Letters

https://doi.org/10.1038/s42255-020-00281-8

Autoantibody-negative insulin-dependent
diabetes mellitus after SARS-CoV-2 infection: a
case report
Tim Hollstein 1,8, Dominik M. Schulte1,8, Juliane Schulz1, Andreas Glück2, Anette G. Ziegler3,
Ezio Bonifacio4, Mareike Wendorff 5, Andre Franke 5, Stefan Schreiber1,2,5, Stefan R. Bornstein6,7
and Matthias Laudes 1 ✉

Here we report a case where the manifestations of with acute diabetic ketoacidosis (DKA) associated with COVID-19
insulin-dependent diabetes occurred following SARS-CoV-2 disease9. The exact time course, causal relationship and pres-
infection in a young individual in the absence of autoanti- ence or absence of autoantibodies were, however, not provided.
bodies typical for type 1 diabetes mellitus. Specifically, a Furthermore, a marked increase in DKA was observed in German
19-year-old white male presented at our emergency depart- children and adolescents during the COVID-19 pandemic10, sug-
ment with diabetic ketoacidosis, C-peptide level of 0.62 µg l–1, gesting a relationship between COVID-19 and new-onset type 1
blood glucose concentration of 30.6 mmol l–1 (552 mg dl–1) diabetes mellitus (T1DM). Therefore, we recommended care-
and haemoglobin A1c of 16.8%. The patient´s case history ful management of patients with diabetes and monitoring for
revealed probable COVID-19 infection 5–7 weeks before new-onset diabetes during the pandemic11.
admission, based on a positive test for antibodies against Here we present the case of a 19-year-old white male patient
SARS-CoV-2 proteins as determined by enzyme-linked immu- admitted to our emergency department with abnormal fatigue,
nosorbent assay. Interestingly, the patient carried a human exhaustion and 12-kg weight loss over several weeks. A detailed
leukocyte antigen genotype (HLA DR1-DR3-DQ2) consi­ timeline of events before presentation at our emergency ward is
dered to provide only a slightly elevated risk of developing presented in Fig. 1. He exhibited increased polydipsia of ~6 l d–1,
autoimmune type 1 diabetes mellitus. However, as noted, nycturia (2–3 times per night) and an intermittent postprandial
no serum autoantibodies were observed against islet cells, left-sided flank pain. Neither fever episodes nor typical chest pain
glutamic acid decarboxylase, tyrosine phosphatase, insulin was reported. Laboratory testing in our emergency department
and zinc-transporter 8. Although our report cannot fully revealed DKA with blood pH 7.1, blood glucose 30.6 mmol l–1
establish causality between COVID-19 and the development (552 mg dl–1), a reduced serum C-peptide level of 0.62 µg l–1 (normal
of diabetes in this patient, considering that SARS-CoV-2 range, 1.1–4.4 µg l–1) and haemoglobin A1c (HbA1c) 16.8%, as
entry receptors, including angiotensin-converting enzyme 2, well as positive urinary ketones and glucosuria. Type 1 diabetes
are expressed on pancreatic β-cells and, given the circum- mellitus was assumed. The family history revealed a maternal
stances of this case, we suggest that SARS-CoV-2 infection, or cousin with autoantibody-positive T1DM and a maternal grand-
COVID-19, might negatively affect pancreatic function, per- mother with type 2 diabetes. Human leukocyte antigen (HLA)
haps through direct cytolytic effects of the virus on β-cells. geno­typing revealed that the patient had no high-risk HLA geno-
The recent COVID-19 pandemic caused by the SARS-CoV-2 type but a DR1-DR3-DQ2 genotype, which is associated with a
virus represents a worldwide health crisis causing severe illness slightly elevated risk of developing autoimmune T1DM (around
and death, especially in people with cardiovascular and metabolic 1.7-fold higher compared with the general population)12,13.
abnormalities1,2. SARS-CoV-2 enters human cells via angiotensin- However, immunological examination yielded an absence of serum
converting enzyme 2 (ACE2)3, a transmembrane glycoprotein with autoantibodies against islet cells (IC-Ab), glutamic acid decarboxy­
proteolytic activity also found in human pancreatic β-cells4, sug- lase (GAD65-Ab), tyrosine phosphatase (IA-2-Ab), insulin (I-Ab)
gesting that SARS-CoV-2 might alter pancreatic β-cell function and zinc-transporter 8 (ZnT8-Ab) in the affected patient (Table 1),
and impair insulin secretion. Several recently published studies suggesting a type 1B diabetes mellitus subtype14.
indicate a link between COVID-19 and diabetes: for example, acute The patient reported that he had had an asymptomatic
hyper­glycaemia has been observed in a large number of individu- SARS-CoV-2 infection 5–7 weeks previously when returning from
als infected with SARS-CoV-2, regardless of any past medical his- vacation in Austria with his family. On 29 April 2020 he tested
tory of diabetes5–8. In another study in Asia, patients were reported positive for IgG—but not IgM—antibodies against SARS-CoV-2

