Secondary Logo

Journal Logo

Original Articles

Relationship between a single measurement at baseline of body mass index, glycated hemoglobin, and the risk of mortality and cardiovascular morbidity in type 2 diabetes mellitus

Brown, Olivera,,*; Costanzo, Pierluigia,,*; Clark, Andrew L.a; Condorelli, Gianluigib; Cleland, John G.F.c; Sathyapalan, Thozhukatd; Hepburn, Davidd; Kilpatrick, Eric S.e; Atkin, Stephen L.f

Author Information
Cardiovascular Endocrinology & Metabolism: December 2020 - Volume 9 - Issue 4 - p 177-182
doi: 10.1097/XCE.0000000000000202



Type 2 diabetes mellitus (T2DM) and obesity are major causes of morbidity and mortality worldwide [1]. Obesity is a major risk factor for both diabetes and cardiovascular disease (CVD) [2,3].

The percentage of hemoglobin that is glycated in the blood (HbA1c) is routinely used in the diagnosis and monitoring of patients with diabetes. Large epidemiological studies in patients with T2DM suggest that having either a high or a low HbA1c is associated with increased all-cause mortality compared with HbA1c in the middle of the range [4,5]. A number of factors are known to affect HbA1c levels, mainly time since diagnosis of diabetes, cholesterol levels, and age [6,7].

The association between HbA1c and body weight is less clear and mainly affected by the interaction of different hypoglycemic treatment which can reduce HbA1c levels and cause either weight loss or gain [8]. Whether the relationship body weight and HbA1c affects clinical prognostic outcomes is unclear. To our knowledge, this association has never been fully explored, but only investigated in the context of interaction analysis between HbA1c and body mass index (BMI) to predict all-cause mortality [5,9]. In particular, van Munster et al. [9] demonstrated a significant interaction between HbA1c and BMI toward risk of mortality in a cohort of patients with T2DM. Similarly, Li et al. [5] in a subgroup analysis of their study demonstrated that higher HbA1c levels were associated with increased risk of mortality in obese patients with diabetes.

The aim of our study was to assess the relationship between a single measurement at baseline of glycated hemoglobin (HbA1c), BMI, and the risk of all-cause mortality, causal mortality, and hospitalization for cardiovascular outcomes in patients with T2DM.


Study population

Patients with a diagnosis of T2DM who attended the outpatient clinic service for diabetes in Kingston upon Hull, UK, were enrolled in a registry between 1995 and 2005. The secondary diabetes service in Kingston upon Hull provided most of the care that is usually provided by primary care service elsewhere, particularly during the years analysed for this cohort. The catchment area for the diabetic service covered about 230 000 people, of whom an estimated 6% have been diagnosed with T2DM [10].

Data were collected by medical and nursing staff and entered into a specifically designed electronic database [Angoss (Westman Medical Software, Manchester, UK)]. More than 99% of patients at the first visit had no known history of CVD (ischaemic heart disease, cerebrovascular, heart failure, or peripheral vascular disease). Data on age, time since diagnosis of diabetes, smoking history, height, weight, and blood pressure were collected at the initial visit. Information on comorbidity [cancer, chronic obstructive pulmonary disease (COPD), and chronic kidney disease (CKD)] was collected at baseline.

Patients were divided by BMI category as recommended by the WHO: underweight (<18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), and obese (>30 kg/m2) [11]. The cohort was followed for clinical events until December 2011.


The study protocol along with all other study documentation was approved by the research ethics committee (reference number 13/SW/0168).

Study outcomes

The primary outcome of the analysis was all-cause mortality. A national register informed the hospital of the death of any patient previously under the hospital’s care regardless of whether the patient had left the region. In addition to all-cause mortality, the cause of death was obtained from the Office of National Statistics. We divided deaths as being due to the following: cardiovascular; cancer; and sepsis (intended as a general term of infective illness). The secondary outcome was cardiovascular hospitalizations, divided into the following: acute coronary syndrome (ACS); heart failure; and cerebrovascular accident (CVA). Information on hospitalizations, coded using the International Classification of Diseases, 10th Revision, Clinical Modification, and mortality was collected through the Patient Information Service of the Hull and East Yorkshire Hospitals National Health Service Trust, the sole hospital provider of emergency medical services in the region.

Statistical analysis

Results are presented as medians and interquartile ranges (IQRs). Kruskal–Wallis tests for nonparametric data and chi-square tests were used to compare continuous and dichotomous data, respectively, between BMI categories.

