Study population
This retrospective cohort study included kidney transplant recipients with COVID-19 from seven centers in Turkey. Initially, four hundred and eighty-eight kidney transplant recipients diagnosed with COVID-19 between 1 March 2020 to 28 February 2021 were enrolled. The exclusion criteria included patients who were lost to follow-up after diagnosis (n = 7) or still under treatment due to COVID-19 (n = 6). Four hundred and seventy-five participants were followed up for at least 28 days or until death, as shown in Fig. 1. Figure 2 shows the distribution of the patients according to their transplant centers. The Medical Ethics Committee of the Istanbul Faculty of Medicine approved this study.
Data collection
Demographic data (age, sex, comorbidities, etiology of primary kidney disease, duration of post-transplant follow-up, type of donor, induction therapy, maintenance immunosuppression, and medication), clinical characteristics (duration of hospitalization, presenting symptoms, contact history, symptoms, examination findings, and laboratory results), and medication during COVID-19 (antiviral, antibiotic, cytokine‐targeted, anticoagulation treatment, and oxygen therapy) were extracted from electronic medical records. Data from out and hospitalized patients were collected. Patients admitted to the hospital excluded those requiring intensive care unit admission.
Patient management
The patients were diagnosed with COVID-19 based on clinical and laboratory findings. Positive results on real-time polymerase chain reaction (RT-PCR) assay of nasopharyngeal swab specimens or compatible findings on a computed chest tomography scan were used to confirm the diagnosis. In the first three months of the pandemic, patients with a high clinical suspicion of disease were diagnosed by computed tomography of the chest due to low RT-PCR sensitivity [9].
All patients underwent blood testing at admission. Laboratory investigations included a complete blood count and serum C-reactive protein, interleukin-6, markers of myocardial damage (creatine kinase, troponin I, lactate dehydrogenase), tests of secondary hemostasis profile (prothrombin time, activated partial thromboplastin time), serum biochemical tests (including renal and liver function, and electrolytes), procalcitonin, fibrinogen, d-dimer. Laboratory tests were repeated daily for inpatients and weekly for outpatients for two weeks. The estimated glomerular filtration rate (eGFR) was calculated by the modification of diet in the renal disease study equation (MDRD).
Patients were divided into four groups according to the severity of the disease per our national guideline [10]. Asymptomatic patients and patients with mild disease (oxygen saturation above 93%, no lung involvement on chest computed tomography) were followed up in the outpatient clinic weekly. Patients with moderate disease (oxygen saturation above 90%, respiratory rate under 30 breaths/min) were hospitalized in the first wave of the pandemic and followed up in the outpatient clinic in the later period because of hospital overcrowding.
Also, indications for hospitalization were severe disease (oxygen saturation under 90%, respiratory rate above 30 breaths/min), cytokine storm (persistent fever, high or increasing CRP, ferritin, and D-dimer, abnormalities in liver function tests, hypofibrinogenemia with cytopenia in the forms of lymphopenia and thrombocytopenia). Criteria for admission to the intensive care unit (ICU) were the low partial pressure of arterial oxygen, the inspiratory oxygen fraction (PaO2 / FiO2) ratio less than 300, oxygen saturation less than 90%, and PaO2 below 70 mm Hg despite 5 L/minute oxygen therapy, and persistent hypotension (systolic blood pressure < 90 mm Hg or mean arterial pressure < 65 mm Hg).
The cytokine storm was diagnosed according to our national guideline. In this way, cytokine‐targeted therapy indications were standardized for all centers, and problems in drug supply were avoided [10]. Acute kidney injury was classified according to the Kidney Disease Improving Global Outcomes (KDIGO) guideline [11].