1
Division of Endocrinology, Diabetes and Clinical Nutrition, Department of Internal Medicine I, University Medical Centre Schleswig-Holstein, Kiel,
Germany. 2Division of Critical Care, Department of Internal Medicine I, University Medical Centre Schleswig-Holstein, Kiel, Germany. 3Institute of
Diabetes Research, Helmholtz Zentrum München and Forschergruppe Diabetes, Klinikum rechts der Isar, Technische Universität München, Neuherberg,
Germany. 4Center for Regenerative Therapies Dresden and Paul Langerhans Institute Dresden, German Center for Diabetes Research, Dresden University
of Technology, Dresden, Germany. 5Institute of Clinical Molecular Biology, Christian-Albrechts University of Kiel, Kiel, Germany. 6Department of
Internal Medicine III, University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany. 7Department of Diabetes, School of
Life Course Science and Medicine, King’s College London, London, UK. 8These authors contributed equally: Tim Hollstein, Dominik M. Schulte.
✉e-mail: matthias.laudes@uksh.de

Nature Metabolism | VOL 2 | October 2020 | 1021–1024 | www.nature.com/natmetab 1021

Letters Nature Metabolism

Week –7 Week –6 Week –5 Week –4 Week –3 Week –2 Week –1 Week 0

Unlikely infection period

Probable infection period
(due to COVID-19 IgM–)

16 March 2020: Excessive weight loss

parents experience Fatigue
first COVID-19 Polydipsia
symptoms
Polyuria

14 March 2020: 6 April 2020: 29 April 2020: 5 May 2020:

parents of patient onset of typical antibody test presentation at
return from Austria diabetes-related (IgG+, IgM–) emergency ward:
symptoms diagnosis of
insulin-dependent
diabetes mellitus

Fig. 1 | Timeline of events before diagnosis of insulin-dependent diabetes mellitus. On 14 March 2020, the patient’s parents had returned from a vacation
in Austria. Two days later, both parents started to develop COVID-19-typical symptoms (dry cough, shivering, fatigue, dyspnoea, joint pain and loss of
smell and taste). No further PCR testing was performed because, at the time, official authorities did not invite them for testing despite both parents
having reported their symptoms. On 6 April 2020, their 19-year-old son (the patient in this report) first noticed symptoms related to diabetes mellitus
including fatigue, polydipsia and polyuria, which worsened over time. He did not show any typical COVID-19 symptoms. Around 20 April 2020, he further
noticed excessive weight loss. In the meantime, both parents started to recover from their complaints. As they were suspected of having COVID-19, both
parents and their two sons underwent a SARS-CoV-2 antibody test on 29 April 2020, which was positive (IgG+, IgM–) for both parents and the patient.
The dizygotic sibling of the patient tested negative for SARS-CoV-2 antibodies and experienced neither COVID-19- nor diabetes-related symptoms. On
5 May 2020 the patient presented at our local emergency ward because his symptoms relating to diabetes mellitus had worsened. He was then diagnosed
with insulin-dependent diabetes mellitus and received treatment according to international guidelines. Based on the information presented in this figure
(adapted from reports on the patient and his parents, and from antibody test results), we assume that the possible infection period of the patient can be
narrowed down to the last 2 weeks of March 2020 (yellow bar) while it is unlikely that the patient had COVID-19 in April 2020 (red bar). This is further
supported by the absence of IgM antibodies detected in the patient’s SARS-CoV-2 antibody test, which have previously been shown to persist for up to
4 weeks after infection with SARS-CoV-2 (ref. 15).

