Single-Centre Experience With Low-Volume Plasma Exchange and Low-Dose Steroid to Treat Patients With Idiosyncratic Drug-Induced Acute Liver Failure

Abstract

Background

Idiosyncratic drug-induced liver injury (iDILI) causing acute liver failure (ALF) carries high short-term mortality and patients who meet King's College criteria for liver transplantation have 1-month survival of 34% without liver transplantation (PMID: 20949552). We present our experience with low-volume plasma exchange (PLEX-LV, 50% of estimated plasma volume exchanged per session) and low-dose steroid to treat iDILI ALF.

Methods

We retrospectively analysed data of patients with iDILI (diagnosed as per RUCAM score), treated with PLEX-LV and low-dose steroid (prednisolone: 10 mg OD, with rapid taper) in our department from 2016 to 2022. Baseline and dynamic parameters (post-PLEX) were assessed as predictors of 1-month liver transplantation-free survival.

Results

Twenty-two iDILI patients [probable: possible iDILI: 20:2, males: 9, age: 30 (14–84) years, median (range); MELD score: 30.5 (19–43)] underwent PLEX-LV for ALF during the study period. Causative agents were complementary and alternative medications (36%), antiepileptics (18%) antimicrobials (14%), antitubercular drugs (14%), antifungal drugs (9%) and others (9%). All patients had jaundice and encephalopathy; 9 patients also had ascites. None of the patients underwent liver transplantation. Study patients underwent 3 (1–7) PLEX sessions and 1.4 (0.6–1.6) litres of plasma was exchanged per session. One-month transplant-free survival was 59% (13/22) in the study population and 63% (12/19) among patients who fulfilled Kings College criteria for liver transplantation. Reduction of ≥25% in plasma von Willebrand factor (VWF) levels after PLEX-LV predicted improved survival (HR: 0.09, 95% CI: 0.01–0.65; AUROC: 0.81; 95% CI: 0.6–1.0).

Conclusion

Low-volume PLEX and low-dose steroid appears a promising treatment option in patients with iDILI-induced ALF not opting for liver transplantation. Dynamic changes in VWF level after PLEX predict 1-month survival in these patients.

Graphical abstract

Acute liver failure (ALF) is a rare medical condition in which patients experience rapid decline in liver function, resulting in altered mental status and coagulopathy in previously healthy individuals.¹ Idiosyncratic drug-induced liver injury (iDILI) is a relatively common cause of ALF across the world, as also in India, with up to 15% cases attributed to it.²

Despite improvement in critical care management, iDILI-ALF continues to have poor prognosis with large multi-center series reporting transplant-free survival of 27%–35% over 3 weeks to 3 months.^3,4

Management is mostly supportive and urgent liver transplantation is the only definitive treatment available. Urgent liver transplantation is often not feasible in resource constrained settings. In a multi-center study from Tamil Nadu, of 451 patients with rodenticidal hepatotoxicity, 159 patients (35%) died, but only one patient could access urgent liver transplantation.⁵

Plasma exchange (PLEX) has been shown to improve survival in patients with ALF.^6,7 While studies have used various volumes of plasma exchanged to treat ALF, we have reported favourable outcomes with low-volume PLEX (PLEX-LV) in both ALF and acute on chronic liver failure patients.^{8, 9, 10}

This study aims to analyse our experience of utilising PLEX-LV and low-dose steroid to treat patients with iDILI-ALF.

Methods

From a prospectively maintained database of patients who underwent PLEX to treat liver disease between 2016 and 2022 in our department, we retrospectively recruited patients with iDILI-ALF for the present study.

