Liver Injury Associated With Drugs and Complementary and Alternative Medicines in India

Idiosyncratic drug-induced liver injury (DILI) is an important differential diagnosis in patients with abnormal liver tests.¹ The presentation is acute in the vast majority of cases, and the diagnosis is challenging. DILI is the most common cause of acute liver failure in the US² and Europe,³ but in India, viral hepatitis is the most common cause, followed by liver injury due to antituberculosis (anti-TB) therapy.⁴ During the last two decades, interest in the study of DILI has increased considerably. Results from several large cohorts of patients with DILI have been reported from Western countries mostly from Spain, Sweden, Iceland, and the US^{5, 6, 7, 8, 9, 10} but also from the East such as from Korea,¹¹ India,^{12, 13, 14} and China.¹⁵

Previous studies from India have been from single centers with a relatively limited number of patients.^{12, 13, 14,16} It is therefore of great importance for the DILI community and for research in liver disease that results from a prospective study of 1288 patients with well-characterized DILI are reported in the current issue of the journal from the great country of India.¹⁷

This was a multicenter study over 5 years, with participating centers over large parts of India. It seems that the vast majority were teaching hospitals, using the same protocol, and the nodal center was St. John's Medical College Hospital in Bangalore. Criteria for the diagnosis of DILI and assessment of severity were based on and adopted from the criteria from the International DILI Expert Working Group.¹⁸ Causality was based on the RUCAM causality assessment method.¹⁹

Patients had relatively severe liver injury as almost 70% presented with jaundice, as might be expected in patients seen in tertiary referral centers. In contrast, in the population-based cohort of DILI patients, only approximately 30% presented with jaundice.^8,20 The overall mortality in the Indian study was 12.3% and higher in those with jaundice. An unfavorable prognosis was independently associated with high INR and creatinine as well as ascites and hepatic encephalopathy.

The most common drugs leading to DILI were the combination of anti-TB drugs in 46% of cases, complementary and alternative medicines(CAMs; 14%), antiepileptic drugs (8%), and non-TB antimicrobials (6.5%).

The results of this large Indian study are very interesting and convincing. The authors are to be congratulated for their efforts. The methodology is sound and in accordance with the highest standard in the field.

In many ways, the results are similar to those of other prospective studies of DILI published in the West.^{5, 6, 7, 8} The prognosis was similar, although patients who fulfilled Hy's law had a somewhat higher mortality of approximately 17%, but in most of the other studies, this has been around 10–12%;^{5, 6, 7, 8} but this is probably related to more severe liver disease at presentation. Interestingly, ascites was present in around 43% of nonsurvivors and in 12.5% of survivors, whereas this has rarely been reported in other DILI cohorts.^{5, 6, 7, 8} Ascites in this context is most likely due to acute liver failure, although it is not entirely clear from the article how many had preexisting chronic liver disease as this could be a manifestation of decompensated chronic liver disease.

The etiology was somewhat different in the Indian study owing to the predominance of the combination of anti-TB drugs in almost 50% of cases. This reflects that TB is still a major health issue in India and also the hepatotoxicity potential of current anti-TB drugs. In fact, anti-TB drugs have been the second most common cause of DILI in the Spanish hepatotoxicity registry,⁵ and isoniazid is also the second most common cause in the DILIN study,⁷ although of much less magnitude than in India. In contrast to the West, where non-TB antimicrobials are the dominant etiology, this was relatively rare in the Indian cohort. Comparison between the results of the present study with those of prospective DILI studies undertaken in Spain, the US, and Iceland, in terms of demographics and the main etiologies of DILI, is shown in Table 1. CAMs were the second most common cause of liver injury as in previous studies from the US and Iceland.^7,8

Table 1.

The Most Common Implicated Agents Causing DILI in Four Prospective Studies on DILI.

