COVID-19 and haematological malignancy: navigating a narrow strait

Patients with haematological malignancies face unique infectious risks. Not only do their cancers typically directly affect the immune system, but therapies can cause severe myelosuppression and lymphodepletion, especially in curative settings. Vigilance to avoid life-threatening infection is a part of life for these patients and is crucial in medical decision making. With this context, the COVID-19 pandemic has understandably shaken this community, and more data to guide management are needed.

In The Lancet Haematology, Francesco Passamonti and colleagues report the results of a multicentre, retrospective study aimed at investigating factors associated with mortality in an Italian cohort of 536 patients with haematological malignancies and laboratory-confirmed, symptomatic COVID-19.¹ They found that mortality in this cohort was meaningfully higher when compared with a cohort of patients with haematological malignancies but not COVID-19 (standardised mortality ratio 41·3, 95% CI 38·1–44·9) and with the general Italian population with COVID-19 (2·04, 1·77–2·34).¹ They used multivariable Cox regression to identify factors independently associated with increased mortality, including older age (hazard ratio 1·03, 95% CI 1·01–1·05), progressive disease (2·10, 1·41–3·12), and several specific cancer diagnoses (hazard ratios ranging from 1·30 to 3·49, using).¹ To our knowledge, this is the largest published cohort study dedicated to the outcomes of patients with haematological malignancies and COVID-19, and informs clinical practice.

The finding that patients with haematological malignancies are at increased risk of mortality due to COVID-19 corroborates other studies.2, 3, 4, 5 The magnitude of the risk has implications for medical decision making. Although appropriate therapy should not be withheld, patients and their physicians can take precautions to reduce risks of COVID-19, such as choosing oral over intravenous regimens where there is equipoise, using growth factor support more judiciously, or reducing surveillance laboratory and radiographical evaluations when possible.6, 7

In addition to the high baseline risk posed by COVID-19 to patients with haematological malignancies, the infectious complications associated with many cancer therapies loom large. Passamonti and colleagues' finding that recency of therapy had no association with mortality¹ provides reassurance of the general safety of cancer treatment in this era. Although this is consistent with studies of patients with cancer in general, including our analyses of the COVID-19 and Cancer Consoritum cohort,⁸ the specific finding that this holds for patients with haematological malignancies is novel and is an important contribution to the literature. It is important to note that this does not guarantee the safety of every specific treatment in every clinical scenario. Receipt of multiple distinct lines of cytotoxic therapy has a known association with increased risk of life-threatening infections other than COVID-19, and this might also hold with COVID-19.⁹ The risk–benefit ratio of later-line therapies with questionable benefit, particularly in light of the finding that patients with progressive disease have higher rates of morbid COVID-19, might therefore not be favourable when studied individually. Widely used non-cytotoxic therapies could pose occult risks—eg, anti-CD38 monoclonal antibodies, which can have deleterious effects on natural killer cell populations.¹⁰ Whether it is safe to deploy such agents during the pandemic remains unclear. Investigations of the detailed associations between specific therapies and clinical scenarios with COVID-19 outcomes should be a priority of future work.

Although informative, Passamonti and colleagues' findings must be interpreted cautiously. The precise estimate of mortality reported is probably higher than that of the global population of patients with haematological malignancy and COVID-19. The composition of this cohort, 84% of whom were inpatients, suggests bias in enrolment favouring patients with severe disease; the relatively low rate of intensive care unit admission (18% of patients) might reflect rationing of health-care resources away from the patients in the cohort (and was well documented in northern Italy during the enrolment period); and the high mortality reported in patients with mild disease (48 [18%] of 268 patients) is inconsistent with previous studies. The degree to which mortality is overestimated is likely to be non-random, which could create apparent differences in mortality between groups that might influence the modelling results. The model reported does not adjust for several known risk factors for COVID-19 mortality, such as smoking and functional status; future studies should account for these where possible. The short median follow-up interval of 20 days highlights that the associations identified are with early mortality and might not reflect an entire COVID-19 course; although it remains too early in the pandemic to collect mature long-term outcome data, this should be recognised when applying these data to patient care.

In conclusion, Passamonti and colleagues have advanced our understanding of the unique risks the COVID-19 pandemic poses to patients with haematological malignancies. Although it is appropriate to fear COVID-19, as many health-care systems return to normalcy, deferring treatment is not the optimal response. Patients and their physicians should be mindful of this when deciding on how best to manage living through the COVID-19 pandemic with haematological malignancies.

© 2020 Fanatic Studio/Gary Waters/Science Photo Library