The disease caused by SARS-CoV-2, named COVID-19, has spread rapidly. Among adult patients, a significant prevalence of thrombotic complications has been described. This risk has been linked to a hypercoagulability state, which has conditioned the publication of guidelines on anticoagulant management in patients with COVID-19.1 2
However, due to the lower experience in COVID-19 disease in children, these guidelines have not included specific recommendations for paediatric patients. Based on adult recommendations, several authors have proposed some indications for anticoagulant prophylaxis that take into account individual risk factors.3 Nevertheless, to date, there is scarce information about thrombotic complications in children with COVID-19. We aimed to describe thrombotic complications among children infected with SARS-CoV-2.
The Epidemiological Study of COVID-19 in children was launched and supported by the Spanish Society of Pediatrics (EPICO-AEP) as a multicentre prospective national cohort study aiming to describe paediatric COVID-19 in Spain. Children younger than 18 years of age infected by SARS-CoV-2 and attended at 49 hospitals are currently being enrolled in this registry. Inclusion criteria include positivity in real-time PCR (RT-PCR), IgM or IgG in lateral-flow rapid test, ELISA or chemiluminescence serology, or disease suggestive of multi-inflammatory syndrome related to SARS-CoV-2 in children (MIS-C). From 1 March to 30 September 2020, children with thrombotic complications associated with SARS-CoV-2 enrolled in EPICO-AEP were included in this study.
Thrombotic complications associated with SARS-CoV-2 infection were defined as any radiologically confirmed thrombosis, arterial or venous, occurring close to the diagnosis of SARS-CoV-2 infection (up to 3 weeks). Microthrombi phenomena related or suspected to be related to SARS-CoV-2 (eg, acral chilblains) were not included. Two-sided 95% Clopper-Pearson exact CIs (95% CI) for binomial proportions were calculated for prevalence. Data were analysed using the Stata V.15 (College Station, Texas, USA).
By 30 September, 537 children were diagnosed with SARS-CoV-2 infection and included in the EPICO-AEP cohort. The median age was 61 months (IQR: 6–140 months), and 293 (54.6%) were boys. Of those, 368 (68.5%) were hospitalised, requiring paediatric intensive care unit admission (PICU), 58 (10.8%) cases. Forty-seven (8.8%) cases were diagnosed with MIS-C.
D-dimer was available from 169 patients, with a median of 1071 µg/L (IQR: 291–2858 µg/L). Anticoagulant drugs (heparin in all cases) were administered to 29 (5.4%) patients: as prophylaxis in 24 (82.8%) cases, and as treatment in 5 (17.2%) cases. Most of the patients on thromboprophylaxis had severe COVID-19 (79.2% admitted to PICU and 56.5% received inotropic support). Four cases, which correspond to 0.7% (95% CI: 0.2% to 1.9%) out of the global cohort and 1.1% (95% CI: 0.3% to 2.8%) out of the hospitalised patients, developed some thrombotic complication. A fifth case receiving anticoagulant treatment was a patient who had a history of venous thrombosis not related to SARS-CoV-2 infection and was not included in this report.
Three out of these four patients were adolescent girls (table 1, cases 1–4). None of the cases was diagnosed with MIS-C. Two patients (patients 1 and 2) had several significant thrombotic risk factors. Patients 3 and 4 did not have any previous personal or family risk factors for thrombosis. However, in patient 3, after thrombosis diagnosis, a thrombophilia work-up was performed, which revealed a positive lupus anticoagulant, remaining positive 5 months later.
The only patient receiving thromboprophylaxis (low-weight heparin) before thrombosis was patient 2. D-dimer levels were severely increased in patients 1 (5953 µg/L) and 4 (35 420 µg/L) but were only slightly increased in the other cases (1194 and 232 µg/L, respectively). All cases except patient 3 had SARS-CoV-2 RT-PCR confirmation coinciding with thrombosis diagnosis. In patient 3, COVID-19 was initially suspected but not microbiologically confirmed until 3 weeks after thrombosis. All cases had thrombosis affecting the limbs, with cerebral venous sinus thrombosis and pulmonary embolism in one of them. All patients were treated with heparin and discharged without sequelae.
