key: cord-1056294-mxmm1eyc
authors: Lesosky, Maia; Myer, Landon
title: Modelling the impact of COVID-19 on HIV
date: 2020-09-30
journal: The Lancet HIV
DOI: 10.1016/s2352-3018(20)30228-9
sha: dcaf0a3c0591cb537e7f3e09d8cf9d6de1004457
doc_id: 1056294
cord_uid: mxmm1eyc

nan

the prospective multicentre observational study that is reported here. There are important differences between HIV-infected deceased organ donors in South Africa, where the vast majority are young trauma victims, mostly either not on ART or on only first-line therapy, and those in the USA, where there is a far higher prevalence of drugresistant HIV mutations. 8 Hence, concerns about donorderived HIV superinfection are heightened in this cohort.

In this study, 22 HIV-positive to HIV-positive liver and kidney transplant recipients in the USA were followed for a period of up to 3 years after transplantation, with blood collection for virological analysis at multiple timepoints. Donor blood samples were collected at the time of transplantation. Total HIV proviral DNA and viral RNA were sequenced to look for evidence of donor-derived HIV superinfection in the recipients using phylogenetic analysis and direct sequence comparison. Reassuringly, HIV viral load remained suppressed for all recipients throughout the period of follow-up, except for one patient who became viraemic 3 years after transplantation in the context of ART non-adherence. Importantly, there was no evidence of sustained donor-derived HIV superinfection in any recipient, even in the viraemic patient who had temporarily ceased ART. This finding is particularly reassuring because three of 11 of the deceased donors had one or more major drug-resistant mutations that were discovered after genotyping, although none had multiclass drug resistance.

There remain additional concerns, particularly once transplantation from an HIV-positive donor to an HIVpositive recipient takes place outside the close monitoring of a clinical trial. The South African group reported that among 27 HIV-positive patients who received kidneys from HIV-positive deceased donors, recurrent HIV-associated nephropathy developed in the transplanted kidney in three patients, even in the absence of detectable viraemia. 9 This observation is consistent with a separate report showing that the transplanted kidney is itself a reservoir for HIV. 10 Nevertheless, given the substantial gap between organ availability and demand, appropriate use of organs from HIV-positive donors can expand the donor pool and is a viable option for HIV-positive candidates for transplantation. This study will reassure transplantation candidates and clinicians, encourage wider uptake of this practice, and also, given that organ donation saves lives, help to challenge the HIV-related stigma that persists in both the transplantation and wider community. 

The global spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and the response to the pandemic by national authorities, has been unprece dented in its rapidity and scope. The virus and our responses to the epidemic have driven profound shifts in social interactions, national economies, and health services across the world. Understanding the implications of these shifts, particularly over the long term, is a massive challenge facing policy makers. In this context, mathematical models or simulations to predict epidemics and the effects of efforts to Comment www.thelancet.com/hiv Vol 7 September 2020 e597 mitigate them are well established tools for planning public health programmes. These models provide broad insights into what could happen in the future under different scenarios and assumptions. In emergent epidemics, these models are especially valu able to generate forecasts relatively rapidly, and, indeed, mathematical models of SARS-CoV-2 transmission and its effects on population health have thrived in the past few months. 1 However, the modelling of epidemics and their impacts is a highly uncertain business. Developing, parameterising, calibrating, and validating a model involves oversimplifications of a poorly understood reality, adaptations of imperfect empirical data, and gross assumptions to explain away the huge unknowns. 2,3 Like all scientists, mathematical modellers often have a vested interest in their own tools and results, and understandings of uncertainty in the modelling process are often underdeveloped. 4 As a result, no single model ever provides an absolute answer; arguably, the findings of any mathematical model should be read with a healthy accompaniment of scepticism.

In The Lancet HIV, the Article by Britta Jewell and colleagues 5 provides a refreshing approach to understanding how responses to SARS-CoV-2 might influence the course of the HIV epidemic. The analysis brings together the results of five different mathematical models, each asking the same question. The results are largely consistent across the models and arrive at a sobering conclusion: relatively short interruptions in HIV prevention and treatment services, and particularly antiretroviral therapy programmes, might lead to substantial increases in mortality. The magnitude of these increases could parallel deaths due to SARS-CoV-2 infection in countries where HIV prevalence is high.

The joint presentation of multiple, independent mathematical models investigating a single question provides a strong basis for understanding the strengths and limitations of each model, and for inference towards public health policy making. This method should be viewed as a gold standard approach to presenting findings as we move away from attempting to make inferences from any single mathematical model in isolation. As this new gold standard emerges, improved standardisation and transparency in the presentation of model structure, assumptions, and results will be required. Furthermore, new methodological questions are likely to emerge, including around whether and how to combine model results using meta-analytical approaches. Jewell and colleagues begin to explore useful directions in their study, but much more work is needed.