It has been known for several years that corona-like viruses

Table 1 | Immunity diagnostics enter human cells via binding to membrane-bound proteases3.
Result (IU ml–1) Reference range For example, the Middle East respiratory syndrome virus, respon-
(IU ml–1) sible for an outbreak of acute respiratory syndrome, has been shown
to use dipeptidylpeptidase-4, a protease known to be involved
GAD65-Ab <5 <10
in the regulation of the incretin system16. Recently it has been
IA-2-Ab <10 <10 demonstrated in a functional assay that ACE2 acts as the major
IC-Ab Negative Negative binding partner for the SARS-CoV-2 spike glycoprotein, mediating
I-Ab <0.1 <0.4 host cell internalization17.
ACE2 is widely expressed in eukaryotic cell membranes, including
ZnT8-Ab <10 <15 those of pancreatic β-cells in mice18–20 and humans4. Using a
COVID-19 (SARS-CoV-2) IgG Positive Negative high-fat diet mouse model, it was shown that β-cell de-differentiation
COVID-19 (SARS-CoV-2) IgM Negative Negative is accompanied by reduction in ACE2 (ref. 20). Deletion of ACE2
in non-obese diabetic mice resulted in hyperglycaemia, decreased
β-cell insulin content and increased β-cell oxidative stress19. In
addition, ACE2 deficiency reduced β-cell mass and impaired
(Table 1), indicating that he had COVID-19 disease >4 weeks before β-cell proliferation in obese C57BL/6 mice18, suggesting overall
antibody assessment15 (Fig. 1). that this glycoprotein is important in β-cell homeostasis. Finally,
Acute pancreatitis was ruled out based on clinical assessment it is currently under debate whether ACE2 is important in regard
by experienced emergency room staff (no history of alcohol to intra-islet paracrine mechanisms in communication by α
or drug abuse, no history of gallstones and no lipid disorders) and β-cells19.
and normal serum lipase of 19 U l–1 (reference range, 13–60 U l–1). The hallmark in the pathology of classical T1DM is β-cell destruc-
Exocrine pancreatic function was not assessed because the patient tion caused by a complex autoimmune process21. Environmental
did not complain of diarrhoea, increased stool volume and/or factors are suspected to be responsible, involving activation of
fatty stools. the immune system by reduction in gut microbiota, by the early
The patient needed treatment in the intensive care unit for introduction to fruit or cow’s milk during childhood, by gluten, by
3 days for recompensating DKA, and was administered intravenous toxins and, especially, by viruses22. Interestingly, besides the induc-
insulin therapy according to international guidelines. At day 4, he tion of autoimmunity, some viruses including enteroviruses might
was transferred to the endocrine ward, and subcutaneous insulin also exert direct cytolytic effects on pancreatic β-cells23.
therapy was initiated. Over the course of the next few days, blood The patient reported here presented several weeks after a viral
glucose levels stabilized at 8.4–10.2 mmol l–1 (reference range, infection with severe DKA and profound loss of β-cell function.
151–183 mg dl–1). The patient received an educational programme However, in contrast to the general idea that viral infections indi-
according to guidelines for insulin-dependent diabetes mellitus rectly affect pancreatic β-cells by triggering autoimmunity, the
management. He was discharged from hospital in good condition absence of five typical antibodies in our patient might argue against
after 10 days. classical autoimmune T1DM although autoantibody-negative

