11 ng/mL. AST and ALT values were obtained from the medical record and have an upper limit of normal of 40 IU/L. The T½ values for α-GST, AST, and ALT were calculated from the peak value for APAP and IH etiologies considered as single time point injuries, using an exponential trendline equation of the serial values.
Results: Median α-GST for all etiologies were increased on day 1, returning to normal by day 3, whereas median AST and ALT values did not return to normal, even at day 7. The median T½ for α-GST, AST, and ALT were 6.4, 22.2, and 33.9 h, respectively.
Conclusions: α-GST is a more responsive marker of liver injury/recovery, allowing for more rapid real-time assessment of improvement or worsening of liver disease.
This study revisits the utility of α-glutathione S-transferase (α-GST) as a biomarker for acute liver injury in comparison to aspartate aminotransferase (AST) or alanine aminotransferase (ALT), the classic liver injury biomarkers. In situations of acute and severe liver injury, such as acetaminophen (APAP) overdose, ischemic liver injury and possibly acute cellular rejection, an enzyme with a shorter half-life (T½) might more rapidly demonstrate worsening or improving (increasing or decreasing) necrosis. We demonstrate here that the T½ of α-GST in these acute settings is approximately 6.4 h compared to 22 h for AST and 34 h for ALT. Thus, α-GST more accurately reflects the degree of hepatic necrosis at the time the test is taken.
Aspartate aminotransferase (AST),4 which has both cytosolic and mitochondrial isoforms, and alanine aminotransferase (ALT) are time-honored liver enzymes that are released from damaged hepatocytes secondary to liver injury of all varieties, including both necrosis and apoptosis. Once released from cells, their clearance from the circulation is relatively slow, occurring over several days after a self-limited liver injury. α-Glutathione S-transferase (α-GST) is a member of a family of enzymes that contain distinct subunits B1 and B2 and make up approximately 3% of cytosolic protein in the hepatocyte. Unlike AST and ALT, which are found predominantly in the periportal hepatocytes, α-GST subunits are distributed uniformly across the liver lobule. The diffuse distribution and other physicochemical properties of the B1 and B2 subunits lead to quick release of large quantities of α-GST from damaged hepatocytes into the plasma. Of all the GSTs, α-GST is the predominant isoenzyme in human hepatocytes. The plasma half-life (T½) of α-GST has been calculated in some studies at <60 min, contrasting with AST or ALT, where the T½ values are between 17 and 47 h, respectively (1, 2). Ozturk el al. (3) found α-GST to be a more sensitive and early indicator of injury compared with transaminases in various clinical settings, including acetaminophen (APAP) toxicity, birth asphyxia, and autoimmune chronic active hepatitis, as well as in liver donors and recipients during and immediately after transplantation. In another study, Trull (4) examined α-GST in a randomized controlled trial of the liver transplant population. Serum α-GST increased more consistently and rapidly than conventional liver enzymes (ALT, alkaline phosphatase, AST, bilirubin) in association with acute rejection. When treatment was given for rejection, α-GST was found to decrease toward the normal range more rapidly than conventional liver enzymes, which remained high for several days after initiation of antirejection treatment. Of note, α-GST is not increased in the setting of muscle injury, in hemolysis, or where there is extrahepatic inflammation such as rheumatoid arthritis and thus appears to be more liver specific than aminotransferases.
To determine whether α-GST might be a more responsive biomarker, we compared performance characteristics of α-GST with those of AST and ALT in a variety of settings associated with acute liver injury of at least moderate severity. While much recent attention has focused on biomarkers of injury, little attention has been paid to enzyme determinations/comparisons in recent years. Thus, our study is novel in raising the issue of whether standard enzyme measurement might be improved with addition of other serum enzyme assays.
The Acute Liver Failure Study Group has, since 1998, collected detailed clinical data and biosamples on patients admitted to study centers with severe liver injury leading to encephalopathy and coagulopathy. Available daily serum samples and laboratory data from these patients were used to perform the study.