Multivariable Cox regression models were constructed for all-cause mortality and cardiovascular events (ACS, CVA, or heart failure) using four groups of HbA1c [<6%, 6.0–6.9% (reference group), 7.0–7.9%, and >8%]. The models were adjusted for age, sex, time since diagnosis of diabetes, smoking history, systolic blood pressure, and comorbidities that could have affected the bodyweight like COPD, cancer, and CKD. If a patient had more than one admission for a given cause, only the first admission was included in the analysis. Each model was constructed for each BMI category. We excluded underweight patients (BMI < 18.5), as their sample size was small (only 44 patients). To check the consistency of the results we performed a sensitivity analysis by assessing the outcomes in reverse, by constructing each model for categories of HbA1c.

All analyses were presented as hazard ratios with 95% confidence intervals (95% CI). A two-tailed P value less than 0.05 was considered statistically significant. Adjustment for multiple tests was made using the Benjamini–Hochberg procedure.

All analyses were performed using STATA 11.0 (Stata Corp. College Station, Texas, USA) and SPSS, 23.0 (IBM Corp. Chicago, Illinois, USA).

Data and resource availability

The data supporting the findings of this study are available from our institution as aggregate data, as restrictions apply to the availability of confidentiality agreements. Under certain circumstances, data are available from the authors upon reasonable request and with permission of the National Health Service relevant institutions.


The cohort (Table 1) included 6220 patients (54% men) who had a median age of 62 (IQR 4–82) and were followed up for a median of 10.6 years (IQR 4.6–16.6). Only 44 patients (0.7%) were classed as underweight, 1354 (21.8%) patients were normal weight, 2316 patients (37.2%) were overweight, and 2505 (40.3%) patients were obese.

Table 1 - Baseline characteristics of study population
Variable Total population <18.5 18.5–24.9 25–29.9 >30 P value
Total (n) 6220 44 1354 2316 2505 <0.001
Age (years) 62.0 (20) 56.0 (37.0) 64.0 (28.0) 64.0 (17.0) 60.0 (17) <0.001
Men (%) 54.1 48.8 52.2 62.4 47.6 <0.001
Time since diagnosis of diabetes (years) 2.0 (7.0) 1.5 (5.3) 4.0 (11.0) 2.0 (7.0) 1.0 (4.0) <0.001
Follow-up duration (years) 10.6 (6.0) 8.3 (10.3) 11.6 (7.1) 11.1 (6.2) 10.8 (5.4) 0.009
Height (m) 1.67 (0.15) 1.64 (0.16) 1.67 (0.15) 1.69 (0.14) 1.65 (0.16) <0.001
Weight (kg) 81.0 (23.4) 48.0 (10.0) 64.2 (12.8) 78.0 (13.3) 95.1 (19.3) <0.001
BMI 28.7 (7.3) 17.6 (0.9) 23.2 (2.4) 27.6 (2.4) 34.0 (5.7) <0.001
SBP (mmHg) 142.0 (29.0) 126.0 (34.5) 136.0 (34.0) 142.2 (28.0) 141.0 (28.0) <0.001
Smoking (%) 15.3 36.3 18.5 14.3 16.3 <0.001
Cancer (%) 14.6 9.1 12.3 15.5 15.2 0.033
CKD (%) 8.2 6.8 6.5 9.9 7.5 0.001
COPD (%) 7.8 15.9 4.9 7.0 8.7 <0.001
Baseline characteristics of study population where not already stated results are shown as median (interquartile range).
CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; SBP, systolic blood pressure.

Of 2187 patients (35.2%) who died during follow-up, 946 (15.2%) died from cardiovascular causes, 423 (6.8%) from cancer, 527 (8.5%) were sepsis-related deaths, and 291 (4.7%) died from other causes (including pulmonary embolism and uncategorized causes). In addition, there were 1305 patients hospitalized for a cardiovascular event: 535 for ACS (8.6%), 356 for heart failure (5.7%), and 414 for CVA (6.7%).

Irrespective of BMI category, a baseline single measurement of HbA1c >8.0 was associated with increased risk of all-cause and cardiovascular mortality (hazard ratio 1.17; 95% CI, 1.04–1.32; P = 0.008) (Fig. 1). However, HbA1c was not shown to significantly predict cardiovascular hospitalization (Fig. 2).

Fig. 1
Fig. 1:
Adjusted Cox regression analysis of HbA1c and all-cause mortality, cardiovascular mortality, cancer mortality, and sepsis mortality. Model adjusted for age, sex, time since diagnosis of diabetes, smoking history, systolic blood pressure, and comorbidities that could have affected the bodyweight like COPD, cancer, and CKD. CI, confidence interval; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; N, number.
Fig. 2
Fig. 2:
Adjusted Cox regression analysis of HbA1c and hospitalization for ACS, CVA and heart failure. Model adjusted for age, sex, time since diagnosis of diabetes, smoking history, systolic blood pressure, and comorbidities that could have affected the bodyweight like COPD, cancer, and CKD. ACS, acute coronary syndrome; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; CVA, cerebrovascular accident; N, number.