1022 Nature Metabolism | VOL 2 | October 2020 | 1021–1024 | www.nature.com/natmetab

Nature Metabolism Letters
T1DM (type 1B subtype) is not a rare condition24. Given the fact in the treatment of the patient or analysis of the available biomaterial. The patient
that ACE2 is expressed on human pancreatic β-cells4 and that it is provided written informed consent for publication of the information presented in
this case report. Patient care and research were conducted in compliance with case
the main receptor for SARS-CoV-2 internalization, we propose that, report guidelines and the Declaration of Helsinki.
in this patient, direct cytolytic β-cell damage due to SARS-CoV-2
infection resulted in insulin-dependent diabetes with no classical Diagnostics. Blood and urine samples (including serum autoantibodies against
autoimmune pathology evident. This assumption is supported by islet cells, GAD65, IA-2, insulin and ZnT8) were analysed at the Department of
Clinical Chemistry and Laboratory Medicine at the University of Kiel, Germany.
a recent mechanistic study demonstrating that adult human pan-
Autoantibody tests for GAD65, IA-2 and ZnT8 were repeated in another laboratory
creatic α- and β-cells are permissive to SARS-CoV-2 pseudo-entry (Helmholtz Institute, Munich, Germany) using internationally standardized and
virus and SARS-CoV-2 virus infection25. Furthermore, SARS-CoV-2 evaluated assays31. These included the 3-Screen enzyme-linked immunosorbent
infection of pancreatic endocrine cells resulted in robust chemokine assay (ELISA) that measures antibodies against GAD65, IA-2 and ZnT8 (RSR)32,
induction, as seen in patients with COVID-19 and upregulation of and radiobinding immunoprecipitation assays with [35S]methionine-labelled
recombinant human proteins measuring antibodies against GAD65, IA-2 and
markers of cell death25; in addition, SARS-CoV-1, the predecessor ZnT8 (refs. 33,34).
of SARS-CoV-2, has been shown to damage islet cells and cause Antibodies against SARS-CoV-2 were determined by an external laboratory
diabetes in humans26. Based on these findings, we hypothesize that using the EDI Novel Coronavirus COVID-19 IgG ELISA Kit (Epitope Diagnostics).
COVID-19 might have caused T1DM in our patient. According to the manufacturer, this assay has a limit of detection of 5 U ml–1, a
Nonetheless, our case report has limitations. We are aware that diagnostic specificity of 100% and a diagnostic sensitivity of 100%.
Further information on the methods used in this manuscript can be found in
the major limitation of this case report lies in the causality miss- the Reporting Summary in the Supplementary information.
ing between SARS-CoV-2 infection and insulin-dependent diabe-
tes. Future studies are warranted to investigate the direct cytotoxic Reporting Summary. Further information on research design is available in the
effects of SARS-CoV-2 on pancreatic islet cells. Furthermore, one Reporting Summary linked to this article.
might argue that the high HbA1c level of our patient at the time of
diagnosis does not support a recently developed T1DM due to Data availability
COVID-19. However, although a high HbA1c level may suggest The authors declare that the data supporting the findings of this study are available
within the paper.
previous hyperglycaemia of several weeks’ duration, diabetic keto-
acidosis is reported with very high HbA1c levels irrespective of the Received: 29 June 2020; Accepted: 18 August 2020;
duration of diabetes in several studies27,28. There are also well-known Published online: 2 September 2020
variations in red blood cell survival that may cause variation in
HbA1c levels for a given mean blood glucose29. These studies indi- References
cate that the markedly elevated HbA1c level in our patient does not 1. Tay, M. Z., Poh, C. M., Rénia, L., Macary, P. A. & Ng, L. F. P. The trinity of
rule out an acute onset of T1DM within the reported time frame. COVID-19: immunity, inflammation and intervention. Nat. Rev. Immunol.
This is supported by the reported onset of diabetes symptoms that 20, 363–374 (2020).
occurred after the probable period of infection with SARS-CoV-2 2. Velavan, T. P. & Meyer, C. G. The COVID-19 epidemic. Trop. Med. Int. Health
25, 278–280 (2020).
(Fig. 1). The latter can be narrowed down to a time between 3. Hoffmann, M. et al. SARS-CoV-2 cell entry depends on ACE2 and