Treatment regimens
A protocol was used to manage immunosuppression, antiviral, and cytokine‐targeted therapy, shown in Fig. 3. If the patient is asymptomatic, the immunosuppressive regimen does not change. Antimetabolites (mycophenolate derivatives and azathioprine) were discontinued in symptomatic cases. Trough levels were adjusted as 4–6 ng/dL for tacrolimus, 25–75 ng/dL for cyclosporine, and 3–7 ng/dL for everolimus for a patient with stable clinic course and were stopped in hypoxemic patients. The calcineurin inhibitor and mammalian target of rapamycin inhibitor levels were also monitored twice a week in hospitalized patients, once a week in outpatients.
The patients were treated with hydroxychloroquine (400 mg twice a day for the first, and then 200 mg twice a day for four days via oral administration) and azithromycin (500 mg once a day for the first, and then 250 mg once a day for four days via oral administration) in the first wave of the pandemic. Then, favipiravir (1600 mg twice a day for the first, and then 600 mg twice a day for four days via oral administration) was used during the second wave of the pandemic.
Tocilizumab (400–800 mg once a day for two days via intravenous infusion) or anakinra (100–600 mg once a day for seven to fourteen days or until hospital discharge via subcutaneous injection) was used for cytokine storm treatment. Convalescent plasma and intravenous immunoglobulin (IVIG) were used in the absence of anakinra or tocilizumab. Antibiotic therapy was administered based on the infection specialist’s decision in the presence of confirmed or suspected invasive bacterial infection. The patients were monitored for adverse drug reactions during the hospital stay. Also, QT interval measurements were obtained in all patients by the 12-lead electrocardiography.
Supportive care
Prophylactic-dose low-molecular-weight heparin was used for inpatients unless there were contraindications. Patients with deep vein thrombosis or pulmonary embolism were treated with therapeutic-dose anticoagulation. Doses were adjusted according to the bleeding risk of the patient.
Oxygen treatment was provided to patients whose oxygen saturation was below 92% via a nasal cannula and with a non-rebreather mask if the former was insufficient. If respiratory failure persisted despite these treatments, invasive mechanical ventilation would be used after non-invasive ventilation. All hypoxemic patients were treated with steroids (40 mg methylprednisolone once a day via intravenous infusion in the first wave of the pandemic and 6 mg dexamethasone once a day intravenously after the RECOVERY trial) [12].
Follow-up protocol after discharge
Hospitalized patients were discharged after resolving hypoxia and cytokine storm. Complete blood count and serum C-reactive protein, serum biochemical tests (including renal and liver function, and electrolytes) were monitored weekly after discharge. The calcineurin inhibitor was added to the patients in good condition during the first-week clinical visit. Antimetabolites were restarted at week two in patients with a stable clinical course, and doses were gradually increased at subsequent clinical visits. Trough levels were set at 5–10 ng/dL for tacrolimus and 50–150 ng/dL for cyclosporine in discharged patients.
Outcomes
The primary endpoint was all-cause mortality. The secondary endpoints were acute kidney injury, cytokine storm, and acute respiratory distress syndrome. All endpoints were classified according to the first and second waves. In addition, longer-term outcomes such as mortality, need for dialysis, and allograft function of the patients who recovered and were discharged was analyzed.
Statistical analysis
Patients were categorized as survivors and non-survivors. All parameters such as patient demographic, clinical presentations, laboratory values at admission, treatment regimens, and outcomes were classified according to the categorization. Categorical variables were summarized with numbers and percentages. Quantitative variables were summarized with means and standard deviations or medians and interquartile ranges where appropriate. While chi-square and Fisher's exact test were performed for qualitative variables, the Mann–Whitney U test was used for quantitative variables with the nonparametric distribution.
Univariate, multivariate logistic regression models were used to determine patient characteristics and COVID-19 complications with death. The variables associated with mortality with a p-value less than 0.05 were entered into this model. We tried a multivariate risk model using only six vital predictions from univariate models due to the limited sample size. Although acute respiratory distress syndrome and immunosuppression withdrawal were strong predictors of death, they were excluded due to collinearity with cytokine storm. Hazard ratios and the corresponding 95% confidence intervals refer to the increase per unit in a continuous variable. A p-value of less than 0.05 is considered significant.