Patient Recruitment

ALF was defined as jaundice, in a patient without pre-existing cirrhosis, complicated within a 26-week period by coagulopathy (prothrombin time-international normalised ratio, PT-INR ≥1.5), and any degree of mental alteration (encephalopathy).¹¹ ALF was subdivided in hyperacute, acute, and subacute based on jaundice to encephalopathy time (hyperacute ALF: <10 days, acute ALF: 10–30 days and subacute hepatic failure: 5–24 weeks).¹²

Roussel UCLAF Causality Assessment Method (RUCAM) score¹³ was utilised for diagnosis of iDILI and classifying the event as probable, possible, or definite iDILI. Patterns of liver injury were classified as hepatocellular, cholestatic and mixed based on R factor.¹⁴ In brief, iDILI diagnosis was based on use of potentially hepatotoxic drugs prior to onset of jaundice and absence of other clinically identifiable triggers (viral hepatitis, Wilson's disease, autoimmune hepatitis, etc.). Patients with alternative aetiology of ALF and with liver disease syndromes other than ALF were excluded. Organ dysfunction and failure were defined based on Sequential Organ Failure Assessment (SOFA) Score.¹⁵

Patients were managed in high dependence unit or intensive care unit with standard medical treatment and supportive measures including cessation of causative hepatotoxic drugs, IV N-Acetyl Cysteine infusion and anti-cerebral oedema measures as needed. All patient families were counselled about need for urgent liver transplantation, however none opted for this. PLEX-LV was offered as an adjunct therapy for patients who did not opt for liver transplantation. None of the patients received fecal microbiota transplantation or granulocyte colony stimulating factor.

Details of patient follow-up were extracted from hospital records and by telephone contact.

PLEX-LV Protocol

Study patients were treated as per our published protocol to treat liver failure.^16,17 Plasma volume was calculated based on the formula 0.065∗bodyweight∗(1- haematocrit).¹⁸ In brief, all study patients underwent PLEX-LV (50% of estimated plasma volume removed at each session and replaced by an equal volume of fresh frozen plasma). All patients had ultrasound guided central line access inserted for PLEX. Three sessions of PLEX were initially targeted in each patient, which were done on consecutive or on alternate days. The number of PLEX sessions was decided based on the daily assessment of clinical response and the patient's condition. Calcium supplementation (tablet calcium carbonate – 1 g was given prior to initiation of each session of PLEX) and injection Calcium gluconate 10%, 20 ml in 100 ml normal saline given IV over 30 min was given at the completion of each PLEX session. Patients had surveillance blood cultures and were given prophylactic antibiotics starting one day prior to PLEX initiation and continued for at least 2 days after completion of PLEX sessions. Low-dose steroid (Prednisolone 10 mg PO once daily) was given from 1 day before initiation of PLEX (for at least 2–4 weeks) to ameliorate the cytokine storm seen in ALF patients, while avoiding the risk of sepsis with higher doses of prednisolone.⁹

Liver Disease Severity Assessment

King's college criteria for non-acetaminophen induced liver injury were used to assess the need to list for urgent liver transplantation.¹⁹ Additionally, fulfilment of Hy's law²⁰ was also noted prior to initiation of PLEX-LV.

All patients underwent laboratory tests at baseline (prior to initiation of PLEX) and 12-h after completion of a PLEX session. Liver disease severity was assessed by model for end-stage liver disease (MELD) score, SOFA score and plasma von Willebrand factor (VWF) antigen levels.

Outcome Assessment

Primary study outcome was liver transplantation-free survival at one-month. Secondary study outcomes were-improvement in hepatic encephalopathy and other liver disease severity parameters after PLEX. We analysed delta (Δ) changes in following parameters: encephalopathy grade, SOFA score, MELD score, serum total bilirubin, INR and VWF levels after 3rd session or at the end of last session if <3 sessions (P3) vis-à-vis baseline (P0).

We also analysed baseline parameters-age, type of inciting agent (complementary and alternative medication v/s allopathy medication), serum bilirubin, MELD score, SOFA score and plasma VWF antigen levels and delta change in MELD score, SOFA score and plasma VWF antigen levels as predictors of response to PLEX-LV + low-dose steroid to treat iDILI-ALF.

Statistical Analysis

Data were entered in MS Excel v.2020 and analysed by SPSS 22.0 statistical software (IBM Corp., New York). Continuous variables were expressed as median (range). Interquartile range (IQR) was mentioned in case of significant outliers. Paired and unpaired non-parametric tests were used to compare groups. Logistic regression was used to assess predictors of 1-month mortality. Receiver operating curve (ROC) analysis was used to derive an optimal cut-off for independent predictor variable of interest. P-value of less than 0.05 was regarded as statistically significant. The study was approved by the institutional review board and ethics committee.