A total of 31 patients enrolled prospectively by the US ALF Study Group were studied. Because patients were encephalopathic by definition, informed consent was obtained from their legally authorized representative. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected by a priori approval by the participating sites' institutional review boards.
All patients met the classic definition of acute liver failure: onset of encephalopathy within 26 weeks of first symptoms and coagulopathy (international normalized ratio ≥1.5) occurring in an individual without underlying chronic liver disease. Detailed data as well as serial serum samples stored at −80 °C were studied from well-characterized patients enrolled in the Acute Liver Failure Study Group (ALFSG) registry with etiologies that specifically demonstrated dramatic aminotransferase elevations and compared the performance of the 3 enzymes in each patient over the same time period.
Patients were selected based on available samples over at least 5 days from admission to the ALFSG study: 20 with APAP toxicity, 5 with ischemic hepatopathy (IH), 3 with severe drug-induced liver injury (DILI), and 3 with autoimmune hepatitis (AIH). Those in the APAP and IH categories were particularly selected for high initial enzyme values, since there were more choices available with multiple day samples. Samples are collected at time of study entry; thus, we are unable to obtain, in the ALF setting, samples closer to exposure, when the liver injury is just beginning. These sera were analyzed to determine the T½ of α-GST. We used AST and ALT values obtained from medical records at the study sites, as recorded in the ALFSG case report forms. AST and ALT upper limits of normal were considered to be equal to 40 IU/L. Although each patient had 5 serial α-GST and 7 AST and ALT values each, a minimum of 3 serial values were used to calculate the T½, and calculations were made between the peak enzyme value and the first normal value obtained or the last measured (increased) value.
Samples were tested by using the Qualigen FastPack α-GST, a chemiluminescent sandwich immunoassay that uses a paramagnetic particle matrix and has a linear range between 0.50 and 200 ng/mL (Qualigen). After thawing, 100 μL of each serum sample was introduced into the individual reagent pack port via a pipette. The pack was placed into the Qualigen FastPack analyzer, which automatically processed the sample within the pack and generated an α-GST value. The value was automatically printed out by the immunoassay analyzer. Serum samples that produced values >200 ng/mL were diluted in standard 1:50 dilution vials provided with the FastPack and reanalyzed.
The Qualigen FastPack α-GST assay has a reference interval of 0.24–11.4 ng/mL that was determined by analyzing the serum samples from 167 apparently healthy blood donors (ranging in age from 18 to 55 years) using the FastPack α-GST immunoassay in conjunction with the FastPack analyzer. The median value for these donors was 1.4 ng/mL, with the reference interval calculated at the 5th and 95th percentiles. The total coefficient of variation values of the low- and high-control standard samples were 10.6% at a mean concentration of 9.3 ng/mL and 9.5% at a concentration of 57.3 ng/mL, respectively.
Statistical analyses were performed using IBM® SPSS® Statistics 21.0 (SPSS).
The study population was not fully representative of acute liver failure patients but included more patients with severe but short interval acute (sometimes termed “hyperacute”) injury (APAP, ischemia) being overrepresented (Table 1). In addition to the 20 APAP and 5 IH patients, we chose 3 patients each with AIH and DILI as representatives of more subacute varieties of injury that still represented severe liver damage but with lower aminotransferase levels. In this small subset overall, 70% were female with a median age of 34.5 years and 90% were white. Coma grades showed 39% with a coma grade of 1, 10% coma grade 2, 19% coma grade 3, and 32% coma grade 4. Across all etiologies, the median day 1 values for all patients were 2333 IU/L for ALT and 1497 IU/L for AST, and these values were 2986 and 2890 IU/L for the APAP patients considered separately. All other results were similar except for the median bilirubin of the AIH patients, which was much higher than the other etiologies (Table 1).
As seen in Fig. 1, median α-GST values across all patients decreased to within normal limits by day 3, while median values of AST and ALT did not decrease to below the upper limits of normal by day 7. As a group, APAP patients had the highest day 1 median α-GST levels as well as the highest AST and ALT levels as expected, with the next highest values being observed in the IH group, followed by AIH (Fig. 2).
