https://doi.org/10.47108/jidhealth.Vol4.Iss4.168 Saha S, Saha S, Journal of Ideas in Health 2021;4(4):573-580 © The Author(s). 2021 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (https://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. e ISSN: 2645-9248 Journal homepage: www.jidhealth.com Open Access Underreporting of treatment outcomes in hospitalized COVID-19 infected diabetes patients: a systematic review, meta-analysis, and meta-regression Sumanta Saha*1, Sujata Saha2 Abstract Background: Prolonged inpatient care requirements and time constraints of research and researchers lead to the non-reporting of the treatment outcome of certain COVID-19 infected diabetes patients in published manuscripts. This study aims to quantify its global burden. Methods: A search for citations addressing the above outcome ensued chiefly in the PubMed, Embase, and Scopus databases, irrespective of the publication date and geographical region. Recruited studies were critically appraised with the National Heart, Lung, and Blood Institute's tool. Using the random-effects meta-analysis with an exact binomial method and Freeman-Tukey double arcsine transformation, the overall and subgroup-wise weighted pooled prevalence of the missing treatment outcome data was determined. The heterogeneity and publication bias assessment utilized I2 and Chi2 statistics, and funnel plot, and Egger's test, respectively. Results: Ten publications (primarily case series; 70.0%) included in this review sourced data from 6687 COVID-19 infected inpatient diabetes patients from Asia, Australia, Europe, and North America. The global pooled prevalence of missing treatment outcome data among these patients was 33.0% (95% CI: 15.0-53.0%; I2: 99.53%; P of Chi2: <0.001). It was highest in Europe (63%; 95% CI: 61.0-66.0%). Publication bias assessment was not suggestive of any small study effect. Conclusion: A considerable proportion of crucial prognosis information of hospitalized COVID-19 patients with diabetes goes underreported. It increases the risk of biasing the contemporary COVID-19-diabetes literature. The reporting of these data in the post-publication era or postponing the primary publication until the availability of all patients' treatment outcome data, when feasible, is recommended to address this enigma. Keywords: Coronavirus Infection, Diabetes Mellitus, Type 1, Type 2, Systematic Review, Meta Analysis, India Background The ongoing coronavirus disease (COVID-19) pandemic started in December 2019 in Wuhan, China [1–3]. As of April 08, 2021, almost 132 million confirmed cases of COVID-19 cases got reported globally, including about 2.8 million deaths [4]. One of the most commonly reported comorbidities determining the morbidity and mortality risk in COVID-19 patients is diabetes. Deaths among hospitalized COVID-19 patients with diabetes are substantial (almost 20% globally) and about two times higher than COVID-19 patients without diabetes [5]. Among hospitalized severe COVID-19 patients, deaths are commoner in those with diabetes than those without diabetes [5]. In the past, during the 2009-H1N1 pandemic influenza and the Middle East Respiratory Syndrome, diabetes was also a crucial determinant of death [6, 7]. The poor disease outcome in COVID-19 patients with diabetes is plausibly attributable to the damage of pancreatic islet cells caused by SARS-CoV-2 entry into the host cell via the angiotensin-converting enzyme-2 receptor [8–10]. Presently, little is known about the clinical outcomes and treatments of inpatient COVID-19 infected diabetes patients, and we must depend heavily on first-hand observational and case series studies for it. The entire COVID- 19 infected inpatient diabetes patient population's treatment outcome data (e.g., morbidity, mortality, recovery, and discharge) remain unavailable in some of these studies since some of these patients remain hospitalized when these studies' manuscripts are prepared or published. Such non-reporting may be due to the time constraints imposed by the study funder, the end of the pre-defined follow-up period of the study, and referral of severe COVID-19 cases to different health facilities ___________________________________________________ sumanta.saha@uq.net.au 1Department of Community Medicine, R. G. Kar Medical College, Kolkata, India. Full list of author information is available at the end of the article https://doi.org/10.47108/jidhealth.Vol4.Iss4.168 