Results

Baseline Characteristics of the Study Cohort

During the study period, of the 403 patients who underwent PLEX to treat liver diseases in our department, 101 patients had ALF and of these 22 (22%) had iDILI (Figure 1). The twenty-two iDILI patients were 30 (14–84), median (range) years old and 59% were females.

Figure 1.

Flow diagram of patients with idiosyncratic drug (iDILI)-induced acute liver failure (ALF) treated with low-volume plasma exchange (PLEX-LV) and low-dose steroid included in this study.

Table 1 summarises the baseline characteristics and laboratory parameters in these patients. Causative agents were complementary and alternative medications (CAM) (8, 36%), antimicrobials (5, 23%), antiepileptics (4, 18%), antitubercular drugs (3, 14%), and others (2, 9%). All patients tested negative for viral hepatitis (A, B, C and E) serology. Based on R factor, 11 (50%) patients were classified as hepatocellular, 4 (18%) as mixed and 7 (32%) as cholestatic type of DILI.

Table 1.

Demographics and Baseline Parameters of iDILI ALF Patients Treated With Low-Volume Plasma Exchange (PLEX-LV) and Low-Dose Steroid.

Parameters	iDILI ALF patients (n = 22)
Age (years)	30 (14–84)
Male: Female	9:13
Presentation (encephalopathy: ascites)	22:9
RUCAM Score (probable: possible DILI)	20:2
Causative drug (CAM: Anti-microbialsa:Antiepilepticsb: Antitubercular drugs: Othersc)	8:5:4:3:2
Hepatocellular: Mixed: Cholestatic DILI	11:4:7
Serum bilirubin (mg/dl)	21.3 (2–48.1)
PT-INR	2.85 (1.51–10.0)
Serum creatinine (mg/dl)	0.9 (0.36–4.8)
Plasma von Willebrand factor (IU/l)	411.3 (242–1214)
MELD score	30.5 (19–43)
SOFA score	8 (5–15)

All values are expressed as median (range) or numbers.

CAM, Complementary and Alternative Medications (individual component is not identified); DILI, Drug-induced liver injury; MELD, Model for end-stage liver disease; PT-INR, Prothrombin Time and International Normalised Ratio; RUCAM, Roussel UCLAF Causality Assessment Method; SOFA, Sequential organ failure assessment.

^aAmoxicillin-clavulanic acid – 3, Voriconazole – 1, Itraconazole – 1.

^bSodium Valproate – 2, Phenytoin – 1, Carbamazepine – 1.

^cNorethisterone – 2.

All the patients presented with jaundice and encephalopathy while 9 subacute liver failure patients (40%) also had ascites. These 9 patients who had ascites were evaluated to rule out other aetiologies and evidence of chronicity. Six of these patients underwent liver biopsy, which did not show cirrhosis/advanced fibrosis. Ultrasound abdomen did not show evidence of chronic liver disease viz altered echotexture, surface irregularity and coarse margins, Upper GI endoscopy done in 5 patients did not show varices.

Hepatic encephalopathy ≥grade II was noted in 17 (74%) patients (9 were in grade 2, 6 in grade III and 2 in grade IV encephalopathy). Jaundice–encephalopathy time interval was 14 (5–62) median (IQR) days. Nine patients had hyper-acute liver failure, 5 had acute liver failure and 9 had sub-acute liver failure.

Trans-jugular liver biopsies done in seven patients showed necrosis (moderate: 4, pan-acinar/sub massive: 3) in absence of advanced/established fibrosis (mild fibrosis: 5, absent fibrosis: 2), micro vesicular steatosis (2 patients) and reticulin collapse (4 patients) (Supplementary Table 1).

Ten (91%) of the 11 patients with hepatocellular type of iDILI fulfilled criteria for Hy's law. Nineteen patients (86%) fulfilled King's College criteria for urgent liver transplantation with following individual components: three patients had PT-INR >6.5, 5 had PT-INR >3.5, five were >40 years old and 13 had jaundice-to-encephalopathy interval >7 days.

