B-Type Natriuretic Peptide Diagnostic Performance Varies by Ethnicity in UK African-Caribbean and South Asian Populations: An Exploratory Analysis from the E-ECHOES Study
Authors:
JeeteshVPatel1
GregoryYHLip2
EduardShantsila2
ElizabethAHughes3
ParamjitSGill1,4✉Email
1
A
A
Warwick Medical SchoolUniversity of Warwick CoventryUK
2Liverpool Centre for Cardiovascular ScienceUniversity of Liverpool, Liverpool John Moores University and Liverpool Heart & Chest HospitalLiverpoolUK
3Sandwell and West Birmingham Hospitals NHS TrustUK
4Warwick Medical SchoolUniversity of WarwickCV4 7ALCoventryUK
Jeetesh V Patel 1, Gregory YH Lip 2 Eduard Shantsila2, Elizabeth A Hughes3, Paramjit S Gill1
Affiliations:
1Warwick Medical School, University of Warwick Coventry, (UK)
2Liverpool Centre for Cardiovascular Science, University of Liverpool, Liverpool John Moores University and Liverpool Heart & Chest Hospital, Liverpool, (UK)
3 Sandwell and West Birmingham Hospitals NHS Trust, (UK)
Paramjit S Gill, Warwick Medical School, University of Warwick, Coventry CV4 7AL, (UK)
p.gill.1@warwick.ac.uk
Corresponding Author:
Abstract
Background
Heart failure (HF) disproportionately affects ethnic minority populations in the UK, with Black African-Caribbean (AC) and South Asian (SA) individuals at increased risk of developing HF at younger ages. Current European guidelines recommend B-type natriuretic peptide (BNP) thresholds < 35 pg/mL for excluding HF in symptomatic patients, but diagnostic performance across ethnic groups remains unclear. This study aimed to evaluate the discriminative performance of BNP for identifying heart failure with reduced ejection fraction (HFrEF) and mildly reduced ejection fraction (HFmrEF) in AC and SA populations residing in the UK.
Methods
A
Cross-sectional analysis of the Ethnic-Echocardiographic Heart of England Screening study (E-ECHOES).
A
We assessed 1,164 participants (457 AC, 707 SA) aged ≥ 45 years. Participants underwent echocardiographic assessment, electrocardiography, and BNP measurement. Abnormal left ventricular ejection fraction (LVEF) was defined as < 50%. Receiver operating characteristic (ROC) analysis determined optimal BNP cutoffs and diagnostic performance. Logistic regression assessed the incremental value of combining BNP with clinical variables.
Results
Eighteen participants (1.5%) had abnormal LVEF. Median BNP levels were similar between ethnic groups (AC: 11.5 pg/mL; SA: 13.6 pg/mL). For the combined cohort, BNP demonstrated moderate discriminatory ability (AUC 77.2%, 95% CI 64.6–89.8) with an optimal cutoff of 35.9 pg/mL (sensitivity 66.7%, specificity 83.6.%). In SA participants alone, BNP performance improved (AUC 86.5%, 95% CI 76.9–96.0) with a lower optimal cutoff of 28.4 pg/mL (sensitivity 81.8%, specificity 75.3%). ROC analysis in AC participants was limited by low event numbers. Combining BNP with diabetes status and ECG findings enhanced diagnostic performance across both groups (AUC 90.5%, 95% CI 84.4–96.7)3012.
Conclusions
BNP maintains a high negative predictive value for ruling out left ventricular dysfunction across ethnic groups (98.9–99.3%).
A
These exploratory findings suggest that SA participants demonstrate optimal diagnostic performance at approximately 20% lower BNP threshold than current European guideline recommendations. Integration of BNP with diabetes status and ECG assessment substantially improves diagnostic accuracy in both groups. These hypothesis generating findings warrant validation in larger studies, particularly in AC populations to establish equitable diagnostic approaches for heart failure detection.
