Therapeutic plasma exchange in patients with COVID-19 pneumonia in intensive care unit: a retrospective study

In patients with COVID-19 pneumonia, high risk of thrombosis became a current issue, and d-dimer levels indicating fibrin degradation products (FDPs) in the plasma were found as a predictor for mortality [1, 2]. Although unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH) decrease the production of FDPs by inhibiting factors Xa and II, they cannot contribute metabolization of existing FDPs. Furthermore, FDPs cannot be filtered by known cytokine filters because of their molecular weight (minimum 240 kDa) [3, 4]. Yet, FDPs can be removed by therapeutic plasma exchange (TPE) [5]. Therefore, recently, three consecutive TPE sessions were performed in selected patients with COVID-19 pneumonia in intensive care units (ICUs) after the assessment of their clinical and coagulation status. In the study, the effect of TPE on outcomes was retrospectively investigated in patients with COVID-19 pneumonia.

All COVID-19 patients admitted to 5 different tertiary ICUs between 10 March and 10 May 2020 were evaluated, and 73 of 91 patients were included in the study. The patients who died within the first 4 days and who were still in the ICUs on 10 May were excluded. According to the Turkish Health Minister Algorithm for COVID-19, all included patients received the same antiviral (favipiravir, hydroxychloroquine, azithromycin) therapy and anticoagulant prophylaxis (UFH infusion 100 mcg/kg or LMWH 0.01 mL/kg). Since two different protocols were used in 5 ICUs, patients with d-dimer ≥ 2 in 3 ICUs had only received therapeutic anticoagulation whereas patients with d-dimer ≥ 2 in the other 2 ICUs had received TPE plus therapeutic anticoagulation. In all ICUs, for all patients in GII, echocardiography, lower extremity venous Doppler, and, if pulmonary thrombosis suspected, thorax computerized tomography angiography were performed. After collecting data, 73 patients were divided into 2 groups as group I (GI) (d-dimer < 2 mg/L) and group II (GII) (d-dimer ≥ 2 mg/L), and then GII was also divided into 2 groups as GIIa (TPE+) and GIIb (TPE−). Patients’ characteristics, respiratory and laboratory parameters, and outcomes were recorded. Propensity score matching (PSM) analysis was conducted on R v4.0.1 (0.2 caliper without replacement and nearest neighbor model, 1:1 ratio) by using 14 covariates (age, gender, CCI, APACHE II, SOFA score, lactate, leucocyte, lymphocyte, d-dimer and creatinine at the ICU admission, maximum respiratory support, the usage of steroid, IL-6 blocker, and cytokine filter).

The total mortality rate was 27.4%. Mortality rates of GI and GII were 5% and 35.9%, respectively. In GII, major thromboembolic events were not detected in any patients. The median (min-max) day for the starting TPE was 3 (2–4). In GIIa, APACHE II, SOFA scores, d-dimer and interleukin-6 (IL-6) levels at the ICU admission, and length of ICU stay were significantly higher than those of GI whereas mortality rates were similar in those groups (Table 1). The median values of the LOS-ICU in survivors and non-survivors in GII were 14 (6.5–21.5) and 15.5 (8–23), respectively (p = 0.630). In GIIa, lactate dehydrogenase (LDH), d-dimer, ferritin, IL-6, C-reactive protein (CRP), and procalcitonin levels were significantly decreased after three consecutive TPEs (Table 2). Furthermore, although ferritin level at the ICU admission was higher in GIIa, the mortality rate in both before and after PSM was higher in GIIb (45.7% and 58.3%) than in GIIa (16.7% and 8.3%) (p = 0.037, p = 0.009, respectively) (Table 1).

Table 1 Comparisons of patients groups

Table 2 Comparisons of laboratory parameters in pre and post-TPE procedure

Some patients with COVID-19 pneumonia have a high risk of thrombosis leading to worse outcomes. Therefore, monitoring d-dimer levels is crucial. In these groups of patients, TPE seems to be a treatment which may improve outcomes by effectively removing FDPs and restoring coagulation status. We are aware that TPE may not be routinely required in these patients [6]. However, we think that it should be featured as a part of the treatment especially in COVID-19 pneumonia patients with a high risk of thrombosis.