13/22 patients (59%) fulfilled criteria for systemic inflammatory response syndrome showing the presence of systemic inflammation.²¹

Details of Plasma Exchange and Low Dose Steroid

None of the 22 study patients underwent liver transplantation. Study patients underwent 3 (1–6) PLEX sessions and the plasma volume exchanged was 1.4 (0.6–1.6) litres per session. Low-dose steroid (prednisolone 10 mg) started a day prior to PLEX was tapered over 3 (1–8) median (IQR) days. One patient had continuous veno-venous hemofiltration along with PLEX.

Primary Outcome

In 22 iDILI-ALF patients treated with PLEX-LV and low-dose steroid, one-month liver transplantation-free survival was 59% (13/22 patients). Of the 19 patients who fulfilled Kings College criteria for urgent liver transplantation one-month transplant-free survival was 63% (12/19 patients). One-month survival rate was 36%, 75%, and 86% among the hepatocellular, mixed, and cholestatic DILI patients, respectively. Of the 10 patients with hepatocellular iDILI who fulfilled Hy's law, one-month survival was 40%.

Table 2 lists individual organ failure at the time of presentation in the study patients. Out of 9 patients who had circulatory failure all had grade 1 circulatory failure (mean arterial pressure < 70 mm of Hg). Higher mortality was noted in patients with respiratory failure and more so in patients who had mechanical ventilation (P-value: 0.09).

Table 2.

Organ Failure at Baseline in 22 Idiosyncratic Drug-Induced ALF Patients Managed With Low-Volume Plasma Exchange and Low-Dose Steroid.

Organ failurea	Survivors (n = 13)	Non-survivors (n = 9)	P value
Liver	13 (100%)	9 (100%)	1
Neurological	13 (100%)	9 (100%)	1.0
Coagulation	9 (69%)	4 (45%)	0.39
Respiratoryb	5 (39%)	7 (78%)	0.09
Circulatory	5 (39%)	4 (45%)	1
Renal	4 (31%)	3 (33%)	1.0

^aOrgan Failure is defined as per Sequential Organ Failure Assessment (SOFA).¹³

^b1/13 and 5/9 patients were on mechanical ventilation in survivor and non-survivor groups respectively (P value: 0.023).

Secondary Outcomes

All 22 patients had hepatic encephalopathy (5 were in grade I, 9 were in grade II, 6 in grade III and 2 in grade IV encephalopathy). None of the 22 study patients had improvement in grade of encephalopathy prior to initiation of PLEX. After PLEX-LV and low-dose steroid, improvement in encephalopathy grade by ≥ 1 was seen after 1st session of PLEX in 14 patients and after 2nd session of PLEX in additional 3 patients. Five patients had worsening or no improvement in the grade of encephalopathy after 3 sessions of PLEX (Figure 2).

Figure 2.

Change in encephalopathy grades in 22 patients with idiosyncratic drug-induced acute liver failure treated with low-volume PLEX and low-dose steroid.

Most laboratory parameters showed immediate significant improvement at 12 h after PLEX- ΔINR (P0: 2.85 (1.51–10) v/s P3: 2.13 (0.85–4.32), P-value: 0.0009); Δserum bilirubin (P0: 21.27 (2–48) mg/dl v/s P3: 16.1 (1.6–36.9) mg/dl, P-value <0.26); Δ MELD score (P0: 30.5 (19–43) v/s P3: 27 (7–43), P-value <0.003); ΔVWF antigen (P0: 411.2 (242–1214) IU/l v/s P3: 290 (129–594.8) IU/l, P-value <0.0006) and ΔSOFA score (P0: 8 (5–15) v/s P3: 7 (1–15), P <0.064).

Safety Outcome

None of the patients required fresh frozen plasma or platelet transfusion for line insertion. One patient had ooze from the port insertion site, which was managed conservatively.

The PLEX procedure was well-tolerated. There were no serious adverse events attributable to PLEX.

Mild decrease in platelet counts was noted in the 22 study patients (P0: 137000, 15000–581000 v/s P3: 115000, 15000–564000, P-value: 0.016). There was no bleeding manifestation due to thrombocytopenia.