http://www.jidhealth.com/ Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 574 making their tracking difficult or impossible for the primary investigators. Quantifying the burden of such patients whose prognostic data go missing from the contemporary COVID-19 literature is crucial to ensure the comprehensiveness and rigor of this literature. This systematic review and meta-analysis aim to quantify this burden by estimating its pooled prevalence. Methods Registration This systematic review is pre-registered in the PROSPERO (CRD42020197319) [11] and reported here according to The Preferred Reporting Items for Systematic Review and Meta- Analysis (PRISMA)2020 statement (Supplementary Table S1) [12]. A pre-published protocol does not exist. Inclusion criteria We included studies that fulfilled the following inclusion criteria: 1. Study population: Hospitalized COVID-19 infected diabetes patients of any age or gender. 2. Study design: Observational studies, including case series conducted in any country. 3. Outcome: The outcome of interest is the number of patients whose post-hospitalization treatment outcome (i.e., discharge from hospital or death) was not reported in the published manuscript. Exclusion criteria 1. Studies conducted on pregnant patients. 2. Experimental study designs, case reports, letters, and editorials. 3. Studies that were reporting of treatment outcomes of its entire sample population. Data Source: We searched the title and abstract of eligible citations published in the English language in three electronic databases (PubMed, Embase, and Scopus) irrespective of the publication date or geographical boundary. Subsequent search terms were used to search the PubMed database: "diabetes mellitus, type 2"(MeSH Major Topic) OR "diabetes mellitus, type 1"(MeSH Major Topic) OR "diabetes mellitus"(MeSH Major Topic) AND "coronavirus infections"(MeSH Major Topic) AND diabetes AND SARS-CoV-2 OR Coronavirus OR COVID-19 NOT "Middle East respiratory syndrome" NOT MERS. Table S2 provides the detailed search strategy used to search different databases. Additional searches ensued in the bibliography of the articles included in this review and the 'Google' search engine. Study selection and data abstraction After uploading the retrieved citations from the database search and additional searches to a reference management software, the review authors independently skimmed through it to identify dubious and seemingly eligible articles for full-text reading and subsequently finalized the list of articles to be reviewed. Data abstraction from the studies included in this review happened for the following components - the nation and continent of the conduct of the study, follow-up duration of the study, the total number of inpatient COVID-19 infected diabetes patients, the total number of these patients whose prognosis data did not get reported in the article, type of diabetes detected in the study population, diagnostic guideline or criteria used to diagnose diabetes, diagnostic techniques used to ascertain COVID-19 infection, the average age of the study population, and the study design. These details are presented in. a tabular form. Pre-piloted data abstraction sheets were used to abstract the data. Risk of bias evaluation The reviewed studies' risk of bias assessment transpired via the National Heart, Lung, and Blood Institute's tool.[13] The 'yes' or 'no' categorization followed for each study's respective risk of bias components, based on if a study did or did not address this, respectively. If such judgment was not possible, 'cannot determine' or 'not applicable' labeling ensued based on whichever was the best applicable categorization. Review authors’ role The review authors conducted the study selection, data abstraction, and critical appraisal independently, and resolved any conflict in an opinion by discussion, and did not require a third-party consultation. Meta-analysis From published manuscripts, estimation of the pooled weighted prevalence of missing treatment outcome data of hospitalized COVID-19 infected diabetes patients ensued by random effect (DerSimonian and Laird) meta-analysis. The 95% confidence interval (CI) and variance stabilization transpired using the exact binomial method and Freeman-Tukey double arcsine transformation, respectively. Heterogeneity estimation happened by I2 statistics (at values 25, 50, and 75% heterogeneity were categorized as low, moderate, and