At present, the report about the risk factors of TILI in ICU patients was rare. This study retrospectively analyzed the changes of liver function in adult ICU inpatients with TGC treatment and the risk factors of TILI. It was observed the overall levels of ALP, TBIL and TBA, but not ALT or AST increased significantly, similar findings were also revealed in severely infected patients [13]. However, Chen [14] et al. observed that levels of ALT and TBIL increased in ICU patients with Ventilator-associated Pneumonia (VAP). The incidence of abnormal increase of liver enzyme in this study was close to a previous study [15] including 92.7% ICU patients (ALT 13.22% vs.18.8%, AST 28.51% vs. 25.0%), but the incidence of abnormal TBIL was reported lower than our study (6.7%vs. 59.09%). Another study [16] showed that TGC-associated serum transaminase elevations occurred in just 2%-5% of patients, which was lower than our study, suggesting that the incidence of abnormal liver enzyme level in ICU patients might be higher than that in general patients.
In the present study, 24 (9.92%) patients treated with TGC experienced TILI. The incidence is within the range reported in other studies [14, 17]. TILI severity graded as 1 to 2 were found in most of the patients. The incidence (0.41% of 242) of grade 4 (liver failure) is lower than that reported in a study including ICU patients with co-infection (1.20%) [18]. TILI is mostly cholestatic, similar to the results of Zhu et al. (62.50%) [19]. The mechanism of liver injury induced by TGC has not been clarified yet. It may be related to the increase of IL-18 level and hepatocyte apoptosis induced by TGC [20]. The concentration of TGC in bile is much higher than that in serum, and its prototype and metabolites are mainly excreted through the bile duct, which may cause cholestatic liver injury [20, 21].
Several studies [14, 22, 23] revealed that there was no significant difference in the incidence of TILI between high-dose and standard dose of TGC. However, a maintenance dose ≥ 200 mg/day was associated with TILI in our study. In addition, the extension of the duration of TGC treatment would also increase the risk of TILI. A study [14] showed that the latest occurrence time of TILI in VAP patients in ICU was (10.28 ± 6.25) days. The median time of TILI in our study was 5 days (IQR 3–7 days), which may cause the attention to whether the patient has TILI in the early stage of TGC treatment. The liver function of patients with liver diseases had been impaired before receiving TGC, which were more likely to be further affected by potential hepatotoxic drugs, and in our study, preexisting liver disease increased the incidence of TILI by 3.2-fold. A pharmacokinetic study [24] demonstrated that variances were observed in Cmax (55% higher), CL (55% lower) and AUC (105% higher) in the patients with Child-Pugh C or severely decompensated cirrhosis. Therefore, it is recommended that patients with liver function impairment should be prescribed with a loading dose of 100 mg followed by a reduced maintenance dose of 25 mg 12-hourly.
In the subgroup analysis of patients with liver disease, patients with TILI had longer ICU stay. Long-stay in ICU may aggravate the liver injury. Patients with TILI had a higher percentage of lung complicated with abdominal infections, which may be related to the higher dose of TGC.
Our study showed that patients with APACHE II score > 15,TILI and TGC maintenance dose 150 mg/day had higher risk of 28-day mortality. The development of TILI further aggravated the severity of the disease in ICU patients. This result suggests that the development of TILI should be reduced to avoid a worse outcome for ICU patients. The TGC maintenance dose ≥ 200 mg/day may increase the risk of TILI, but the effect on 28-day mortality was not statistically significant, this finding needs further study and discussion. Interesting, we also found TGC treatment duration ≥ 7 days was associated with a reduction in 28-day mortality. This was possibly because adequate treatment can help to treat infections and thus reduce the risk of death.
The present study had several limitations. Firstly, our study lacked PK evaluation. Nonetheless, according to the linear PK characteristics of TGC metabolism in the liver, the exposure will increase with the increase of dose, which is more likely to cause liver injury. Secondly, this study did not exclude patients with liver injury before TGC treatment. But the liver function ahead of TGC therapy was taken as baseline to evaluate the changes between before and after medication to exclude other factors as much as possible. In addition, we also performed a stratified analysis of patients with and without liver disease. Thirdly, this study was a retrospective study, the frequency and timing of liver examinations during TGC treatment were not constant due to retrospective design. Thus, prospective, randomized controlled studies are needed to further confirm the conclusions.