Predictors of 1-month Transplant-Free Survival in iDILI–ALF Patients

Table 3 depicts baseline parameters and dynamic changes in parameters (change expressed as percentage of baseline) during PLEX-LV, as predictors of 1-month outcome in these patients.

Table 3.

Predictors of Survival in 22 Idiosyncratic Drug induced ALF Patients Treated With Low-Volume Plasma Exchange (PLEX-LV) and Low-Dose Steroid.

Parameters		Survivors (n = 13)	Non-survivors (n = 9)	Univariate
				HR	95% CI	P value
Baseline	Age (years)	33 (14–60)	26 (17–84)	1.0	0.94–1.04	0.82
	CAM	7 (54%)	1 (11%)	0.107	0.10–1.1	0.06
	Bilirubin (mg/dl)	22.2 (5.7–48.1)	19.8 (5.5–27.5)	0.94	0.86–1.0	0.28
	INR	2.4 (1.51–10)	3.8 (2–10)	1.32	0.86–2.1	0.21
	Creatinine (mg/dl)	0.8 (0.36–4.83)	1.08 (0.51–2.45)	0.92	0.38–2.2	0.86
	Plasma VWF (IU/L)	424.6 (292–1214)	398 (242.6–634.1)	0.98	0.99–1.0	0.15
	MELD score	30 (19–34)	33 (24–43)	1.21	0.98–1.51	0.08
	SOFA score	8 (5–12)	9 (6–15)	1.42	0.92–2.29	0.10
Dynamica	Δ bilirubin	25.2	−7	0.29	0.07–1.2	0.09
	Δ VWF	48.9	7.6	0.003	0.0–0.63	0.03
	Δ MELD score	15.6	2.7	0.10	0.001–7.65	0.29

All parameters are expressed as median (range) or number (%).

Δ, delta; CAM, Complementary and Alternative Medications; CI, Confidence Interval; HR, Hazards Ratio; MELD, Model for End-stage Liver Disease; SOFA, Sequential Organ Failure Assessment; VWF, Von Willebrand factor.

^aFormula: $\left[Baselinevalue\left(P0\right)-Post3rdPLEXsessionvalue\left(P3\right)\right]/Baselinevalue\left(P0\right)\times 100\%$.

ΔVWF was the only significant predictor of outcome after PLEX-LV.

Area under ROC for ΔVWF was 0.81 (95% CI: 0.6–1.0) for predicting survival (Supplementary Figure 1). ≥25% fall in plasma VWF levels from baseline predicted significantly better survival (HR: 0.09, 95% CI: 0.01–0.65, P-value: 0.02, sensitivity: 77%, specificity: 78%).

Follow Up

Patient were followed up physically and by telephonic means wherever physical visit was not possible. 90 days survival in total cohort was 55% (12/22 patients).

Of 9 patients with subacute liver failure 8 patients (89%) were alive at 90 days. Serum bilirubin at the end of 90 days was 1.6 (0.49–6.8) mg/dl and INR was 1.01 (0.64–1.09) (in 7 patients who had physical follow up at 90 days).

Discussion

In this retrospective study comprising patients with iDILI ALF (n = 22), we report a 1-month transplant-free survival of 59% with PLEX-LV and low-dose steroid treatment. Post PLEX dynamic changes in plasma VWF levels (reduction by ≥25% from baseline value) was an independent predictor of 1-month survival in these patients.

Sub-analysis of US-ALF study group data reported a 27% transplant-free survival in iDILI at 3 weeks.³ Similar dismal survival rate was noted in a multi-centre study from India.⁴ Table 4 compares the present study to other published studies of iDILI-ALF patients in terms of patient selection, disease severity and transplant-free survival.^{2, 3, 4} All studies differ in the implicated drugs and proportion of CAM usage (relatively high CAM use is noted in Indian sub-continent). Despite comparable liver disease severity (similar MELD scores), the current study, utilising PLEX-LV and low-dose steroid demonstrates significantly better overall short-term transplant-free survival as compared to other large cohorts of iDILI-ALF patients, Table 4.

Table 4.