Keywords
Heart failure
B-type natriuretic peptide
ethnic minorities
diagnostic performance
health equity
South Asian
African-Caribbean
A
A
Background
Heart Failure (HF) is a complex clinical syndrome with increasing prevalence, posing major challenges for health services worldwide [1]. Early detection is an immense challenge - in the UK, approximately 920,000 people live with diagnosed HF, and an estimated 200,000 additional cases remain undiagnosed [2]. HF disproportionately affects ethnic minorities with Black African-Caribbean (AC) and South Asians (SA) facing an increased risk of developing HF at younger ages and from different causes compared to the White UK population [3–7].
AC populations develop HF approximately 10 years earlier than White populations, where ‘more’ severe hypertension is a primary cause [4, 5]. SA populations have a 2–4 fold higher susceptibility to diabetes and coronary heart disease, both occurring at younger ages and contributing to an elevated HF risk [6, 7]. Compared to other ethnic groups, SA with HF have distinct patterns of left ventricular remodelling and diastolic dysfunction [8].
B-type natriuretic peptide (BNP) has emerged as a key biomarker in HF diagnosis, and circulating levels reveal the real-time cardio-endocrine responses to myocardial stretch, remodelling and hypoxia [9–11]. Current European Society of Cardiology guidelines recommend BNP measurement (with a threshold < 35 pg/mL) to help exclude HF in symptomatic patients presenting in primary care and emergency settings [12]. However, the diagnostic performance of BNP across different ethnic populations remains poorly characterised, as ethnic minority groups have been typically underrepresented in cardiovascular research [13].
Substantial variations in natriuretic peptide levels are seen within multi-ethnic populations, with lower BNP concentrations in Black populations relative to White populations even after adjustment for clinical variables [14, 15]. However, data specifically looking at the diagnostic performance of BNP in ethnic minority populations is limited. Given the known ethnic differences in HF aetiology, phenotype, and risk factor profiles, it is plausible that universally applied BNP thresholds may not perform equally across all ethnic groups, potentially contributing to delayed diagnosis and health inequities.
The Ethnic-Echocardiographic Heart of England Screening study (E-ECHOES) systematically assessed cardiovascular health in a large, community-based sample of AC and SA living in Birmingham, UK [16]. The primary objective of this analysis was to evaluate the discriminative performance of BNP for identifying heart failure with reduced ejection fraction (HFrEF) and mildly reduced ejection fraction (HFmrEF) in these populations, and to determine whether ethnicity-specific BNP thresholds may improve diagnostic accuracy.
Methods
Study Design and Population
We conducted a cross-sectional sub-study of the Ethnic-Echocardiographic Heart of England Screening study (E-ECHOES), which documented community HF prevalence amongst SA and AC groups living in Birmingham, UK [16].
A
Briefly, participants were recruited from 20 general practices in Birmingham between 2006 and 2009. Eligible individuals were aged ≥ 45 years and self-identified as being of AC or SA ethnicity. This sub-study included E-ECHOES participants with valid left ventricular ejection fraction (LVEF) assessment and BNP measurement.
Trained research teams collected data using a standardised protocols including health questionnaires, clinical examination (anthropometry and blood pressure), and 12-lead electrocardiography (classified abnormal in the presence of ventricular hypertrophy, pathological Q waves, artrial fibrillation, bundle branch block or ST segment or T wave abnormalities.
Echocardiography
Participants underwent two-dimensional and Doppler trans-thoracic echocardiography (VIVID-i, GE Healthcare, Chalfont St Giles, UK). Left ventricular function was measured objectively using the area-length method from the apical four-chamber view [17]. Abnormal LVEF was defined as < 50%, encompassing both heart failure with reduced ejection fraction (HFrEF, LVEF < 40%) and heart failure with mildly reduced ejection fraction (HFmrEF, LVEF 40–49%), consistent with ESC classification [12].
Blood samples were processed within 8 hours of collection, with serum and plasma stored at -70°C for batch analysis including total cholesterol, HDL cholesterol and creatinine (Cobas Integra 400, Roche Diagnostics, UK). Chronic kidney disease was determined based on estimated glomerular filtration rate < 60 ml/min/1.73m (calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [18]).
BNP measurement
BNP was measured using commercial automated immunoassay (ADVIA Centaur, Bayer Healthcare, UK) with coefficient of variation < 5%. BNP levels expressed (pmol/l) were converted to pg/ml by multiplying by 3.47.