mid-March and early April 2020, due to the positive IgG, but nega- TMPRSS2 and is blocked by a clinically proven protease inhibitor. Cell 181,
tive IgM, in the SARS-CoV-2 antibody test15. 271–280. (2020).
Despite tremendous progress in biomedical research over the 4. Fignani, D. et al. SARS-CoV-2 receptor angiotensin I-converting enzyme
centuries, the complex immunopathology of T1DM is still not type 2 is expressed in human pancreatic islet β-cells and is upregulated by
inflammatory stress. Preprint at bioRxiv https://doi.org/10.1101/2020.07.
completely understood. Detailed understanding of the induction 23.208041 (2020).
and progression of pancreatic β-cell failure, however, is crucial 5. Chen, N. et al. Epidemiological and clinical characteristics of 99 cases of
in the development of therapies to prevent and/or cure this impor- 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study.
tant metabolic disease. In that regard, the detailed examina- Lancet 395, 507–513 (2020).
6. Iacobellis, G., Penaherrera, C. A., Bermudez, L. E. & Bernal Mizrachi, E.
tion of patients developing β-cell failure and insulin-dependent
Admission hyperglycemia and radiological findings of SARS-CoV-2
diabetes after SARS-CoV-2 infection might yield new insights in patients with and without diabetes. Diabetes Res. Clin. Pract. 164,
into the hitherto-unknown role of ACE2 in β-cell function, 108185 (2020).
and also into the potential virus-induced cytolytic processes in 7. Sardu, C. et al. Outcomes in patients with hyperglycemia affected by
T1DM development. Thus, a global registry has been launched COVID-19: can we do more on glycemic control? Diabetes Care 43,
1408–1415 (2020).
to provide a systematic platform for collection and analysis of 8. Wu, J. et al. Elevation of blood glucose level predicts worse outcomes
patients with new-onset diabetes30. It will be critical to explore in hospitalized patients with COVID-19: a retrospective cohort study.
the frequency, severity and duration of these complications BMJ Open Diabetes Res. Care 8, e001476 (2020).
and to investigate more deeply the mechanisms involved in 9. Chee, Y. J., Ng, S. J. H. & Yeoh, E. Diabetic ketoacidosis precipitated by
primary human tissue. Because our case may illustrate the sys- Covid-19 in a patient with newly diagnosed diabetes mellitus. Diabetes Res.
Clin. Pract. 164, 108166 (2020).
temic nature of COVID-19-related organ damage, we think that a 10. Kamrath, C. et al. Ketoacidosis in children and adolescents with newly
population-representative assessment of endocrine syndromes post diagnosed type 1 diabetes during the COVID-19 pandemic in Germany.
COVID-19 would be necessary. The recently launched COVIDOM JAMA 324, 801–804 (2020).
study of University Hospital Schleswig-Holstein represents an 11. Bornstein, S. R. et al. Practical recommendations for the management
epidemiological campaign with the scope of a 10-year follow up of diabetes in patients with COVID-19. Lancet Diabetes Endocrinol. 8,
546–550 (2020).
after SARS-CoV-2 infection to assess the long-term consequences 12. Sabbah, E. et al. Glutamic acid decarboxylase antibodies in relation to other
of COVID-19 in a larger population. autoantibodies and genetic risk markers in children with newly diagnosed
As a conclusion, our report indicates that diabetologists should insulin-dependent diabetes. Childhood Diabetes in Finland Study Group.
be aware of the possibility of insulin-dependent diabetes as an acute J. Clin. Endocrinol. Metab. 81, 2455–2459 (1996).
complication in patients infected with SARS-CoV-2. 13. Lambert, A. P. et al. Absolute risk of childhood-onset type 1 diabetes defined
by human leukocyte antigen class II genotype: a population-based study in
the United Kingdom. J. Clin. Endocrinol. Metab. 89, 4037–4043 (2004).
Methods 14. Imagawa, A., Hanafusa, T., Miyagawa, J. & Matsuzawa, Y. A novel subtype of
Clinical study. The patient reported here was treated in both the intensive care unit type 1 diabetes mellitus characterized by a rapid onset and an absence of
and the Division of Endocrinology, Diabetes and Clinical Nutrition, Department diabetes-related antibodies. Osaka IDDM Study Group. N. Engl. J. Med. 342,
of Medicine 1 at the University of Kiel, Germany. All authors were directly involved 301–307 (2000).