Short-term Liver Transplantation-Free Survival in Patients With iDILI Induced ALF Who Underwent PLEX-LV (n = 22) Compared With Other Large Published Series.

Parameters		Present study (2016–2022)	Harshad et al.2 (1997–2018)	INDILI4 (2013–2018)	USALF3 (1998–2010)
		n = 22	n = 128	n = 124	n = 133
Age (years)		34 ± 17	38 ± 19	49 ± 16.2	43.8 ± 14.1
Gender: Female		13 (59%)	68 (53%)	71 (57%)	94 (71%)
Inclusion	Idiosyncratic DILI	Y	Y	Y	Y
	Acute liver failure	Y	Y	Y	Y
Severity	Fulfilling King's College criteria	19 (86%)	54 (42%)	Not mentioned	Not mentioned
	MELD score	30.5 ± 5.7	30 ± 11.7∗	28.5 ± 10∗∗	33 ± 9.2∗∗∗
Treatment given		SMT + PLEX + steroid	SMT	Not mentioned	SMT + liver transplantation
Short-term transplant-free survival	In all patients	At 1 month	At 70 daysa	At 3 months	At 21 days
		59% (13/22)	34% (43/128)€	35% (45/124)#	27% (36/133)¶
	In patients fulfilling King's College criteria	63% (12/19)	39% (21/54)¥	Not Mentioned	Not Mentioned

DILI, Drug-induced liver injury; INDILI, Indian Network of Drug-Induced Liver Injury; PLEX, Plasma exchange; SMT, Standard medical therapy; USALF, US Acute Liver Failure Study Group.

All parameters are expressed as median (range) and (standard deviation) or number (%).

On comparing with the present study, P values were as follows.

∗P-value = 0.5 ∗∗P-value: 0.3 ∗∗∗P-value: 0.2.

^€P value = 0.03^#P value = 0.05^¶P value = 0.005^¥P value = 0.11.

^aMean duration of treatment.

Paracetamol-induced ALF has a significantly better survival in the US-ALF study group data 64–66%,³ probably secondary to early recognition and effective supportive management. IV N-acetyl cysteine also serves as a specific antidote for paracetamol overdose. In India, paracetamol overdose contributes only to a small proportion of all DILI-ALF.⁴ In the current study we included patients with idiosyncratic (not dose dependent) drug-induced liver injury only.

Unlike other studies, where baseline factors like age, total bilirubin, and MELD score predicted survival in iDILI ALF patients,^{2, 3, 4,19,22} in our study none of the baseline factors predicted survival. This may be secondary to limited sample size, but also suggests that PLEX-LV partially negates the effect of conventional severity parameters.

Plasma VWF antigen levels are a good predictor of short-term outcome in patients with ALF.²³ It has been postulated that PLEX-LV improves survival in ALF patients by helping to remove inflammatory debris (esp. macromolecules e.g., VWF) and possibly thus improving liver micro-circulation.²⁴ In this study, although baseline VWF did not predict survival, ≥25% drop in VWF after PLEX was an independent predictor of survival.

This study is limited by small sample size, retrospective nature, lack of control arm and heterogeneity of implicated drugs. Plasma ammonia levels were not systematically analysed in study patients. CAM intake provides a unique challenge to decipher the active components and needs to be further studied.

In conclusion, low-volume PLEX and low-dose steroid appears a promising treatment option in patients with iDILI induced ALF not opting for liver transplantation. Assessment of post PLEX-LV dynamic changes in plasma VWF antigen levels predict eventual outcome in these patients.

Credit authorship contribution statement

Conceptualized and designed the study, collected and analysed data, reviewed literature and prepared initial draft of the manuscript: KAS, SEK, UGZ, DD, VD, SK, KP, EJ, TK, CEE and AG.

Data analysis and manuscript preparation: KAS, UZ, CEE and AG.

All authors critically revised and approved the final version of the manuscript.

All authors have agreed to be accountable for the accuracy and integrity of the work.

Conflicts of interest

All authors have none to declare.

Acknowledgements

Nil.

Funding

None.

Ethics approval statement

IRB no -13113 dt.24/06/2020.

Appendix A. Supplementary data

The following are the Supplementary data to this article: