Diabetes, ethnic minority groups and COVID‐19: an inevitable storm

Abstract

The risk of type 2 diabetes (T2DM) is two‐ to four‐fold higher in ethnic minority populations compared to White populations in the UK and is also associated with an increased risk of certain macrovascular and microvascular complications. Additionally, T2DM has an earlier onset in ethnic minority groups of around 10–12 years than in White populations. The exact reasons for the higher prevalence are unclear but include the complex interplay of biological, lifestyle, environmental and socioeconomic factors. This is further compounded by disparities in care received by ethnic minority populations. The UK was the first country to report on the disproportionate impact of COVID‐19 on ethnic minority groups. Diabetes is also a major risk factor for severe COVID‐19 and, combined with pre‐existing ethnic disparities in diabetes care, has been a significant contributor to inequalities in COVID‐19 outcomes for ethnic minority populations with diabetes including disproportionate hospitalisation and mortality. Major ethnic disparities in diabetes care in the US and UK, especially intermediate outcomes and diabetes complications, were evident prior to the COVID‐19 pandemic. However, the COVID‐19 pandemic has exposed these pre‐pandemic health disparities for ethnic minority populations with diabetes. Similar to the higher risk of T2DM in ethnic minority populations, the exact reasons for higher risk of COVID‐19 in minority ethnic groups are complex and include comorbidities, risk factor control, deprivation and access to care including wider structural issues. As we now plan for recovery, it is imperative that those delivering diabetes care urgently address the disproportionate impact the pandemic has had on ethnic minority populations. Reducing these inequalities will require a greater understanding of the causes. Copyright © 2022 John Wiley & Sons.

Background

The risk of type 2 diabetes (T2DM) is two‐ to four‐fold higher in ethnic minority populations compared to White populations, and is also associated with an increased risk of certain macrovascular and microvascular complications. Major ethnic disparities in diabetes care were evident prior to the COVID‐19 pandemic. ¹ The UK was the first country to highlight the disproportionate number of cases and adverse outcomes of SARS‐CoV‐2 infection in ethnic minority populations, and as the pandemic progressed, it has become clear that large disparities exist for hospital admissions and mortality in people from ethnic minority populations. ² Diabetes is a major risk factor for severe COVID‐19 and the combination of ethnic disparities in people with diabetes has been a major contributor to the inequity in COVID‐19 outcomes for people with diabetes. ¹

In this 2022 Arnold Bloom lecture, I compared and contrasted some of our research work showing similarities in outcomes in ethnic minority populations, with people with diabetes and COVID‐19, as COVID‐19 appears to be a cardiometabolic condition, rather than our initial understanding that it was primarily a respiratory condition.

The intersection of ethnicity, diabetes and COVID‐19

The NHS Health Check programme was introduced in England in 2009 to improve the primary prevention of coronary heart disease, stroke, diabetes and chronic kidney disease. ³ An evaluation of the programme showed that people of ethnic minority background had a lower body mass index (BMI) and higher HbA_1c when they were screened. ⁴ These data also showed that South Asians had a higher prevalence of diabetes, prediabetes and high cardiovascular risk compared to White Europeans in the UK. ⁵ This increased disease and risk factor exposure in ethnic minority populations has become an important factor in understanding differing COVID‐19 risk profiles by ethnicity.

A meta‐analysis we published early in the pandemic reported on pooled prevalence of diabetes and other cardiovascular diseases in people with COVID‐19, showing that of those admitted to hospital with COVID‐19, approximately 23% had hypertension, 11% diabetes and around 10% cardiovascular disease. This highlighted that COVID‐19 was indeed affecting people with cardiometabolic diseases. ⁶ Furthermore, this meta‐analysis also showed that in individuals with COVID‐19, the presence of these comorbidities was associated with severe COVID‐19 outcomes including mortality. ⁶

Further analyses using National Diabetes Audit data reported that the adjusted hazard ratio for COVID‐19 related death for people with type 1 (T1DM) and T2DM, particularly T2DM, was exceptionally high in Asian and Black populations, even after adjustment for various confounding variables. ⁷ Similarly, using the largest secondary data analysis involving 61 million people in the UK, we also showed that Asian, Black, and Mixed ethnic groups had a much higher hospital related COVID‐19 mortality compared to White populations. ⁸ We replicated this work using the large OpenSAFELY data set of 17 million individuals from the UK and found that there was a higher association for hospital related admission for COVID‐19 and intensive care unit admission for COVID‐19 in ethnic minority groups, and that COVID related mortality was higher by about 20–50% in ethnic minority groups compared to the White population. ⁹

As data availability grew, a number of systematic reviews were conducted. One such review, which included over 50 studies and 18 million patients, reported greater rates of intensive care unit (ICU) admission in South Asian ethnic groups, and a 26% increased rate of ICU admission for non‐White compared to White populations. Importantly, mortality risk was not increased for any ethnic minority populations, suggesting that although ICU admission was higher, improved care and delivery of care received had led to improved outcomes. ²

Complications of COVID‐19 in ethnic minority populations

In relation to COVID‐19 complications in ethnic minority populations, our research group has previously shown that cardiovascular events (e.g. myocardial infarction and all cardiovascular disease) and mortality are higher in people with T2DM compared to those without, regardless of ethnicity. However, the cardiovascular event risk associated with T2DM was markedly higher in South Asian compared to White groups, mainly driven by increased risk of myocardial infarction. ¹⁰ Our previous work in diabetes populations also reported that the incidence of end‐stage kidney disease is higher in Black individuals, but not South Asian. ¹¹

These ethnic inequalities in disease risk have continued in COVID‐19 infected populations. Analysis of data from the International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) cohort on 64,000 people admitted to UK hospitals found that South Asian and Black individuals had greater risk of any cardiovascular or renal complications. In comparison to the White group, South Asians also displayed greater risk of cardiac ischaemia, cardiac arrest and renal injury, whereas Black populations displayed greater risk of coagulation complications and renal injury. ¹² Further work from our team using UK Biobank data has found that COVID‐19 hospitalisation and mortality, in those with and without chronic kidney disease (CKD), was higher in people of Black and South Asian ethnicity whether they had CKD or not. ¹³

Beyond acute complications, the UK Office for National Statistics (ONS) have carried out a programme of work on the long‐term persistence of COVID‐19. In a study of 47,000 COVID‐19 patients who were discharged from hospital, compared to a group who did not have COVID‐19, 29.4% of patients were readmitted and around 12% had died at a mean follow‐up of 140 days. Readmission rates were higher in non‐White populations, and there were higher cardiovascular event rates in non‐White populations, but these were not significantly different compared to the White group. ¹⁴

Prior to the pandemic, our group studied the patterning of multiple long‐term conditions (multimorbidity) and depression by ethnic group, finding that the prevalence of depression is high in South Asians, and even greater in those with T2DM. Utilising a US dataset of adults with T2DM, we showed that cardiometabolic multimorbidity and depression are associated with greater risk of cardiovascular disease and major adverse cardiovascular events in those with and without depression. The risk seems to be substantially greater in younger groups compared to older groups, showing that young T2DM is quite a severe phenotype. ¹⁵ During the pandemic, our work as part of the UK‐REACH study on the mental health of health care workers has also reported increased post‐traumatic syndromes (adjusted for various confounders) in ethnic minority workers compared to White populations. ¹⁶

In relation to risk factors for severe COVID‐19 outcomes, there has been a range of work conducted on multiple long‐term conditions. Using UK Biobank data, Chudasama et al. found that multiple long‐term conditions were associated with worse outcomes as a consequence of severe COVID‐19 infection in ethnic minority groups compared to White. The highest risk of infection, in those with CKD and diabetes, were seen in ethnic minority populations. ¹⁷ Further analyses using ISARIC data on the impact of cardiometabolic multimorbidity and ethnicity on cardiovascular and renal complications, showed that having one, two or three multiple long‐term conditions is associated with increased risk of any cardiovascular or renal hospital outcomes and all‐cause death. Across all‐cause outcomes, there was a gradual increase in risk with increasing number of comorbidities, but there were no interactions in ethnic minority groups compared to the White European population. ¹²