Statistical Analysis
Sample sizes for receiver operating characteristic (ROC) curves were estimated from published tables [19]. At 5% type I error rate, and 80% power, 229 patients were sufficient to detect 10% differences in area under the curve (AUC) for BNP discriminating left ventricular systolic dysfunction (LVSD).
Data were analysed using SPSS version 29.0.20. Parametric distribution was tested using Kolmogorov–Smirnov plots. Normally distributed data are presented as mean ± standard deviation and compared using independent t-tests. Non-normally distributed data are presented as median (interquartile range) and compared using Mann-Whitney U tests. Categorical variables are presented as frequencies (percentages) and compared using chi-square. BNP performance discriminating abnormal LVEF (HFrEF and HFmrEF) was analysed using ROC curves. Optimal cutoff points were determined using Youden's index. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for optimal cutoffs. Separate ROC analyses were performed for the combined cohort and for each ethnic group individually.
Logistic regression analysis was performed to identify independent predictors of abnormal LVEF and reported with Odd Ratio (OR). Variables with P < 0.10 in univariate analyses were included in the multivariable model. The incremental diagnostic value of combining BNP with clinical variables (diabetes status and ECG abnormalities) was assessed using ROC analysis for the combined model.
All tests were two tailed with P < 0.05 considered significant.
Results
Study Population
From a total of 1,959 subjects recruited for this E-ECHOES sub-study, 1,164 (59.4%) had a valid assessment of LVEF and BNP measurement, of whom 457 (39.3%) were AC and 707 were SA (60.7%). Of these, 1146 subjects had normal LVEF and 18 had abnormal LVEF and their demographics, co-morbidities, clinical and biochemical measures are shown in Table 1.
Table 1
Demographics, co-morbidities, clinical and biochemical measures of participants of the Ethnic-Echocardiographic Heart of England Screening study (E-ECHOES)
 
African Caribbean
South Asian
Left Ventricular Ejection Fraction > = 50% (n = 450)
Left Ventricular Ejection Fraction < 50% (n = 7)
Left Ventricular Ejection Fraction > = 50% (n = 696)
Left Ventricular Ejection Fraction < 50% (n = 11)
Age (yrs)
60.1
(11.7)
69.1
(9.7)
56.7
(9.3)
64.3
(12.2)
% female
55.1
(248)
14.3
(1)
44.0
(306)
27.3
(3)
% smokers
41.3
(186)
42.9
(3)
19.5
(136)
18.2
(2)
Body-mass index (kg/m2)
29.7
(5.9)
31.5
(4.7)
28.1
(4.6)
26.9
(4.1)
% diabetes
22.9
(103)
71.4
(5)
25.0
(174)
54.5
(6)
% coronary artery disease
3.1
(14)
0.0
(0)
8.8
(61)
54.5
(6)
% arrythmia
2.9
(13)
0.0
(0)
1.4
(10)
9.1
(1)
% hypertension
52.7
(237)
100.0
(7)
39.4
(274)
72.7
(8)
Systolic BP (mmHg)
114
(19)
162
(13)
140
(19)
142
(31)
Diastolic BP (mmHg)
82.8
(10.6)
88.1
(8.5)
82.6
(10.8)
77.2
(12.0)
% ACE inhibitor/Angiotensin II receptor inhibitor therapy
24.0
(108)
85.7
(6)
20.1
(140)
63.6
(7)
% Beta-blocker therapy
10.9
(49)
42.9
(3)
12.2
(85)
36.4
(4)
% Diuretic therapy
28.9
(130)
71.4
(5)
16.1
(112)
18.2
(2)
% normal ECG
70.4
(317)
14.3
(1)
81.5
(567)
9.1
(1)
% chronic kidney disease
4.4
(20)
0.0
(0)
2.9
(20)
9.1
(1)
% eGFR < 30
0.7
(3)
0.0
(0)
0.4
(3)
0.0
(0)
eGFR ml/min/1.73m
112
(39)
97
(33)
113
(38)
108
(40)
% statin therapy
32.9
(148)
71.4
(5)
37.5
(261)
54.4
(6)
Total cholesterol (mmol/l)
4.48
(1.36)
3.55
(1.30)
4.45
(1.2)
4.14
(0.91)
HDL cholesterol (mmol/l)
1.30
(0.47)
1.16
(0.42)
1.07
(0.35)
1.07
(0.25)
Data are mean (SD) or Percent (n)
Six subjects with LVEF < 40% (AC = 2), and 12 subjects with LVEF 41–49% (AC = 5) were identified. Among participants with abnormal LVEF, 11 were male (61.1%) and 7 were female (38.9%). The gender distribution was similar between ethnic groups (AC: 4 male, 3 female; SA: 7 male, 4 female).