high, respectively) [14] and p-value of Chi2 statistics (statistically significant at p<0.1). The meta-analysis findings are presented using a forest plot and table. Subgroup analysis The subgroup-wise weighted prevalence estimation of missing prognosis data transpired for continents, countries, diabetes types, and sample size (≤100 versus >100). Publication bias Small study effects got evaluated using visually and statistically by funnel plots and Egger’s test, respectively. Heterogeneity assessment A univariate meta-regression analysis (random-effect) ensued for each of the above-stated subgrouping variables to explain heterogeneity, and its statistical significance was determined at p<0.1. As none of these models produced a statistically significant outcome, we did not include these variables in an adjusted meta-regression model. Sensitivity analysis We repeated the overall pooled prevalence meta-analysis by dropping a study each time to see how each study contributed to the meta-analysis model. Stata statistical software (version 16) of StataCorp, College Station, Texas, USA, and MetaProp [15] package was used for the analysis. Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 575 Results Scope of this review The database search and additional searches retrieved altogether 996 citations, of which 779 records got skimmed following the elimination of duplicates. Out of the 54 articles requiring full- text reading, ten publications published in 2020 got included in this review (Figure 1) [16–25]. The primary reason for excluding papers read in full-text was non-reporting the outcome data of interest (61.0%; n=27). The last date of the search was 19-Nov-2020. Majority of the recruited studies were case series (70.0%) [16–22], followed by cross-sectional studies (20.0%) [23, 24] and retrospective cohort study (10.0%) [25]. The data of the recruited studies came from 6687 COVID-19 infected inpatient diabetes patients from four continents (Asia, Australia, Europe, and North America). Most of these patients belonged to the US (76.8%; n=5137). About 20.7% (n=1386) of the hospitalized COVID-19 patients with diabetes had either type 1 or 2 diabetes, another 1.3% (n=87) had type 2 diabetes, and for the remaining study participants, the exact diabetes type remains unknown. The salient features of the reviewed studies got presented in Table 1. Figure 1. The PRISMA 2020 flow diagram [12] Risk of bias evaluation Upon critical appraisal, the case series were of fair quality [16– 22], whereas the remaining study types were of good quality [23–25]. Table 2 depicts the study design-specific risk of bias assessment for the respective studies. Meta-analysis findings The overall pooled weighted prevalence of inpatient COVID-19 infected diabetes patients whose outcome data did not get reported in COVID-19 literature was 33.0% (95% CI: 15.0- 53.0%; I2: 99.53%; P of Chi2: <0.001) (Figure 2). Subgroup- wise, among the four continents, it was highest in studies conducted in Europe (63%; 95% CI: 61.0-66.0%). The latter was about three times higher than North America (24%; 95% CI: 3.0-55.0%; I2: 99.75%; P of Chi2 <0.001). The proportion of missing treatment outcome data reporting among hospitalized COVID-19 infected diabetes patients was marginally higher (5%) in studies with a larger sample size (i.e., n >100) (Table 2). Records identified from: Databases (n = 994) (PubMed: 411; Embase: 165; Scopus: 418) Records removed before the screening: Duplicate records removed (n = 215) Records screened (n = 779) Records excluded (n = 725) Reports sought for retrieval (n = 54) Reports not retrieved (n = 0) Reports assessed for eligibility. (n = 54) Reports excluded: (n=44) Reasons of exclusion 1. All participants' final outcomes reported (n=3) 2. Unclear outcome data (n=27) 3. Missing prognosis data not reported (n=8) 4. Participant data from the same hospital over an overlapping period (n=5) 5. Wrong study population (n=1) Records identified from: Websites (n = 2) Reports assessed for eligibility (n = 2) Reports excluded (n=0) Studies included in the review. (n = 10) Reports of included studies (n = 10) Identification of studies via databases and registers Identification of studies via other methods Id e n ti fi c a ti o n S c r e e n in g In c lu d e d Reports sought for retrieval (n = 2) Reports not retrieved (n = 0) Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 576 Publication bias and heterogeneity assessment On visual inspection, the funnel plots appeared somewhat symmetrical (Figure 3). The statistical evaluation of the small study effect did not suggest any publication bias (p = 0.617). The univariate meta-regression analysis was not statistically significant for any of the predictors (Table 3). Sensitivity analysis On iterating the meta-analysis while dropping a study each time, the prevalence varied between 29-37%. Discussion Altogether, this review included ten articles published in 2020 reporting of 6687 COVID-19 infected diabetes patients sourcing from Asia, Australia, Europe, and North America. Meta-analysis suggested a considerable underreporting of the treatment outcome data of hospitalized COVID-19 infected diabetes patients. This non-reporting was highest in Europe. Juxtaposing this review's findings with other review articles on COVID-19 was beyond the scope due to conceptual novelty and the non-availability of identical review articles. Implications While the number of COVID-19-diabetes-related publications soars at an unprecedented rate during the ongoing SARS-CoV-2 pandemic, it is vital to evaluate the completeness and rigor of this novel evidence. In this regard, the findings of this paper may serve as an identifier and reminder of the bulk of crucial prognosis data lost from the contemporary COVID-19-diabetes literature due to underreporting and may encourage researchers to take initiatives to ensure completeness of prognosis data reporting among COVID-19 infected hospitalized diabetes patients. It emphasizes the plausible constraints of COVID-19 research in the context, like limitations in funding or available time to ensure complete reporting of studies. Given the substantial burden of underreported prognostic data, policymakers may consider fetching regular updates from the researchers to calibrate the existing policies accordingly. Strengths and weaknesses The key strength of this study is its uniqueness in exploring an unexplored area of COVID-19-diabetes literature. Besides, this review is likely to be comprehensive as its literature search did not get restricted to any date range or geographic boundary. Despite these strengths, our systematic review has a few weaknesses. This review could not include potential studies published in the non-English language since the authors are not adept in any other language. Besides, our estimates are based on observational study designs, considered to be a weaker source of evidence than randomized clinical trials. Conclusion Globally, the under-reporting of hospitalized COVID-19 infected diabetes patients’ treatment outcomes is substantial. It increases the threat of biasing the expanding COVID-19 literature. The researchers may consider releasing such initially non-published prognostic data as adjunct reports in the post- publication period to decrease the risk of such bias. Journals might also take the initiatives to permanently identify such updated supplementary reports by providing digital object identifiers and electronically linking these to the parent publication. Alternatively, when feasible, the researchers may defer their manuscript drafting until the treatment outcomes of all admitted patients are known. Abbreviation CI: Confidence interval; COVID-19: coronavirus disease; PRISMA: The Preferred Reporting Items for Systematic Review and Meta-Analysis Declaration acknowledgment The authors would like to express gratitude to participants who helped in filling the google form. Funding The authors received no financial support for their research, authorship, and/or publication of this article. Availability of data and materials Data will be available by emailing sumanta.saha@uq.net.au on receiving a legitimate request Authors’ contributions Sumanta Saha designed and conceptualized this study analyzed and drafted this manuscript's first and final draft. Both authors participated in study selection, data abstraction, and critical appraisal. All authors have read and approved the final manuscript. Ethics approval and consent to participate We conducted the research following the Declaration of Helsinki. However, Review Articles need no ethics committee approval. Consent for publication Not applicable Competing interest The authors declare that they have no competing interest. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Author details 1Department of Community Medicine, R. G. Kar Medical College, Kolkata, India. 