Only 14 patients (12 APAP and 2 IH) qualified for calculation of α-GST T½ over the study days, by displaying steep and consistent declines during the evaluation period. The T½ values for α-GST and for the other enzymes were calculated starting with the peak value of the 5 days (day 1 or day 2) and using an exponential trend line. Fig. 3 shows representative serial α-GST, ALT, and AST values for a single APAP patient with the corresponding trend line and exponential equation. This patient was chosen as being typical of a severe overdose who was admitted to study relatively early, before the peak of AST or ALT had occurred, a pattern we observe in approximately 15% of patients, allowing the best chance to follow the enzyme pattern over time. This patient had ingested 100 extra-strength APAP tablets between 48 and 72 h before the first sample being taken. It is noteworthy that α-GST declined immediately, whereas AST and ALT levels were still rising between days 1 and 2 in this individual. Peak α-GST and aminotransferase levels were 40250 ng/mL, 12943 IU/L, and 15925 IU/L, respectively. The T½ values for this patient were calculated to be 7.3, 9.8, and 29.0 h, respectively. Despite the very high aminotransferase levels and international normalized ratio of 8.2, she made a full recovery.
The median T½ for α-GST, AST, and ALT for the 14 α-GST evaluable patients was calculated to be 6.8, 15.4, and 29.1 h, respectively. When all evaluable ALT and AST results were examined, the median T½ values were 22.4 and 33.9 h for AST and ALT, respectively. These latter values, taking into account patients with subacute injury, are unlikely to be accurate, given the ongoing necrosis and relatively low baseline values observed, compared to the APAP or ischemic patient values.
On closer examination of the APAP patient data, 7 were considered to represent intentional overdose (suicide) and 13 cases were considered as unintentional overdoses (5). Intentional patients represent a single time point ingestion, while unintentional overdose patients take smaller amounts typically in excess of the recommended daily limit doses over a longer period of time, but present with similarly high aminotransferase levels. Of the intentional overdose patients, 5 of the 7 had calculable T½ values. The 2 patients for whom we were unable to calculate T½ values had normal α-GST values in the initial sample, representing late presentation to the study site, although AST and ALT levels were relatively high (1132 and 226 IU/L and 1095 and 2428 IU/L, respectively). In similar fashion, of the 13 unintentional APAP overdose patients, 6 had α-GST results that did not allow T½ calculation, likely because their presentation was later than suicidal patients. Of these patients, 5 of the 6 had normal to very low α-GST values (<20 ng/mL) on day 1. For unclear reasons, the α-GST values for the remaining patient rose and fell throughout the 5 days measured.
In both APAP subgroups, those for whom a T½ could not be calculated generally had longer times between the date of last dose of APAP and hospitalization or admission to study (2 days or more). These patients also had lower APAP levels on admission (0–30 mg/L). Most of the patients (8 of 12) for whom T½ could be calculated had APAP levels that were ≥40 mg/L.
Of the 5 IH patients analyzed, only 2 had serial values for which a T½ could be calculated. These 2 patients had very high day 1 α-GST values (>3000 ng/mL) and did not survive. Their day 1 samples were collected 1 or 2 days after initial hospitalization, while the other 3 patients had much lower day 1 values (<48 ng/mL), and this second group had samples collected 4–6 days after initial hospitalization. While AST and ALT levels were increased on presentation (>570 and >795 IU/L, respectively), they were lower than the initial 2 patients studied.
As expected, most patients with AIH or DILI, examples of subacute ALF, had α-GST values that were unsuitable for T½ calculations. Two DILI patients had normal α-GST values for all 5 days, with modestly increased ALT and AST values (200–300 IU/L or less). All other AIH and DILI patients had day 1 α-GST values that were either <100 ng/mL, decreasing to normal values by days 3–5, or their values fluctuated from 50 to 150 ng/mL for all 5 days of the study. While the ALT and AST values followed similar patterns, the values of both enzymes were initially elevated and remained elevated at day 7. Day 1 of the study for all of these patients ranged from 2 to 38 days after their initial hospitalization, emphasizing their slower disease evolution, with 4 of the 6 patients being enrolled in the study >6 days after their hospitalization.