For patients with a normal BMI, an increased baseline single measurement of HbA1c was not associated with all-cause, cause-specific mortality (Fig. 1) or hospitalization for ACS, heart failure, or stroke (Fig. 2).

For patients who were overweight, a baseline single measurement of HbA1c <6.0 was associated with an increase in all-cause mortality (hazard ratio, 1.35; 95% CI, 1.02–1.79; P = 0.03), which might have been driven by an increase in sepsis-related death (hazard ratio, 1.80; 95% CI, 1.04–3.14; P = 0.04). HbA1c >8.0 was associated with a borderline-significant increase in cardiovascular death (hazard ratio, 1.32; 95% CI, 1.00–1.77; P = 0.05). HbA1c was not associated with an increase in cancer-related deaths (Fig. 1). There was no significant association between HbA1c and rate of ACS, heart failure or stroke (Fig. 2).

Similarly to patients with normal weight, for obese patients, there was no significant association between HbA1c and any of the clinical outcomes (Figs 1 and 2).

The multiple-test procedure for type I error control gave an adjusted P < 0.001 to be possibly considered statistically significant. Therefore, we cannot exclude that in view of multiple testing the statistically significant results of these studies are the result of type I error.

Finally, as sensitivity analysis, we assessed the outcomes in reverse by constructing each model for categories of HbA1c. This did not show a substantial difference in the results (Supplemental Tables and Supplemental Figure 1, Supplemental digital content 1,


In a large cohort of patients with T2DM, enrolled consecutively at their first-attendance at a diabetes clinic, increasing HbA1c levels were not consistently associated with increased risk of all-cause mortality and cardiovascular death. A single baseline measurement of HbA1c >8.0% was associated with increased mortality only in overweight patients (BMI 25–29.9 kg/m2). In the other weight categories (normal and obese), a trend of increased all-cause and cardiovascular mortality was observed with a HbA1c >8.0%, although this did not reach statistical significance.

In our study, single baseline measurement of HbA1c did not significantly predict cardiovascular hospitalizations, although there was a trend of higher HbA1c levels and risk of CVA and heart failure hospitalizations.

It is well acknowledged that higher levels of HbA1c are associated with increased mortality in diabetic patients, as we have confirmed in our study [5,9,12]. Previous studies have investigated BMI as an interacting factor with HbA1c toward prognostic outcomes [5,9]. Compared to these studies, ours is the first study that systematically investigated the prognostic value of HbA1c according to different weight categories for different causes of death and adjusting for comorbidities that could have affected the BMI.

In view of these differences in study design, our results are not directly comparable with the available literature; however, some consistency in the results can be observed with Li et al. [5] In particular, in their subgroup analysis of a study in which they investigated the association of HbA1c with all-cause mortality in a T2DM population, they reported increased mortality with an HbA1c >9.0% in obese patients (BMI ≥ 30 kg/m2).

Our results are consistent with those from van Munster et al. Similarly, they investigated the relationship of HbA1c with cardiovascular events and mortality in a cohort of patients with T2DM. Similarly, they did not find a significant association between higher levels of HbA1c and cardiovascular events or all-cause mortality, although they reported interaction with BMI as continuous variable [9]. Compared to our study, van Munster et al. [9] had a smaller sample size (1753 patients) and did not have the possibility to include comorbidities in their statistical analysis.

Higher HbA1c is associated with an increased risk of macrovascular and microvascular complications in patients with diabetes [13–15], and so a relation between HbA1c and mortality might be expected. We found such an association in the overweight subset of patients (BMI 25–29.9 kg/m2). Whereas, only a trend towards increased risk of all-cause and cardiovascular mortality with HbA1c level >8% was observed in normal and obese patients. It could possible that with a bigger sample size statistical significance could have been reached. Interestingly, intermediate levels of glycemic control like HbA1c 7–7.9% were not associated with increased risk of cardiovascular hospitalizations, cardiovascular, or all-cause mortality in any of the BMI categories.

A recent systematic review reported that raised HbA1c is associated with an increase in cancer mortality [16], in particular pancreatic, colorectal, respiratory, and gynecological cancers. In our study, HbA1c levels were not associated with cancer-related mortality (Fig. 1), in the overall population and across the weight categories.

In our study, the association between HbA1c <6.0 and increased risk of death amongst patients who were overweight was possibly driven by sepsis. In a study of patients with T2DM and sepsis, higher HbA1c at admission was associated with increased hospital mortality, but this was a very different population from ours. In our study, the link between low HbA1c and increased mortality might reflect frailty, which increases the risk of sepsis and impairs recovery from illness [17].


There is a risk of type II error as some of the subgroups were small. Some true associations between HbA1c and clinical outcome may thus have been missed due to low statistical power.

