Discussion
DITP diagnosis is complicated owing to multiple causes of acquired thrombocytopenia in ICU patients. Among thrombocytopenic patients who present with severe infection, DITP is a very important distinction to make. Firstly, we roughly excluded primary blood diseases and autoimmune diseases through medical history and laboratory examination results. Then we focused on the infection and medication factors. As a priority factor suspected to thrombocytopenia, the severe infection had been alleviating (Table 1) while PLT decreased. Thus, vancomycin caught our attention. It offered further clinical evidence as thrombocytopenia appeared when vancomycin was re-administered, and disappeared as vancomycin discontinued.
Thrombocytopenia was occurred in 7.1 % patients treated with vancomycin18, however, it could be easily overlooked as its description and incidence have not been labelled. There were some case reports of vancomycin-induced thrombocytopenia before. Back in 1985, Walker presented a 48-year-old female patient with secondary peritonitis after peritoneal dialysis. During the treatment period, vancomycin (500mg, i.v.gtt.) was given first, followed by intraperitoneal injection of 120 mg/day, and the patient showed a significant thrombocytopenia 6 days later.21 Howard reported in 1997 that a 58-year-old male patient admitted to hospital due to osteomyelitis caused by MRSA, who manifested thrombocytopenia after vancomycin administration.22 In 2013, a systematic review including pediatric patients (<18 years old)23 found that 32 substances had potential pathogenic effects in thrombocytopenia, including vancomycin (in 3 of 21 cases). In 2017, 30 case reports with 30 patients were included in a scoping review reported vancomycin-induced thrombocytopenia.24 It can occur on the 1st to 10th days after drug exposure. The time window for the platelet count dropping to the lowest point ranged from 4 hours to 10 days after drug exposure. The median platelet count was often dropping < 20 × 109/L25,26, with those <2 × 106/L to 1 × 108/L occurred bleeding.27 The degree and occurring time (the time for the platelet count to drop to 7 × 109/L was 2 days of vancomycin use) of thrombocytopenia in our case was basically consistent with the previous reports.
Significantly, it shows that vancomycin causing thrombocytopenia by accelerating platelet destruction via drug-dependent platelet reactive antibodies. Compared with linezolid, another agent commonly used for MRSA, vancomycin-induced immunemediated thrombocytopenia occurs more rapidly, while linezolid-induced mye-losuppression-mediated thrombocytopenia occurs for a longer time (usually 2 weeks).28 Previous experience showed that the incidence of thrombocytopenia caused by linezolid was high. Notably, based on our systematic review, there was no significant difference in leading thrombocytopenia (defined as platelets<150 x 109/L) between vancomycin and linezolid (RR=1.01, 95%CI=0.64-1.58).
Usually, DITP was alleviative after vancomycin withdrawal, with the severe one required platelet infusion.24 There is no evidence for the efficacy of immunosuppression in treating DITP. In previous reports, vancomycin was discontinued in 29 of 30 patients, and platelet counts in 17 patients recovered within 5-6 days.25, 26 In this process, patients’ infection control needs to be taken into account when discontinuing vancomycin. In our case, luckily, the patient was definitely infected with Escherichia Coli, and we decisively discontinued vancomycin after thrombocytopenia was observed again. As thrombocytopenia leads to an increased risk of bleeding, the transfusion of platelets was a common management though it did not always result in expected increases in platelet counts of affected patients (67% transfusion resistant of 20 patients).24 In this case, platelet infusion was given when platelet count was at 8 × 109/L and 14 × 109/L, and platelet 2U infusion was given at 3 × 109/L.
There are some limitations in this study. Firstly, only postoperative drainage fluid was collected for culture specimens after starting antimicrobial therapy, but no blood samples obtained before at emergency. We should continue to improve the implementation of sepsis bundle, emphasizing the importance of retaining blood samples to monitor and analyze the bacteria before the empirical antibiotic treatment.27 Secondly, we couldn’t confirm whether the cause was DITP or not, as our hospital doesn’t perform tests of drug-dependent platelet reactive antibodies.29 In order to better identify DITP and improve patient prognosis, it is suggested to improve the accessibility of platelet reactive antibody technology in the future. Thirdly, only three major databases were searched for systematic reviewed, therefore some literatures might be overlooked.