Nature Metabolism | VOL 2 | October 2020 | 1021–1024 | www.nature.com/natmetab 1023

Letters Nature Metabolism
15. Liu, X. et al. Patterns of IgG and IgM antibody response in COVID-19 30. Rubino, F. et al. New-onset diabetes in Covid-19. N. Engl. J. Med. 383,
patients. Emerg. Microbes Infect. 9, 1269–1274 (2020). 789–790 (2020).
16. Fleischer, B. CD26: a surface protease involved in T-cell activation. Immunol. 31. Ziegler, A.-G. et al. Yield of a public health screening of children for islet
Today 15, 180–184 (1994). autoantibodies in Bavaria, Germany. JAMA 323, 339–351 (2020).
17. Letko, M., Marzi, A. & Munster, V. Functional assessment of cell entry and 32. Amoroso, M. et al. 3 Screen islet cell autoantibody ELISA: a sensitive and
receptor usage for SARS-CoV-2 and other lineage B betacoronaviruses. specific ELISA for the combined measurement of autoantibodies to GAD65,
Nat. Microbiol. 5, 562–569 (2020). to IA-2 and to ZnT8. Clin. Chim. Acta 462, 60–64 (2016).
18. Shoemaker, R., Yiannikouris, F., Thatcher, S. & Cassis, L. ACE2 deficiency 33. Achenbach, P. et al. Autoantibodies to zinc transporter 8 and SLC30A8
reduces β-cell mass and impairs β-cell proliferation in obese C57BL/6 mice. genotype stratify type 1 diabetes risk. Diabetologia 52, 1881–1888 (2009).
Am. J. Physiol. Endocrinol. Metab. 309, E621–E631 (2015). 34. Bonifacio, E. et al. Harmonization of glutamic acid decarboxylase and
19. Roca-Ho, H. et al. Angiotensin-converting enzyme 2 influences pancreatic islet antigen-2 autoantibody assays for National Institute of Diabetes and
and renal function in diabetic mice. Lab. Invest. 100, 1169–1183 (2020). Digestive and Kidney Diseases Consortia. J. Clin. Endocrinol. Metab. 95,
20. Xuan, X. et al. Activation of ACE2/angiotensin (1-7) attenuates pancreatic 3360–3367 (2010).
β cell dedifferentiation in a high-fat-diet mouse model. Metabolism 81,
83–96 (2018).
21. Saberzadeh-Ardestani, B. et al. Type 1 diabetes mellitus: cellular and Author contributions
molecular pathophysiology at a glance. Cell J. 20, 294–301 (2018). T.H. attended the patient on the ward and wrote the manuscript. D.M.S. attended the
22. Knip, M. & Simell, O. Environmental triggers of type 1 diabetes. Cold Spring patient in the emergency room and wrote the manuscript. J.S. and A.G. attended the
Harb. Perspect. Med. 2, a007690 (2012). patient in the intensive care unit and edited the manuscript. M.W. and A.F. performed
23. Rodriguez-Calvo, T., Sabouri, S., Anquetil, F. & von Herrath, M. G. The viral HLA genotyping and edited the manuscript. S.S. attended the patient in the emergency
paradigm in type 1 diabetes: who are the main suspects? Autoimmun. Rev. 15, room and edited the manuscript. E.B., A.G.Z. and S.R.B. performed further antibody
964–969 (2016). testing, critically reviewed the case and edited the manuscript. M.L. attended the patient
24. Wang, J. et al. Prevalence of autoantibody-negative diabetes is not rare at all on the ward and wrote the manuscript.
ages and increases with older age and obesity. J. Clin. Endocrinol. Metab. 92,
88–92 (2007). Competing interests
25. Yang, L. et al. A human pluripotent stem cell-based platform to study The authors declare no competing interests.
SARS-CoV-2 tropism and model virus infection in human cells and
organoids. Cell Stem Cell 27, 125–136 (2020).
26. Yang, J.-K., Lin, S.-S., Ji, X.-J. & Guo, L.-M. Binding of SARS coronavirus to Additional information
its receptor damages islets and causes acute diabetes. Acta Diabetol. 47, Supplementary information is available for this paper at https://doi.org/10.1038/
193–199 (2010). s42255-020-00281-8.
27. Ekpebegh, C., Longo-Mbenza, B. & Blanco-Blanco, E. Glycosylated
haemoglobin is markedly elevated in new and known diabetes patients with Correspondence and requests for materials should be addressed to M.L.
hyperglycaemic ketoacidosis. Afr. Health Sci. 14, 526–532 (2014). Peer review information Primary Handling Editor: Christoph Schmitt.
28. Henderson, D. C. et al. Elevated hemoglobin A1c as a possible indicator of Reprints and permissions information is available at www.nature.com/reprints.
diabetes mellitus and diabetic ketoacidosis in schizophrenia patients receiving
atypical antipsychotics. J. Clin. Psychiatry 68, 533–541 (2007). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
29. Cohen, R. M. et al. Red cell life span heterogeneity in hematologically normal published maps and institutional affiliations.
people is sufficient to alter HbA1c. Blood 112, 4284–4291 (2008). © The Author(s), under exclusive licence to Springer Nature Limited 2020

1024 Nature Metabolism | VOL 2 | October 2020 | 1021–1024 | www.nature.com/natmetab