Participants with abnormal LVEF were more likely to have diabetes (66.7% vs 42.8%, P = 0.04), abnormal ECG findings (83.3% vs 28.7%, P < 0.001), and hypertension (72.2% vs 62.3%, P = 0.38) compared to those with normal LVEF. There were no significant differences in age, body mass index, or renal function between groups.
Ethnic differences in BNP levels
Median (IQR) BNP in those with normal LVEF was 11.5 (6.3–19.9) pg/ml for AC and 13.6 (8.1–22.2) pg/ml in SA. There were no significant ethnic differences in BNP levels for those with and without normal LVEF.
ROC analysis: Combined Cohort
On ROC analysis of the combined group, BNP levels significantly discriminated the presence of abnormal LVEF with an AUC of 77.2%, (95% CI 64.6–89), P < 0.001, and a concentration of 35.9pg/ml maximised sensitivity (66.7%) and specificity (83.6%), with a positive predictive value of 21.4% and negative predictive value of 98.9%.
ROC analysis: Ethnic Specific Cohorts
In AC, the ROC model quality was poor, limited by the small numbers of abnormal LVEF (n = 7). The AUC was 71.4% (95% CI 48.9–93.9), and the wide confidence intervals preclude reliable determination of BNP performance for this population.
In SA, BNP levels significantly discriminated the presence of abnormal LVEF with an AUC of 86.5% (95%CI 76.9–96.0), P < 0.001. The optimal cutoff was a concentration of 28.4 pg/ml, which maximised sensitivity (81.8%) and specificity (75.3%), with a positive predictive value of 18.2% and negative predictive value of 99.3%.
Multivariate Analysis
In multivariate logistic regression analysis (Table 2), abnormal ECG findings (OR 17.9, 95% CI 3.9–81.2, P < 0.001), presence of diabetes (OR 1.17, 95% CI 1.15–9.09, P = 0.03), and BNP level (OR 1.01, 95% CI 1.00-1.01, P < 0.001) were independently associated with abnormal LVEF after adjusting for other variables
Table 2
Logistic regression of variables independently associated with abnormal left ventricular ejection fraction.
 
Exp Beta
95% CI for Exp Beta
Sig
Abnormal ECG
17.9
3.9–81.2
< 0.001
Diabetes
1.17
1.15–9.09
0.026
BNP (mg/dl)
1.01
1.00-1.01
< 0.001
Combined Diagnostic Model
ROC analysis using a combined model incorporating BNP levels, diabetes status, and ECG status demonstrated improved diagnostic performance compared to BNP alone. The combined model achieved an AUC of 90.5% (95% CI 84.4–96.7, P < 0.001).
When tested separately in each ethnic group, the combined model maintained performance in both AC (AUC 87.6%, 95% CI 76.9–98.3, P < 0.001) and SA participants (AUC 91.9%, 95% CI 84.3–99.6, P < 0.001).
Discussion
These data provide exploratory evidence regarding the diagnostic performance of BNP for identifying left ventricular systolic dysfunction in a UK community cohort of AC and SA. BNP demonstrated moderate discriminatory ability (AUC 77.2%, sensitivity 66.7%, specificity 83.6.%) across the combined cohort.
A
In SA alone, BNP performed better (AUC 86.5%), with an optimal diagnostic threshold approximately 20% lower (28.4 pg/mL) than current European guideline recommendations (35 pg/mL) - universal thresholds may lack sensitivity in this population. Combining BNP levels with diabetes status and ECG abnormalities substantially improved diagnostic accuracy (AUC 90.5%) performance across both ethnic groups.
Epidemiological studies have reported lower BNP/NT-proBNP concentrations in Black populations vs. White populations [14, 15], though the focus has been on absolute levels rather than diagnostic performance.