Pre‐existing ethnic health and social disparities

A range of analyses conducted over previous decades have demonstrated that there is an interaction between ethnicity, deprivation and T2DM. For example Coles et al. utilised UK primary care records (Clinical Practice Research Datalink, CPRD), and found that White ethnic groups tend to reside in less deprived areas than both Black and South Asian groups. Similarly, age at diagnosis (of T2DM) was substantially lower in ethnic minority groups compared to White. ¹⁰ In relation to deprivation, ethnicity and COVID‐19, our research group has conducted a range of work in this area. Utilising UK Biobank data, Razieh et al. reported that people of South Asian and Black ethnic groups are more likely to reside in more deprived areas (calculated using the Townsend Score), and cases of severe COVID‐19 and COVID‐19 mortality were higher among ethnic minority groups. ¹⁸ ONS colleagues then extended this work by conducting mediation analysis on data from over 10 million people across England, finding that the main predictors of increased mortality were geographical location and socio‐economic status. Additional predictive factors included multi‐generational households, and associations for all of these factors were stronger in ethnic minority populations, particularly in the Bangladeshi, Pakistani and Black groups. ¹⁹

It has long been known that the age of onset of diabetes is much lower in ethnic minority populations. However, there are more contemporary data that suggest that in comparison to White populations South Asians are about 12 years younger when they develop diabetes, and Black populations are 10 years younger, and at a lower BMI but higher HbA_1c and living in more deprived areas. ²⁰ In relation to COVID‐19 we have observed the same pattern where non‐White populations have a much lower age and lower BMI, for severe infection and adverse outcomes, compared to the White population. ¹⁸

There are likely multiple factors that may contribute to this increased risk. For example, analyses of physical activity data suggest that South Asian populations are less physically active than White, and they display 40% less total physical activity compared to White populations when measured using device‐based measures (i.e. accelerometry). In addition, and overall, South Asians are two to three times less likely to meet the weekly physical activity guidelines compared to White populations. ²¹ Further device‐based estimates of physical activity also show that South Asian and Black populations are less physically active, and during the pandemic ethnic minority groups were even less active, which is of concern regarding the future negative impact on health outcomes. ²²

A second contributory risk factor is increased body fatness. For a number of years we have examined ethnic differences in body composition and obesity cut points for South Asians both in the UK and India compared to White populations. This work has found that glycaemia, at an equivalent BMI for a White European of obesity cut point of 30kg/m², was evident at a BMI of around 23kg/m² for South Asians. These findings informed NICE guidance for BMI waist circumference cut points for ethnic minority populations. ²³ , ²⁴ In relation to COVID‐19 risk, in a large English population using ONS data, BMI was associated with COVID‐19 mortality among all ethnic groups, but the positive association appeared to be stronger in ethnic minority groups. ²⁵ The nadir for mortality for White populations was at a BMI of 27kg/m² while it was about 23–26kg/m² for ethnic minority populations. The hazard ratio for COVID‐19 related mortality at a BMI cut point for White groups of 40kg/m² was found at a significantly lower BMI of around 27kg/m² for South Asians and around 30kg/m² for Black ethnic groups. ²⁵

Sleep quality may also be a relevant factor in the development of an increased risk for poorer COVID‐19 outcomes. Data collected prior to the pandemic using accelerometers suggest that sleep quality is poorest in South Asians, and sleep variability is greatest in South Asian men and women, compared to other ethnic groups. These sleep indices are also worse for Black compared to White populations. ²⁶ In relation to COVID‐19 mortality, and using UK Biobank data, Rowlands et al. found that for health care workers there is an increased association with the severity of COVID‐19 outcomes such as hospitalisation and mortality, regardless of ethnic group. The strength of association is greater, however, for South Asian and Black populations, particularly for those who were both a health care and shift worker. ²⁷

Other pre‐pandemic studies from our team have focused on hospital admissions for hypoglycaemia. It appears most ethnic minority populations seem to have lower odds of admission to hospital with hypoglycaemia, except for the Caribbean group who display higher odds compared to White ethnic groups. ²⁸ Extending this work during the pandemic to examine hypoglycaemia and hyperglycaemia by ethnic groups (for patients admitted to hospital with COVID‐19), we found there was no significant difference in rates of hypoglycaemia between ethnic minority groups and White populations, but admission glucose was significantly higher in Black and South Asian populations.

