Baseline characteristics and outcomes of the whole population (n=258) are shown in Table 1. Among the type of ketamine administered, 60.7% of the patients received exclusively K, 26.7% exclusively ESK and 12.6% both K and ESK. Cumulative dose was expressed only for the patients who received exclusively K or ESK, as no exact equivalence could be determined in case of combined therapy.
Baseline values of ALP, GGT, bilirubin, AST and ALT are shown in Table II. At baseline, 40 (15%) patients had abnormal ALP, 160 (61.5%) had abnormal GGT, 10 (3.8%) had abnormal bilirubin, 163 (62.7%) abnormal AST, and 117 (45%) abnormal ALT.
The comparison between K/ESK versus no K/ESK group showed that the peak value for ALP, GGT and bilirubin was significantly higher in the K/ESK group (p value respectively <0.0001, <0.0001, =0.0005). (Table 2, Figure 1) No significant difference existed for peak AST (p=0.069) and ALT (p=0.059) values, despite a trend towards higher value in the K/ESK group. When looking at subgroups, patients with K/ESK and mechanical ventilation had a higher peak ALP value than the patients with IMV but without K/ESK (p=0.018) and the patients who did not require neither mechanical ventilation nor K/ESK (p<0.0001 ). For the other peak values (GGT, bilirubin, AST, ALT), the difference was only significant between the mechanically ventilated patients with K/ESK and the patients without mechanical ventilation and K/ESK. There was no significant difference regarding peak values of ALP, GGT, bilirubin, AST and ALT between patients who received K, ESK or both (Supplemental material). Besides, there was a moderate correlation between the cumulative dose of K/ESK over the ICU stay and the peak value for ALP (Pearson’s correlation coefficient R=0.33, p<0.0001)’s, GGT (R=0.41, p<0.0001) and bilirubin (R=0.37, p<0.0001) (Figure 2). The same observation applied to the correlation between the duration of K/ESK therapy and peak values for ALP (R=0.3, p=0.0004) and GGT (R=0.36, p<0.0001) (Figure 2). The median “time to peak value” of ALP (for the values at least two times higher than ULN, n=85) was 14 days (IQR 7-23).
Among the patients who developed abnormal cholestatic biomarkers during ICU stay, 38 were investigated by abdomen ultrasonography. Examination was normal in 16 patients, abnormal in 22 patients, predominantly in the K/ESK group. The most frequent anomaly was gallbladder sludge (n=14), followed by liver steatosis (n=5), gallbladder hydrops (n=5), and acute cholecystitis (n=2). In the K/ESK group, two patients needed percutaneous drainage of the gallbladder and one patient required additional surgical drainage for biliary peritonitis.
Among the patients who were excluded from the analysis for pre-existing liver disease, a 30-year-old woman had a Child-Pugh B liver cirrhosis as an unusual complication of Turner syndrome (non alcoholic steatohepatitis). When mechanically ventilated for COVID-19 related ARDS, she received a cumulative dose of 2040 mg of K and 2370 mg of ESK over a period of 28 days. We noted a marked increase in plasma ALP levels (Figure 3, left). Due to the progression of liver failure, a liver transplantation was required 28 days after the cessation of ESK. Ultrastructural examination of the explanted liver was consistent with a cholestatic injury (Figure 3, right). The responsibility of K/ESK remained speculative as the patient also experienced infectious complications, but without shock.
Otherwise, in this case series of COVID-19 patients with K/ESK, no patient required a liver transplantation and a progressive decrease of ALP levels was noted in most of the patients who survived.