In this retrospective cohort study, we found potential associations between cefoperazone/sulbactam treatment and higher risks of PT prolongations, aPTT prolongations, coagulation disorders, and bleeding. However, no significant associations were observed with higher risks of decreased PLT and TT prolongations. Furthermore, the results of logistic regression analysis, performed before and after PSM in the two groups, consistently showed that vitamin K1 supplementation reduces the risks of PT prolongation and coagulation dysfunction induced by cefoperazone/sulbactam. Additionally, we documented a total of eight bleeding events, with seven occurring in the non-vitamin K1 group. These bleeding events were observed in patients with coagulation disorders, reinforcing the notion that coagulation disorders may contribute to major bleeding [6].
A retrospective study by Wu et al. also reported that cefoperazone/sulbactam causes significant prolongation of PT and aPTT but not TT in patients with long-term aspirin use. These observations can be attributed to the cefoperazone/sulbactam interference with vitamin K metabolism and synthesis, leading to the disruption in the synthesis of various vitamin K-dependent coagulation factors, while having no impact on fibrinogen synthesis and secretion. However, unlike our study, they also found that patients experienced a significant decrease in platelet count following cefoperazone/sulbactam treatment, which might be related to aspirin-induced platelet apoptosis [14]. Furthermore, the concurrent use of cefoperazone/sulbactam and aspirin might enhance the therapeutic effect of aspirin by regulating aspirin's pharmacokinetic [17].
Cefoperazone can induce hypoprothrombinemia, but the mechanism is not fully understood. Some experts proposed that the main cause is associated with cefoperazone’s N-methyltetrazolium sulfide (NMTT) side chain interfering with vitamin K1 metabolism. Two potential mechanisms have been suggested:1) it can inhibit vitamin K1-dependent gamma-carboxylation of glutamic acid, thereby inhibiting K1 epoxide reductase; and 2) because the drug is mainly excreted through the biliary tract, and therefore, can inhibit vitamin K1-producing intestinal microflora [8].However, when comparing with other cephalosporins with NMTT or even cefoperazone-tazobactam, cefoperazone/sulbactam was found to have a higher risk of coagulopathy[7–8]. This suggests that sulbactam may intensify cefoperazone’s effect on coagulation function [7]. A study on the PK/PD of cefoperazone/sulbactam in cirrhotic patients showed that the total clearance of cefoperazone was often decreased in cirrhotic patients [10], which would theoretically exacerbate vitamin K1 deficiency. Based on this pathogenic mechanism, vitamin K1 supplementation has been suggested as a resolution [18]. In the present study, the SHAP method revealed 17 predictors of coagulation dysfunction, with “drugs” identified as the most important predictor variables and vitamin K1 supplementation having a negative impact, pushing the prediction towards normal coagulation function, which further verified this hypothesis.
The Child-Pugh score is widely used to assess the severity of cirrhosis in patients [19]. In our study, the SHAP method revealed that Child-Pugh C had a positive impact and pushed the prediction towards abnormal coagulation function, which is consistent with our findings from multivariate analysis within the vitamin K1 group (Child-Pugh C: OR = 17.931, 95% CI: 1.646-342.818, P = 0.03; Supplementary Table S1). This observation aligns with previous studies, such as the one conducted by Penget al., where they observed gradual prolongation of PT and aPTT and an increase in the Child-Pugh score (P < 0.01) [3]. Additionally, Liet et al. also found that hemostatic parameters, including PT, INR, aPTT, and TT, gradually increased with the increase of Child-Pugh score [2].
In our study, the SHAP analysis revealed that the daily dose of cefoperazone/sulbactam had little predictive value for clotting abnormalities, which is consistent with our multivariate analysis results within the vitamin K1 group (T/X2 = 0.309, P = 0.579, Supplementary Table S1). This finding may be explained by the previous study of Dong Y et al., which reported no significant difference in the Cmin content of cefoperazone between the two administration regimens of 3g q8h and 3g q12h in patients with liver cirrhosis. This outcome may be related to the increase of creatinine clearance through high frequency administration [10]. On the other hand, the SHAP analysis revealed the total duration of antibiotic ranked 6th in the list of influencing factors and showed some value in predicting coagulation abnormalities.
In previous studies, various recommendations have been made regarding the dosage of vitamin K1 for prophylaxis. Wu et al. suggested a subcutaneous injection of 10 mg of vitamin K1 before each dose of cefoperazone/sulbactam, while Hunt et al. recommended a weekly dose of 10 mg for critical care patients at risk of bleeding [20]. However, none of these studies have explored the association between the drug dose of vitamin K1 and its impact on coagulation. In our study, the vitamin K1 group received daily subcutaneous injections of vitamin K1 ranging from 10 mg to 30 mg for more than 48 hours. The SHAP analysis indicated that the total duration and daily dose of vitamin K1 ranked 12th and 17th, respectively, in the list of influencing factors, suggesting limited predictive value in assessing coagulation abnormalities. Considering the potential risk of serious adverse reactions associated with higher doses of vitamin K1, such as severe allergic reactions, cardiovascular system damage, and respiratory system damage [21], we recommend a daily prophylactic dose of 10 mg of vitamin K1.
In this novel retrospective analysis, we investigated the impact of cefoperazone/sulbactam on coagulation in patients with liver cirrhosis, while also exploring the potential improvement in coagulation function through vitamin K1 prophylaxis. To mitigate the influence of confounding factors, we utilized PSM to balance baseline characteristics, resulting in 60 patients in each group. Additionally, we established an association between the daily dose and course of cefoperazone/sulbactam, vitamin K1 supplementation, and coagulation, which compensated the shortcomings of the previous study [14]. We also employed the SHAP model to ascertain the importance of each predictor variable in coagulation disorders. Despite these valuable findings, several limitations must be acknowledged. Firstly, this study was conducted at a single center, potentially limiting the generalizability of our findings. Secondly, the sample size was modest. These limitations highlight the need for future studies with larger sample sizes and a multi-center approach to bolster the robustness of our conclusions.