In terms of cardiovascular risk factors, prior to the pandemic, Gholap et al. examined survival of South Asian and White European patients after myocardial infarction, observing that South Asian patients were younger and, in a fully adjusted model, survival rates following myocardial infarction were similar between South Asian and White Europeans. ²⁹ Data from the ISARIC cohort, with 36,000 individuals admitted to hospital and their admission glucose across England and Scotland during the COVID‐19 pandemic, reveal evidence for a modifying effect of ethnicity for heart failure and cardiac arrest, but not for any other cardiovascular outcomes. ³⁰

As well as variations in health outcomes prior to the pandemic, various groups have studied differences in care provision and implementation of therapies. During the introduction of pay for performance (Quality and Outcomes Framework, QOF) within primary care in the UK and despite the intention of this scheme to reduce unwarranted variations in care, ethnic minority groups – particularly the Indian, Pakistani and Bangladeshi groups – were less likely to be initiated on insulin following the introduction of pay for performance compared to prior. ³¹ Similar data, focused on the delay in first‐line treatment in newly diagnosed people with T2DM, also reveal that Black ethnic groups were 14% less likely to get their first medication following diagnosis compared to the White European. There was, however, no difference observed for South Asian populations. ³² Recent National Diabetes Audit data from the UK also demonstrate variations in glucose‐lowering therapy prescription prior to the pandemic. Asian populations are more likely to receive dipeptidyl peptidase‐4 inhibitors or sulphonylureas and meglitinide compared to White ethnic groups, and they are less likely to be prescribed glucagon‐like peptide‐1 (GLP‐1) receptor agonists. Black populations appear less likely to receive sodium‐glucose co‐transporter‐2 inhibitors and GLP‐1 receptor agonists, but more likely to be prescribed meglitinide. ³³

In children and young people, ethnic variations in health behaviours also exist. A number of studies from our group suggest that there is low physical activity and high sedentary behaviours and cardiometabolic risk in South Asian schoolchildren in comparison to their White peers. ³⁴ , ³⁵ , ³⁶ Likewise for body fatness outcomes, for a given equivalent BMI, percent body fat has been shown to be higher in South Asian children compared to children of White ethnicity. ³⁷ Children and young people were, of course, also affected by COVID‐19. In a large analysis of over 2.6 million children in England (using the QResearch database) with COVID‐19, results showed a higher risk of a positive COVID‐19 test result in Asian and Black populations, and critically greater levels of mortality in these ethnic minority populations. ³⁸ Ethnic variation in COVID‐19 outcomes clearly affects populations across the life course.

One way of identifying individuals at risk of poorer outcomes is to develop risk assessment frameworks. Our team have developed a number of risk scores, including the first multi‐ethnic diabetes risk score which has been rolled out nationally via paper and automatic practice computer‐based screening for people with or at high risk of T2DM. ³⁹ The Leicester Diabetes self‐assessment Risk Score has since been used by over 2 million people on the Diabetes UK website and has been translated into a range of South Asian languages to increase accessibility for ethnic minority populations. ⁴⁰ Utilising our expertise in risk scoring, our team has contributed to the development of the QCovid Risk score, which included ethnicity, along with a range of other risk factors to predict the likelihood of COVID‐19 infection. This scoring algorithm was used to prioritise targeting of the national vaccination programme across England. ⁴¹

Cultural considerations in COVID‐19 research and COVID‐19 response efforts

Knowledge, beliefs and attitudes of and towards lifestyle behaviours and other risk factors for cardiometabolic disorders are an area of concern among ethnic minority populations, and our Centre for Ethnic Health Research in Leicester has worked for many years to develop culturally adapted self‐management interventions. For example, we studied concerns and perceptions of insulin therapy for people with both T1DM and T2DM and we identified numerous barriers to injectable therapies in ethnic minority populations. ⁴² , ⁴³ The implementation of COVID‐19 infection prevention measures, as well as more‐targeted interventions such as vaccination, has therefore required careful consideration of culturally‐driven variance in beliefs, knowledge and attitudes between ethnic groups. ⁴⁴ As an example, we utilised close and long‐standing relationships with the multi‐ethnic community in the City of Leicester, as well as large data analyses of ethnic inequality in COVID‐19 vaccination coverage (using ONS data) to identify that ethnic minority groups (South Asian and Black) were significantly less likely to have a COVID‐19 vaccination compared to White populations. ⁴⁵ To address this locally, qualitative interviews and focus groups were run with local ethnic minority populations and stakeholders to identify barriers and facilitators to vaccine uptake (some of which were similar to injectable therapies), which have then been fed back into the national vaccination programme via the Cabinet Office and national Clinical Research Networks. ⁴⁶