2Department of Mathematics, Mankar College, Mankar, India. Article Info Received: 27 August 2021 Accepted: 28 October 2021 Published: 23 November 2021 Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 577 Table 1. Salient findings of the reviewed studies Author, year Country Continent Dates Total diabetes admissions Missing prognosis data Diabetes type Diabetes diagnosis method COVID-19 diagnosis method Mean age of the study population* Study design Agarwal, 2020[23] US North America March 11 to May 07, 2020 1279 87 Unclear Clinical Modification code or HbA1c ≥6.5% RT-PCR Mean±SD:18 6 20 (n=1,279) Cross-sectional Cariou, 2020[16] France Europe March 10 to April 10, 2020 1317 877 Type 1 and 2 diabetes Personal history or HbA1c ≥6.5% RT-PCR Mean±SD:169.8 ± 13.0 (n=1,317) Case series Ciceri, 2020[17] Italy Europe February 25 to March 24, 2020 69 5 Type 1 and 2 diabetes Unclear RT-PCR Median (IQR): 65 (56– 75) (n= 410) Case series Croft, 2020[18] US North America Unclear 5 1 Type 2 diabetes Unclear RT-PCR Mean: 49 years; (n=5) Case series Liu, 2020[24] China Asia January 16, 2020, to March 16, 2020 19 13 Unclear Guidelines for the Prevention and Treatment of Type 2 Diabetes in China (2017 edition) seventh Trial Version of the Novel Coronavirus Pneumonia Diagnosis and Treatment Guidance DM patients (mean ±SD): non-critical (61.57 ± 12.01), critical (59.36 ± 12.31) Cross-sectional Marcello, 2020[19] US North America March 05 to April 16 2045 420 Unclear Unclear RT-PCR Median (IQR): 50.2 (36.6-61.9); (n=22176) Case series Richardson, 2020[20] US North America March 01 to April 04 1808 1051 Unclear Unclear RT-PCR Median (IQR): 63 (52- 75) (n=5700) Case series Wu, 2020[21] Australia Australia March 20 and May 01, 2020 8 2 Type 2 diabetes Unclear Unclear Mean±SD: 55±11.9 years (n=8) Case series Zhang, 2020a[22] China Asia January 03 to April 14, 2020 74 10 Type 2 diabetes Unclear Chinese National Health Committee (version 5). Median (IQR): 62(56– 72) (n=74) Case series Zhang, 2020b[25] China Asia January 29 to February 12 63 40 Unclear medical history and guidelines for the prevention and control of T2DM in China World Health Organization interim guidance Median (IQR): 65 (57– 71) (n of diabetes patients=63) Retrospective cohort study *n is the total sample size for which demographic data are presented in the respective studies Abbreviations: IQR: interquartile range; RT-PCR: Reverse transcription-polymerase chain reaction; SD: standard deviation Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 578 Table 2. Overall and subgroup weighted prevalence of missing prognosis data among inpatient COVID-19 patients with diabetes Subgroup Category Number of Studies Number of admitted COVID- 19 patients with diabetes Number of admitted COVID-19 patients with diabetes with missing prognosis data Mean prevalence of missing prognosis data in COVID-19 infected patients with diabetes Heterogeneity measures % 95% CI I2 (%) Chi2 (p-value) Continent Asia 3 156 63 46 11.0-84.0 - - Australia 1 8 2 25 3.0-65.0 - - Europe 2 1386 882 63 61.0-66.0 - - North America 4 5137 1559 24 3.0-55.0 99.75 <0.001 Country Australia 1 8 2 25 3.0-65.0 - - China 3 156 63 46 11.0, 84.0 - - France 1 1317 877 67 64.0-69.0 - - Italy 1 69 5 7 2.0-16.0 - - US 4 5137 1559 24 3.0-55.0 99.75 <0.001 Diabetes type Both type 1 and 2 2 1386 882 63 61.0-66.0 - - Type 2 3 87 13 12 5.0-21.0 - - Unclear 5 5214 1611 40 16.0, 67.0 99.68 <0.001 Sample size ≤100 6 238 71 31 8.0-59.0 93.43 <0.001 >100 4 6449 2435 36 11.0-66.0 99.84 <0.001 Overall 10 6687 2506 33 15.0-53.0 99.53 <0.001 Abbreviation: CI: confidence interval Table 3. Univariate meta-regression analysis for the prevalence studies on missing prognosis data of COVID-19 patients with diabetes. Subgroup Category Univariate model OR P-value 95% CI Continent North America 1 Asia 2.95 0.389 0.17, 50.72 Australia 1.17 0.929 0.02, 75.60 Europe 1.386 0.812 0.06, 34.95 Country US 1 Australia 1.17 0.919 0.03, 51.58 China 2.95 0.332 0.22, 39.06 France 7.00 0.243 0.16, 307.79 Italy 0.27 0.420 0.01, 12.06 Diabetes type Unclear 1 Both type 1 and 2 0.63 0.707 0.04, 10.47 Type 2 0.37 0.376 0.03, 4.37 Sample size ≤100 1 >100 1.21 0.841 0.15, 9.86 Abbreviations: CI: confidence interval; OR: odds ratio Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 579 Figure 2. Forest plot depicting the overall pooled prevalence of missing prognosis data of COVID-19 infected diabetes patients; Zhang, 2020a[22], Zhang, 2020b[25] The diamonds are centered on the summary of the overall and subgroup-wise prevalence estimates, and their widths indicate the corresponding 95% CI. Figure 3. Funnel plot assessing small study effect on pooled prevalence among hospitalized COVID-19 patients with diabetes. Saha S, Saha S, Journal of Ideas in Health (2021); 4(4):573-580 580 References 1. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet 2020; 395: 497–506. 2. Zou L, Ruan F, Huang M, et al. SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected Patients. N Engl J Med 2020; 382: 1177–1179. 3. Liu Y, Yan L-M, Wan L, et al. Viral dynamics in mild and severe cases of COVID-19. Lancet Infect Dis 2020; 20: 656–657. 4. World Health Organization. WHO Coronavirus (COVID-19) Dashboard, https://covid19.who.int/ (2021). 5. Saha S, Al-Rifai RH, Saha S. Diabetes prevalence and mortality in COVID-19 patients: a systematic review, meta‐analysis, and meta‐ regression. J Diabetes Metab Disord. Epub ahead of print March 31 2021. DOI: 10.1007/s40200-021-00779-2. 6. Yang JK, Feng Y, Yuan MY, et al. Plasma glucose levels and diabetes are independent predictors for mortality and morbidity in patients with SARS. Diabet Med 2006; 23: 623–628. 7. Alqahtani FY, Aleanizy FS, Ali El Hadi Mohamed R, et al. prevalence of comorbidities in cases of Middle East respiratory syndrome coronavirus: a retrospective study. Epidemiol Infect 2019; 147: e35. 8. Letko M, Marzi A, Munster V. Functional assessment of cell entry and receptor usage for SARS-CoV-2 and other lineage B betacoronaviruses. Nat Microbiol 2020; 5: 562–569. 9. Song Z, Xu Y, Bao L, et al. From SARS to MERS, Thrusting Coronaviruses into the Spotlight. Viruses; 11. Epub ahead of print 2019. DOI: 10.3390/v11010059. 10. Yang J-K, Lin S-S, Ji X-J, et al. Binding of SARS coronavirus to its receptor damages islets and causes acute diabetes. Acta Diabetol 2010; 47: 193–199. 11. Saha S, Saha S. A systematic review and meta-analysis of observational studies and case series to determine the prevalence of missing prognosis data in hospitalized COVID-19 patients with diabetes mellitus. PROSPERO 2020 CRD42020197319, https://www.crd.york.ac.uk/prospero/display_record.php?ID=CR D42020197319 (accessed 27 November 2020). 12. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021; n71. 13. Study Quality Assessment Tools | National Heart, Lung, and Blood Institute (NHLBI), https://www.nhlbi.nih.gov/health- topics/study-quality-assessment-tools (accessed June 14, 2020). 14. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–60. 15. Nyaga VN, Arbyn M, Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Arch Public Heal 2014; 72: 39. 16. Cariou B, Hadjadj S, Wargny M, et al. Phenotypic characteristics and prognosis of inpatients with COVID-19 and diabetes: the CORONADO study. Diabetologia 2020; 63: 1500–1515. 17. Ciceri F, Castagna A, Rovere-Querini P, et al. Early predictors of clinical outcomes of COVID-19 outbreak in Milan, Italy. Clin Immunol 2020; 217: 108509. 18. Croft A, Bucca A, Jansen JH, et al. First-time Diabetic Ketoacidosis in Type 2 Diabetics With Covid-19 Infection: A Novel Case Series. J Emerg Med. Epub ahead of print July 2020. DOI: 10.1016/j.jemermed.2020.07.017. 19. Marcello RK, Dolle J, Grami S, et al. Characteristics and Outcomes of COVID-19 Patients in New York City’s Public Hospital System. medRxiv 2020; 2020.05.29.20086645. 20. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA 2020; 323: 2052. 21. Wu L, Girgis CM, Cheung NW. COVID‐19 and diabetes: Insulin requirements parallel illness severity in critically unwell patients. Clin Endocrinol (Oxf) 2020; 93: 390–393. 22. Zhang Q, Wei Y, Chen M, et al. Clinical analysis of risk factors for severe COVID-19 patients with type 2 diabetes. J Diabetes Complications 2020; 34: 107666. 23. Agarwal S, Schechter C, Southern W, et al. Preadmission diabetes-specific risk factors for mortality in hospitalized patients with diabetes and coronavirus disease 2019. Diabetes Care 2020; 43: 2339–2344. 24. Liu D, Lan L, Luo D, et al. Lymphocyte subsets with the lowest decline at baseline and the slow lowest rise during recovery in COVID-19 critical illness patients with diabetes mellitus. Diabetes Res Clin Pract 2020; 167: 108341. 25. Zhang Y, Cui Y, Shen M, et al. Association of diabetes mellitus with disease severity and prognosis in COVID-19: A retrospective cohort study. Diabetes Res Clin Pract 2020; 165: 108227.