Recently, there has been great interest in the development of new biomarkers that might better predict outcomes or result in earlier recognition of liver injury and its resolution (6). In this study, we compared α-GST with AST and ALT as markers of severe acute liver injury, highlighting situations based on high aminotransferase levels where T½ could most readily be calculated, i.e., APAP overdose and ischemia. In every patient studied, acute liver failure had already developed so that very early disease situations could not be observed. However, in patients whose sera were collected relatively early after injury, all 3 enzyme levels were very high and led to calculation of shorter T½ values, based on the nature of the T½ calculation. Those patients that were admitted later on during their injury frequently exhibited low or normal values of α-GST such that a T½ calculation was not possible. While our studies did not reveal an α-GST T½ as short as had been reported (60–90 min), our median α-GST T½ (6.8 h) was still a fraction of that observed for AST or ALT, with T½ values of 15.4 and 29.1 h, respectively. For all ischemia and APAP patients, declining values were observed from the beginning, and, in all, α-GST values were normal by day 3, indicating that the ongoing necrosis had completely resolved. These results resemble those observed by McGill et al. (7) for argininosuccinate synthetase. Additionally, some of the subacute injury patients with lower but sustained AST and ALT activity demonstrated normal α-GST levels for all study days. This result suggests that minimal ongoing injury that occurs in these subacute settings (DILI or AIH) is even lower than it seems, since the aminotransferase values observed reflect, in part, the slower systemic clearance of AST and particularly ALT over time.
All T½ calculations assume no further enzyme release and ongoing steady-state decay. Although APAP and ischemic liver injury best approximate this ideal situation (a one-time injury), the differences observed between individuals and within individual patient curves likely reflect some degree of ongoing enzyme release into the circulation; this is a less than ideal situation for calculation purposes. Thus, our T½ calculations should be understood to be rough approximations, indicating in a general way the relative performance characteristics of these 3 enzymes.
We are not aware of any direct comparison of performance characteristics of the various enzyme assays in recent years. We believe there is considerable variability in T½ measurement and acknowledge that the nature of our study may not allow optimal T½ calculation, since most patients were on the downslope of the curve when samples were collected. The shorter T½ of α-GST more rapidly signals the decline/end of liver necrosis. Changes in status are detected earlier, but our study design did not allow us to observe rapid elevations in α-GST reported by others (3). It has been postulated that shorter T½ enzyme assays would detect more rapidly onset of acute liver rejection, for example (4). While the median value for α-GST had fallen to normal levels by day 3 for all patients, AST and ALT did not reach normal levels for any of the study patients by day 7 of the study. Of note, in the case example in Fig. 3, both aminotransferase levels were initially rising while α-GST levels were declining sharply. In essence, no patient was identified for whom the α-GST level increased after the initial sample. All our results were taken from typical ALF patients. Thus, by the time liver failure develops, enzyme levels are generally on the decline, with some exceptions as noted for the early APAP case depicted. As markers of injury, all enzyme levels depend on the interval between initial liver injury and sample collection. Because of the short T½, time from the initial injury to sample collection was particularly important in determining α-GST. In 8 of 20 later-presenting APAP patients, T½ could not be calculated because most of the decline in α-GST had already occurred. By contrast, for AST and ALT, we were able to readily calculate T½ for 28 of 31 patients, with 3 prolonged subacute cases being the expected outliers (2 DILIs and 1 AIH). As with α-GST, the patients with the lowest day 1 AST and ALT enzyme values had the longest T½, substantiating the effect of late presentation and prolonged but limited degree of necrosis in confounding the T½ calculation. In short, if the etiology of the liver injury indicates that there is ongoing, even moderate, liver damage, it is impossible to calculate T½. We understood this when we designed the study but wanted to determine the value of α-GST in clinical decision-making; hence, we considered a variety of etiologies, all in the setting of patients reaching the degree of severity indicated by acute liver failure. Once again, T½ was best calculated assuming that the injury was due to a single time point toxic ingestion (APAP toxicity or ischemia). Of interest, recent analysis has suggested that even multiple time point APAP ingestions appear to reach a threshold where liver injury occurs abruptly, much like the single time point suicidal overdose ingestion (8).