In terms of developing interventions, our research groups have employed previous learning from many years of work with ethnic minority populations such as barriers to insulin use ⁴⁷ to generate a number of recommendations relating to the management (and self‐management) of diabetes during the pandemic. These have largely focused on the need to utilise approaches that are culturally competent/sensitive and that draw upon culturally adapted materials (e.g. multiple languages and written and audio formats) to increase uptake and engagement with developed guidance. For example, the South Asian Health Foundation (SAHF) has worked in conjunction with the Centre for Ethnic Health to develop a series of infographics and videos in different languages focused on COVID‐19 protection measures and the self‐management of diabetes during the pandemic. These have been disseminated nationally and internationally, and drew upon previously developed recommendations for people fasting during Ramadan. Therefore when Ramadan fell during the COVID‐19 pandemic, the SAHF were in a position to rapidly develop guidance for people fasting during Ramadan in the context of the COVID‐19 pandemic. ⁴⁸

Throughout the pandemic, and with the renewed focus on ethnic health inequalities, the issue of minority inclusion within health research and how ethnicity is conceptualised and referred to has received particular attention. Prior to the pandemic, previous work has shown the ethnic minority populations are not well represented in health research, which limits the applicability of evidence generated from trials and evidence reviews. ⁴⁹ The issue of low ethnic minority representation has not changed during the pandemic. For example, a systematic review of all COVID‐19 studies registered on clinicaltrials.gov reported that only six out of the 1518 studies were even collecting data on ethnicity. ⁵⁰ In the UK, our Centre for Ethnic Health Research worked with the National Clinical Research Network to increase uptake to COVID‐19 trials. There are also now several frameworks that can help researchers with minority inclusion. For example, the INCLUDE Ethnicity Framework (www.trialforge.org/trial-forge-centre/include), which was co‐developed by the National Institute for Health Research, the Medical Research Council and the Centre for Ethnic Health Research, can aid research teams on how to think about the ethnicity of people who need to be involved in a study, and how to include these populations. ⁵¹ In relation to language, early COVID‐19 analysis utilised the phrase Black, Asian and Minority Ethnic (BAME), before later being phased out and replaced by ‘Ethnic Minority groups’ as a result of public feedback and consultation regarding the use of the terminology BAME. ⁵² , ⁵³ Likewise, there has been greater emphasis on the importance of standardised and more expansive and specific ethnicity coding to ensure that ethnic inequalities are not masked by high level groupings (for example, breaking down South Asian into Indian, Pakistani, Sri Lankan). ⁵⁴

Summary

Prior to the pandemic a number of research groups worked together to examine potential mechanisms for increased risk of T2DM in ethnic minority groups and there are clearly a wide range of complex and interacting contributory factors. ⁵⁵ , ⁵⁶ Likewise, the mechanisms behind the greater burden of COVID‐19 in ethnic minority populations is particularly complex. ⁵⁷ Increased risk is likely due to poor control measures, increased exposure to the virus, increased vulnerabilities to disease, and residence in worse physical environments and social conditions. ¹

The latter may be particularly influential. In a mediation analysis, when moving 25% of the most deprived populations from deprivation in a substitution analysis, there is around a 50% reduced risk of COVID‐19 infection, development of severe disease and mortality. When relocating 50% of the population from the most deprived areas, the increased risk of severe disease and COVID‐19 mortality is almost entirely removed.

Clearly, there is growing evidence of marked ethnic inequalities in diabetes which have been exacerbated with COVID‐19. Due to the complexity of contributory mechanisms, there is no single explanation. Social inequalities are likely to play a major role, as are structural drivers of risk behaviours. To tackle these structural inequalities we need a comprehensive multi‐sectoral approach supported by strong policy action. In particular, we need to ensure disparities are not widened following our recovery. In terms of research on ethnicity and diabetes, while we in the UK have the best research data globally and the research methodologies and hypothesis currently being generated are transferable to other diseases, what is most important is strong research collaboration. Some 18 years ago we wrote that if the epidemic of coronary heart disease in South Asians is a disease of migration and occurred within a generation, there is no reason why we should not be able to reverse it. ⁵⁸ I still believe this holds true.

Declaration of interests

KK is Director of the University of Leicester Centre for Ethnic Health Research, Trustee of the South Asian Health Foundation, Chair of the Ethnicity Subgroup of the UK Scientific Advisory Group for Emergencies (SAGE), and Member of SAGE.

KK has acted as a consultant, speaker or received grants for investigator‐initiated studies on behalf of Astra Zeneca, Bayer, Novartis, Novo Nordisk, Sanofi‐Aventis, Lilly, Merck Sharp & Dohme, Boehringer Ingelheim, and Bayer.