We included patients only recruited before 2006 and the management of diabetes has subsequently changed substantially, which might have an impact on outcome. Particularly in the UK, in 1996 sulphonylureas were the main treatment. However, by 2005 the use of this drug class had decreased substantially, and there had been an increase of treatment with metformin and glitazone [18].

Patients will have had different time since diagnosis of T2DM prior to referral, which may have affected both HbA1c and weight. We did not have information on medications, cholesterol levels, education, or ethnicity, although our population is almost exclusively of European descent. We do not have information on hospitalizations outside the Hull and East Yorkshire Hospital Trust, but the rate of emigration from the area amongst adults is low.

We used only measurements made at the time of referral. Treatment and the evolution of disease will have altered risk factors, weight, and HbA1c, which may have disrupted the relationship with initial measurements.


P.C. and O.B. are the guarantors of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Conflicts of interest

There are no conflicts of interest.


1. Centers for Disease Control and Prevention. National Diabetes Fact Sheet, 2011: National Estimates and General Information on Diabetes and Prediabetes in the United States. 2011, Atlanta, GA: Centers for Disease Control and Prevention;. Accessed at www.cdc.
2. Adams KF, Schatzkin A, Harris TB, Kipnis V, Mouw T, Ballard-Barbash R, et al. Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med. 2006; 355:763–778
3. Bogers RP, Bemelmans WJ, Hoogenveen RT, Boshuizen HC, Woodward M, Knekt P, et al.; BMI-CHD Collaboration Investigators. Association of overweight with increased risk of coronary heart disease partly independent of blood pressure and cholesterol levels: a meta-analysis of 21 cohort studies including more than 300 000 persons. Arch Intern Med. 2007; 167:1720–1728
4. Nicholas J, Charlton J, Dregan A, Gulliford MC. Recent HbA1c values and mortality risk in type 2 diabetes. population-based case-control study. PLoS One. 2013; 8:e68008
5. Li W, Katzmarzyk PT, Horswell R, Wang Y, Johnson J, Hu G. HbA1c and all-cause mortality risk among patients with type 2 diabetes. Int J Cardiol. 2016; 202:490–496
6. Martono DP, Hak E, Lambers Heerspink H, Wilffert B, Denig P. Predictors of HbA1c levels in patients initiating metformin. Curr Med Res Opin. 2016; 32:2021–2028
7. Balkau B, Calvi-Gries F, Freemantle N, Vincent M, Pilorget V, Home PD. Predictors of HbA1c over 4 years in people with type 2 diabetes starting insulin therapies: the CREDIT study. Diabetes Res Clin Pract. 2015; 108:432–440
8. Mavian AA, Miller S, Henry RR. Managing type 2 diabetes: balancing HbA1c and body weight. Postgrad Med. 2010; 122:106–117
9. van Munster SN, van der Graaf Y, de Valk HW, Visseren FL, Westerink J; SMART Study Group. Effect modification in the association between glycated haemoglobin and cardiovascular disease and mortality in patients with type 2 diabetes. Diabetes Obes Metab. 2017; 19:320–328
10. Hull Geo-type: Clinical Commissioning Group. DIABETES UK. 2014, England: Geowise Ltd. [Accessed 17 March 2020].
11. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser. 1995; 854:1–452
12. Nichols GA, Hillier TA, Javor K, Brown JB. Predictors of glycemic control in insulin-using adults with type 2 diabetes. Diabetes Care. 2000; 23:273–277
13. Awua-Larbi S, Wong TY, Cotch MF, et al. Microalbuminuria and chronic kidney disease as risk factors for cardiovascular disease. Am J Kidney Dis. 2006; 17:2317–2324
14. Selvin E, Marinopoulos S, Berkenblit G, Rami T, Brancati FL, Powe NR, Golden SH. Meta-analysis: glycosylated hemoglobin and cardiovascular disease in diabetes mellitus. Ann Intern Med. 2004; 141:421–431
15. Stratton IM, Adler AI, Neil HA, Matthews DR, Manley SE, Cull CA, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000; 321:405–412
16. Hope C, Robertshaw A, Cheung KL, Idris I, English E. Relationship between HbA1c and cancer in people with or without diabetes: a systematic review. Diabet Med. 2016; 33:1013–1025
17. Abdelhafiz H, Sinclair J. Low HbA1c and increased mortality risk-is frailty a confounding factor? Aging Dis. 2015; 6:262
18. Gonzalez E, Johansson S, Wallander M, Rodriguez L. Trends in the prevalence and incidence of diabetes in the UK : 1996–2005. J Epidemiol Community Health. 2009; 63:332–336

body mass index; obesity; obesity paradox; type 2 diabetes mellitus

Supplemental Digital Content

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.