India (n = 1288) Mean age (years): 43 Male (%): 51.4%	Spain (n = 843) Age (years): 53 Male (%): 51%	USA (n = 899) Age (years): 49 Male (%): 41%)	Iceland (n = 96) Age (years): 55 Male (%): 44%
Antituberculosis drugs (46%)	Amoxicillin-clavulanate (22%)	Amoxicillin-clavulanate (10%)	Amoxicillin-clavulanate (22%)
Antiepileptics (8.1%)	Antituberculosis (4.5%)	Isoniazid (5.3%)	Diclofenac (6.3%)
Antibiotics (non-TB) (6.5%)	Ibuprofen (3%)	Nitrofurantoin (4.7%)	Nitrofurantoin (4%)
Antimetabolites (3.8%)	Flutamide (2.6%)	Sulfam-trimeth (3.4%)	Azathioprine (4%)
Antiretrovirals (3.5%)	Atorvastatin (1.9%)	Minocycline (3.1%)	Infliximab (4%)
NSAIDs (2.6%)	Diclofenac (1.8%)	Cefazolin (2.2%)	Isotretinoin (3%)
Hormones (2.5%)	Ticlopidine (1.4%)	Azithromycin (2%)	Atorvastatin (2%)
Statins (1.3%)	Azathioprine (1.3%)	Ciprofloxacin (1.8%)	Doxycycline (2%)
Others (11.3%)	Fluvastatin (1.3%)	Levofloxacin (1.4%)	Imatinib (1%)
	Simvastatin (1.3%)	Diclofenac (1.3%)	Isoniazid (1%)
CAM (13.9%)	CAM (3.4%)	CAM (16.1%)	CAM (16%)

DILI = drug-induced liver injury; Sulfam-trimeth = Sulfamethoxazole-trimethoprim; CAM = complementary and herbal supplements; TB = tuberculosis.

Although CAMs are commonly used in India, the proportion of patients with CAM-induced liver injury was not higher than that in series from the West. A total of 14% patients had liver injury due to CAM in the Indian study, which was very similar to the 16% frequency in the US and Iceland.^7,8 As pointed out by the authors, there was a male predominance among patients with CAM-induced liver injury, in contrast with most previous studies from the West, showing more women having liver injury from CAMs. In the West, the indication of supplements found to have caused liver injury was frequently weight loss.^21,22 Apparently, it was difficult to tease out the indication for CAMs in India, and in very few instances was the prescription of the CAM traced. Therefore, the information was lacking on the indication, and it was very difficult to detect and characterize the ingredients. However, it is unlikely that CAMs were used to induce weight loss in the present study as the mean body mass index was normal in these patients, although this might be associated with severe liver injury. Anabolic steroids are also relatively a common cause of liver injury among nonprescription drugs in the West,²¹ but not a single case due to anabolic steroids was reported in the article from India.

One of the interesting observations in the Indian study was the relatively low age of patients with DILI, who only had a mean age of 42 years, which is lower than that in studies from the West. It is acknowledged by the authors that this is probably reflecting the young population of India, with a median age of only approximately 28 years.

Although direct toxicity was not a focus of the study and paracetamol hepatotoxicity was excluded, only 10 cases were found, and the authors suggested this as a marker of very low frequency of paracetamol intoxications in India, in contrast with studies from the West, wherein paracetamol is the most common cause of acute liver failure.^2,3 However, it is not entirely clear if results would have been the same if the study aims had been to investigate the etiology of acute liver failure in India.

Hypersensitivity skin reactions were observed in 19.4% of the Indian patients with DILI. This is a much higher frequency than that in previous DILI studies. In the DILIN study, only 9 of 899 (1%) presented with severe cutaneous reactions, Steven–Johnson syndrome, or toxic epidermal necrolysis.⁹ However, hypersensitivity reactions are not always severe, so it is difficult to compare these two studies in that respect. The reason for the high frequency of skin reactions in almost 20% of the patients is unclear. It is conceivable that Indian doctors were more thorough in their physical examination than their colleagues in the studies from the West.^{5, 6, 7, 8, 9} However, another reason might be the higher frequency of DILI due to antiepileptics than in other DILI studies, and these medications are commonly associated with cutaneous reactions.²³ Indeed, skin rash was observed in 61% of those who suffered from DILI owing to antiepileptic drugs.

As pointed out previously, markers of severe liver disease were predictors of mortality such as a high INR and creatinine as well as ascites and encephalopathy. The Model for End-Stage Liver Disease (MELD) was shown to be a relatively good predictor of outcomes, yielded a C-statistics of 0.811. This is in agreement with a Korean study, showing the MELD to be a useful predictor of outcomes in patients with DILI.²⁴

In summary, a large and impressive DILI study has come from the second largest country in the world, which is predicted to be the largest country within a few years. The results of the study do increase the knowledge of DILI in general and represent an important contribution to the study of DILI.

Conflicts of interest

The author has none to declare.