α-GST levels might be clinically advantagous. First, the finding of normal or near-normal values early in the clinical course of illness suggests that the injury is already resolving. This result was particularly evident in the early APAP cases, where the AST and ALT values were continuing to rise although α-GST was falling precipitously, indicating the apparent end of ongoing damage. It has been argued that declining values of enzymes may indicate lack of additional hepatocytes; however, it is unlikely that this scenario occurs or if it does, it is quite infrequent. We suspect that α-GST appears earlier in the evolution of injury but did not have sera collected early enough to convincingly demonstrate this.
Where active treatments such as antirejection medications are administered, improvement in α-GST would be likely to occur within hours and thus might indicate a positive response to treatment, much earlier than AST or ALT. A much larger study would be necessary to determine whether measurement of α-GST might aid in prediction of outcome after APAP or ischemic injury; it is unlikely to be of value in ongoing chronic liver injury settings but might indicate response to therapy in AIH or subsiding damage in the DILI setting. An algorithm for prediction of APAP recovery might be developed using the early time point data for α-GST (8, 9). Determining whether α-GST might be a better indicator of APAP dose ingested (or prognosis) compared to AST or ALT would be of interest, since accurate dosing information is rarely available (10). In both circumstances, patients experience effects of overdose hours after ingestion, but ALT and AST have delayed onset of elevation and delayed decline of several days and do not reflect the degree of injury except in retrospect. α-GST increases earlier, peaks hours before AST and ALT, and normalizes much sooner (1).
The reason this enzyme displays a shorter T½ is not well understood. It is possible that this difference reflects its ubiquity across the liver lobule. Injury due to either ischemia or APAP occurs predominantly in the centrilobular region (zone 3), while aminotransferases are found most abundantly in the periportal (zone 1) region. Perhaps AST and ALT, since they are largely found at a distance from the injury, are actually less directly affected by the toxic/necrotic process. It is unclear why the T½ calculated in our study is longer than that reported in studies performed nearly 20 years ago.
Limitations of the study included the relatively small number of patients examined to date and the use of existing AST and ALT values from different but standardized clinical chemistry laboratories around the US. All samples were obtained on first-morning blood draws but the exact timing of each sample was not known. Our general findings confirm past reports that α-GST is a more responsive biomarker that may more accurately reflect the state of injury at a given moment, rising more rapidly in the presence of injury and declining when injury is over. A larger controlled side-by-side trial would be necessary to fully discern its value across a wider variety of liver injury settings and establish its utility in the point-of-care setting. The FastPack IP System is available in the US, Europe, and other countries but is not currently not approved by the Food and Drug Administration.
In summary, α-GST is an interesting biomarker that may reflect more readily the dynamic changes in liver injury over time. The assay is inexpensive and easy to perform and is currently available as a point-of-care assay. Aminotransferases by comparison are standard and well understood by clinicians worldwide and are likely to remain the day-to-day mainstay for some time, although they may not reflect the degree of liver injury as effectively as α-GST does in real time.
The authors acknowledge the help of the NIDDK Biosample Repository in storing and triaging the samples for this study.
↵4 Nonstandard abbreviations:
- aspartate aminotransferase
- alanine aminotransferase
- α-glutathione S-transferase
- Acute Liver Failure Study Group
- ischemic hepatopathy
- drug-induced liver injury
- autoimmune hepatitis.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Employment or Leadership: M. Poirier, Qualigen.
Consultant or Advisory Role: None declared.
Stock Ownership: M. Poirier, Qualigen.
Honoraria: None declared.
Research Funding: National Institutes of Health grant U-01 58369 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), an American Gastroenterological Association Student Research Fellowship Award, and Qualigen.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
- Received April 12, 2016.
- Accepted July 1, 2016.
- © 2016 American Association for Clinical Chemistry