Immunoglobin G/total antibody testing for SARS-CoV-2: A prospective cohort study of ambulatory patients and health care workers in two Belgian oncology units comparing three commercial tests

BACKGROUND: Coronavirus disease (COVID-19) is interfering heavily with the screening, diagnosis and treatment of cancer patients. Better knowledge of the seroprevalence and immune response after Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection in this population is important to manage them safely during the pandemic. METHODS: 922 cancer patients, 100 non-cancer patients and 94 health care workers (HCW) attending the Multidisciplinary Oncology Unit of Antwerp University Hospital from 24th of March 2020 till 31st of May 2020, and the Oncology Unit of AZ Maria Middelares Hospital, Ghent, from 13th of April 2020 till 31st of May 2020 participated in the study. The Alinity® (A; Abbott) and Liaison® (D; DiaSorin) commercially available assays were used to measure SARS-CoV-2 IgG, while total SARS-CoV-2 Ig was measured by Elecsys® (R; Roche). RESULTS: In the overall study population IgG/total SARS-CoV-2 antibodies were found in respectively 32/998 (3.2%), 68/1020 (6.7%), 37/1010 (3.7%) and of individuals using the A, D or R test. Forty-six out of 618 (7.4%) persons had a positive SARS-CoV-2 polymerase chain reaction (RT-PCR) test. Seroprevalence in cancer patients (A:2.2%, D:6.2%, R:3.0%), did not significantly differ from that in non-cancer patients (A:1.1%, D:5.6%, R:0.0%), but was lower than the HCW (A:13%, D:12%, R:12%; respectively Fisher’s exact test p = 0.00001, p = 0.046, p = 0.0004). A positive SARS-CoV-2 RT-PCR was found in 6.8% of the cancer patients, 2.3% of the non-cancer patients and 28.1% of the HCW (Fisher’s exact test p = 0.0004). Correlation between absolute values of the different Ig tests was poor in the cancer population. Dichotomising a positive versus negative test result, the A and R test correlated well (kappa 0.82 p McNemar test = 0.344), while A and D and R and D did not (respectively kappa 0.49 and 0.57; result significantly different p McNemar test = <0.0001 for both). The rate of seroconversion (>75%) and median absolute antibody levels (A: 7.0 versus 4.7; D 74.0 versus 26.6, R: 16.34 versus 7.32; all >P Mann Whitney U test = 0.28) in cancer patients and HCW with a positive RT-PCR at least 7 days earlier did not show any differences. However, none (N = 0/4) of the patients with hematological tumours had seroconversion and absolute antibody levels remained much lower compared to patients with solid tumours (R: 0.1 versus 37.6, p 0.003; D 4.1 versus 158, p 0.008) or HCW (all p < 0.0001). CONCLUSION: HCW were at high risk of being infected by SARS-CoV-2 during the first wave of the pandemic. Seroprevalence in cancer patients was low in the study period. Although Ig immune response in cancer patients with solid tumours does not differ from healthy volunteers, patients with hematological tumours have a very poor humoral immune response. This has to be taken into account in future vaccination programmes in this population. SARS-CoV-2 antibody tests have divergent results and seem to have little added value in the management of cancer patients.