A
This study looked at the diagnostic accuracy in UK minority populations and suggests potential implications for guideline thresholds. High negative predictive values (98.9–99.3%) support the role of BNP to rule out LVSD across ethnic groups, while low positive predictive values (18.2–21.4%) indicate the need for confirmatory testing when elevated. The improved performance when combining BNP measures with diabetes and ECG assessment aligns with the American Diabetes Association recommendations for natriuretic peptide screening in diabetic patients [20]. This is particularly relevant given the 2–4 fold higher diabetes rates in AC and SA groups [21, 22]. Comparable median BNP levels between AC and SA suggest similar synthesis and clearance mechanisms. However, differences in the phenotypical presentation, comorbidities and HF aetiology may explain ethnic-specific differences in diagnostic performance here.
Limitations
BNP performance assessment in AC was limited by low abnormal LVEF rates; adequately powered studies are needed in this population. Sample size calculations were based for the overall diagnostic accuracy, rather than subgroup analyses, making these findings exploratory and hypothesis generating rather than definitive.
Conclusion
This exploratory analysis of the E-ECHOES study demonstrates that BNP maintains consistently high negative predictive value for ruling out LVSD across UK AC and SA populations.
A
However, SA participants demonstrated optimal diagnostic performance at approximately 20% lower BNP thresholds than current European guideline recommendations, suggesting that universal diagnostic thresholds may lack sensitivity for this population. These data support the integration of BNP measurement alongside systematic diabetes and ECG assessment to improve diagnostic accuracy in UK minority populations. These hypothesis-generating findings highlight important questions about whether ethnicity-informed diagnostic approaches could improve equity in heart failure detection.
Declarations
Ethics approval
A
The E-ECHOES study received ethical approval from the South Birmingham Research Ethics Committee (Reference: 06/Q2707/11).
A
All participants provided written informed consent prior to study participation.
A
All methods were performed in accordance with the relevant guidelines and regulations, and the study adhered to the Declaration of Helsinki.
A
Data Availability
The dataset analysed during the current study is not publicly available due to ethical restrictions and regulations, but are available from the corresponding author on reasonable request and subject to appropriate data sharing agreements.
Competing interests
Jeetesh V Patel : none declared
Gregory YH Lip : none declared
Eduard Shantsila: none declared
Elizabeth A Hughes : none declared
Paramjit S Gill : none declared
A
Funding
This study is funded by the British Heart Foundation (PG/05/036), Heart of Birmingham Teaching Primary Care Trust, and through the National Health Service R&D support funding (Primary Care Research Network-Central England). Primary Care Clinical Sciences is a member of the NIHR National School for Primary Care Research. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
A
Author Contribution
PSG conceived and designed the E-ECHOES study. JVP, GYHL and PSG designed this sub-study analysis. JVP performed the statistical analyses. All authors coordinated data collection and management. ES and GYHL contributed to study design and interpretation of echocardiographic data. EAH co-ordinated BNP measurement and sample storage. All authors critically reviewed and revised the manuscript. All authors read and approved the final manuscript.
A
Acknowledgement
We are grateful for the contribution from Julia Chackathayil for biochemical analysis. We are grateful to all the subjects, and the following practices for taking part in this study:Rotton Park Medical Centre, City Road Medical Practice, Cavendish Medical Practice, Ann Jones Family Health Centre, Shanklin House Surgery, Burbury Street Surgery, Heathford Group Practice, Broadway Health Centre, Victoria Road Medical Centre, Churchill Medical Centre, St Clements Surgery, Handsworth Medical Centre, Soho Health Centre, Church Road Surgery, Bloomsbury Health Centre, Al-Shafa Medical Practice, Enki Medical Practice, Aston Pride Health Centre, Newtown Health Centre, Hockley Medical Centre.
References
1.
Khan MS, Shahid I, Bennis A, Rakisheva A, Metra M, Butler J. Global epidemiology of heart failure. Nat Rev Cardiol. 2024;21:717–34.
2.
British Heart Foundation. Heart & Circulatory Disease Statistics 2025. London: British Heart Foundation. 2025. Available from: https://www.bhf.org.uk/what-we-do/our-research/heart-statistics
3.