To the best of our knowledge, this is the first study that describes thrombotic complications in children with SARS-CoV-2 infection in a large cohort. Only four patients developed a thrombotic complication (1.2% of hospitalised children with non-MIS-C SARS-CoV-2 infection). Although this prevalence is not as high as in hospitalised adults with COVID-19,1 it is slightly higher than that previously described in hospitalised children without COVID-19 (0.13%–0.55%). Among the factors associated with a lower incidence of thromboembolic complications in children with COVID-19, a different balance in coagulation homeostasis, with a higher concentration of antithrombotic serum factors (eg, alpha-2-macroglobulin) in children compared with adults, a protective ‘healthy’ endothelium or fewer comorbidities have been proposed.
Only two of our four cases had significant thrombotic risk factors (central venous catheter, obesity, malignancy or asparaginase treatment), highlighting the difficulty of predicting thrombotic complications in children with COVID-19 to initiate anticoagulant prophylaxis. Thrombosis is very uncommon in patients under 18 years old, and systemic anticoagulant prophylaxis may have adverse events that outweigh its benefit. Of note, one patient was diagnosed with antiphospholipid syndrome, showing lupus anticoagulant persistently elevated. Transient lupus anticoagulant has been described in patients with COVID-19, but its significance is unclear.
We conducted a systematic search in order to review reported paediatric patients diagnosed with SARS-CoV-2 infections with any thrombotic complication (search strategy in online supplemental material). Ten cases were found (full references in online supplemental material); one of them included in our series (case 4). Table 1 (cases 5–13) shows the main characteristics of the reported cases. Additionally, three (3.8%) and four (2.2%) cases diagnosed with MIS-C from two other national cohorts developed a venous thrombosis, but their full clinical information was not available.4
The central nervous system was the most commonly affected site (six cases), followed by the lung (five cases) and lower limbs (three cases). Similar to our data, most of the patients were adolescents. This age distribution is also seen in patients with MIS-C, in which the prevalence of thrombosis increases with the age of the patients: 0 of 66 (0%), 1 of 75 (1.3%) and 3 of 45 (6.7%) in patients aged <5 years, 5–12 years and 13–20 years, respectively.4 This peak of cases with thrombosis affecting adolescents is similar to what is seen in hospitalised children without COVID-19.
Markedly elevated plasma D-dimer levels (eg, ≥5 times the upper limit of normal values) has been proposed as an indication for anticoagulant prophylaxis in children hospitalised with SARS-CoV-2-related illness.3 In our cohort, 68 of 169 (40.2%) had a value >1500 µg/L. However, only 2 of 68 (2.9%) cases developed a thrombotic complication. Therefore, the use of the D-dimer value is not specific enough to make decisions regarding anticoagulant prophylaxis in children. In our opinion, other factors, such as the age (considering adolescents especially vulnerable), coexistence of risk factors (eg, malignancy, indwelling central venous catheter, obesity, immobility, etc),5 or MIS-C diagnosis, should be considered when initiating anticoagulant prophylaxis in children with COVID-19. Of note, a clinical trial is currently evaluating the safety and efficacy of thromboprophylaxis in children hospitalised with signs and/or symptoms of SARS-CoV-2 infection (ClinicalTrials.gov NCT04354155).
Our study has some limitations, such as the variability of tests that were used to diagnose infection, not only the gold standard PCR, which could make false results. Additionally, since ultrasound was not routinely performed in all patients, subclinical thrombosis may not have been diagnosed.
In conclusion, except in the MIS-C, unlike adults, thrombotic complications seem very uncommon in children with SARS-CoV-2. Adolescence and previous thrombotic risk factors may be considered when initiating thromboprophylaxis in children with COVID-19. Further studies are needed to clarify risk factors among children with COVID-19 in order to develop specific recommendations.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplemental information.
This study was approved by the Ethics Committee of Hospital 12 de Octubre (20/101).