The convergence of findings across these five models provides robust insights into a concern held by many individuals at risk of or living with HIV, and clinicians and policy makers involved in HIV services. 6 With the work of Jewell and colleagues, these concerns can shift from an anecdote that provokes anxiety to a more tangible phenomenon that policy makers can work to prevent. However, notably, this work addresses only the effect of the SARS-CoV-2 response on HIV services, and does not go into the potential effects on other health conditions and services, including other infectious diseases, 7 maternal and child health, 8 and chronic disease care. 9 Considering these issues, we must recognise that not all health sequelae are negative and some notable health benefits might result from the response to SARS-CoV-2, 10 linking back to the importance of mathematical modelling in weighing risks and benefits across a population.

In its broadest interpretation, the study by Jewell and colleagues asks the uncomfortable question of whether the response to SARS-CoV-2 might lead to greater long-term morbidity and mortality than the virus itself. Such a possibility could have profound implications for public health policies and programmes globally. However, answering this uncomfortable ques tion will require much more than mathematical modelling and will no doubt drive empirical investigations for years to come.

Cancer and HIV have been inextricably linked since the first report in 1982 of biopsy-confirmed Kaposi's sarcoma among previously healthy homosexual men living in Los Angeles and Orange counties, California, USA. 1 The cases established concerns for what later would be defined as AIDS. The report speculated: "One hypothesis consistent with the observations reported here is that infectious agents are being sexually transmitted among homosexually active males. Infectious agents not yet identified may cause the acquired cellular immunodeficiency that appears to underlie KS [Kaposi's sarcoma] and/or PCP [Pneumocystis carcinii pneumonia] among homosexual males." The discovery of HIV proved it to be the aetiological agent. Given the known association between immunodeficiency and cancer and oncogenic virus and cancer, the discovery of human herpesvirus 8 (also known as Kaposi's sarcomaassociated herpesvirus) completed the puzzle on the mysterious outbreak of Kaposi's sarcoma. 2 Non-Hodgkin lymphoma was the second cancer linked to HIV and soon it was included, along with cervical cancer, in the Center for Disease Control definition of AIDS. Later, various non-AIDS-related cancers would be recognised as having an increased incidence in those living with HIV, including Hodgkin lymphoma. The cause of non-Hodgkin lymphoma is diverse and includes B-cell dysregulation, suppressed immune surveillance, oncogenic viruses, and possibly direct effects of HIV itself. Thankfully, the incidence of AIDS-related cancer is markedly diminished, though incompletely eradicated, with the use of multiagent antiretroviral therapy (ART). Those living with HIV are still at increased risk, particularly as life expectancy is now that of the general population. One approximation is a 4% lifetime risk of non-Hodgkin lymphoma. 3 In the early days of HIV infections, those diagnosed with HIV invariably developed AIDS at a median of 8 years after infection. The inexorable destruction of the immune system led to a multiplicity of opportunistic infections and the development of cancers. Patients with AIDS-related cancers fared poorly for several reasons. The body was already weak, which made chemotherapy difficult to withstand. The ongoing risk of opportunistic infections during chemotherapy was daunting. The absence of robust immunosurveillance increased the risk of relapse of even curable cancers like lymphoma. Finally, a pronounced shorted life expectancy reduced the benefit of aggressively treating cancers. For lymphoma, a randomised trial established reduced-dose chemotherapy as standard of care. 4 This situation changed with the advent of ART. Most notably, the normal life expectancy of individuals with HIV and state-of-the-art care makes addressing cancer a compelling public health concern. Many patients presenting with lymphoma are already on ART and with distant or no history of opportunistic infections. Multiple trials established full-dose chemotherapy that could be safely given to patients with curative intent and with robust outcomes. Regardless, the treatment paradigms remained suspect for decades, as the bias against those with HIV was so ingrained in the general and medical psyche. Research has established full-dose treatment as standard of care, except in the instance where patients present with lymphoma and neglected and typically undiagnosed HIV present with non-Hodgkin lymphoma. However, even in this setting, ART institution and reduced-dose chemotherapy can lead to rapid partial immune reconstitution and the

Potential effects of disruption to HIV programmes in sub-Saharan Africa caused by COVID-19: results from multiple mathematical models

Contracting HIV or contracting SAR-CoV-2 (COVID-19) in pregnancy? Balancing the risks and benefits

Potential impact of the COVID-19 pandemic on HIV, tuberculosis, and malaria in low-income and middle-income countries: a modelling study

Early estimates of the indirect effects of the COVID-19 pandemic on maternal and child mortality in low-income and middle-income countries: a modelling study

The public health response to COVID-19: balancing precaution and unintended consequences

South Africa's COVID-19 alcohol sales ban: the potential for better policy-making