Sosin MD, Bhatia GS, Davis RC, Connolly DL, Lip GY. Heart failure: treatment and ethnic origin. Lancet. 2003;362:919–20.
4.
Yancy CW. Heart failure in African Americans. Am J Cardiol. 2005;96:i3–12.
5.
Bibbins-Domingo K, Pletcher MJ, Lin F, Vittinghoff E, Gardin JM, Arynchyn A, et al. Racial differences in incident heart failure among young adults. N Engl J Med. 2009;360:1179–90.
6.
Almulhem M, Chandan JS, Gokhale K, Adderley NJ, Thayakaran R, Khunti K, et al. Cardio-metabolic outcomes in South Asians compared to White Europeans in the United Kingdom: a matched controlled population-based cohort study. BMC Cardiovasc Disord. 2021;21:320.
7.
Srinivasan N, Gullapalli N, Shah KS. Highlighting the South Asian Heart Failure Epidemic. Card Fail Rev. 2024;10:e07.
8.
Shantsila A, Shantsila E, Gill PS, Lip GYH. Predictors of diastolic dysfunction in ethnic groups: observations from the Hypertensive Cohort of The Ethnic-Echocardiographic Heart of England Screening Study (E-ECHOES). J Hum Hypertens. 2018;32:477–86.
9.
Bayés-Genis A. Population screening with natriuretic peptides is ready for prime time. Circulation. 2025;151:1547–9.
10.
Pascual-Figal DA, Antolinos MJ, Bayes-Genis A, Casas T, Nicolas F, Valdés M. B-type natriuretic peptide release in the coronary effluent after acute transient ischaemia in humans. Heart. 2007;93:1077–80.
11.
Fu S, Ping P, Zhu Q, Ye P, Luo L. Brain natriuretic peptide and its biochemical, analytical, and clinical issues in heart failure: a narrative review. Front Physiol. 2018;9:692.
12.
McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42:3599–726.
13.
Taylor CJ, Roalfe AK, Iles R, Hobbs FR, Barton P, Deeks J, et al. Primary care REFerral for EchocaRdiogram (REFER) in heart failure: a diagnostic accuracy study. Br J Gen Pract. 2017;67:e94–102.
14.
Gupta DK, de Lemos JA, Ayers CR, Berry JD, Wang TJ. Racial differences in natriuretic peptide levels: the Dallas Heart Study. JACC Heart Fail. 2015;3:513–9.
15.
Bajaj NS, Gutiérrez OM, Arora G, Judd SE, Patel N, Bennett A, et al. Racial differences in plasma levels of N-terminal pro-B-type natriuretic peptide and outcomes: the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. JAMA Cardiol. 2018;3:11–7.
16.
Gill PS, Davis R, Davis M, Freemantle N, Lip GYH. Rationale and study design of a cross sectional study documenting the prevalence of heart failure amongst the minority ethnic communities in the UK: the E-ECHOES Study (Ethnic Echocardiographic Heart of England Screening Study). BMC Cardiovasc Disord. 2009;9:47.
17.
Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2015;16:233–71.
18.
Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–12.
19.
Obuchowski NA. Sample size tables for receiver operating characteristic studies. AJR Am J Roentgenol. 2000;175:603–8.
20.
American Diabetes Association Professional Practice Committee. 10. Cardiovascular Disease and Risk Management: Standards of Care in Diabetes-2025. Diabetes Care. 2025;48:S207-38.
21.
Dreyer G, Hull S, Aitken Z, Chesser A, Yaqoob MM. The effect of ethnicity on the prevalence of diabetes and associated chronic kidney disease. QJM. 2009;102:261–9.
22.
Mathur R, Palla L, Farmer RE, Chaturvedi N, Smeeth L. Ethnic differences in the severity and clinical management of type 2 diabetes at time of diagnosis: A cohort study in the UK Clinical Practice Research Datalink. Diabetes Res Clin Pract. 2020;160:108006.
Yes
Total words in MS: 2430
Total words in Title: 22
Total words in Abstract: 330
Total Keyword count: 7
Total Images in MS: 0
Total Tables in MS: 2
Total Reference count: 22