Vous êtes sur la page 1sur 12

The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Risk Factors, Mortality, and Cardiovascular


Outcomes in Patients with Type 2 Diabetes
Aidin Rawshani, M.D., Araz Rawshani, M.D., Ph.D., Stefan Franzén, Ph.D.,
Naveed Sattar, M.D., Ph.D., Björn Eliasson, M.D., Ph.D., Ann‑Marie Svensson, Ph.D.,
Björn Zethelius, M.D., Ph.D., Mervete Miftaraj, M.Sc.,
Darren K. McGuire, M.D., M.H.Sc., Annika Rosengren, M.D., Ph.D.,
and Soffia Gudbjörnsdottir, M.D., Ph.D.​​

A BS T R AC T

BACKGROUND
Patients with diabetes are at higher risk for death and cardiovascular outcomes than From the Department of Molecular and
the general population. We investigated whether the excess risk of death and cardio- Clinical Medicine, Institute of Medicine
(Aidin Rawshani, Araz Rawshani, B.E., A.
vascular events among patients with type 2 diabetes could be reduced or eliminated. Rosengren, S.G.), and the Health Metrics
Unit, Sahlgrenska Academy (S.F.), Uni-
METHODS versity of Gothenburg, and the Swedish
In a cohort study, we included 271,174 patients with type 2 diabetes who were reg- National Diabetes Register, Center of
Registers in Region (Aidin Rawshani, Araz
istered in the Swedish National Diabetes Register and matched them with 1,355,870 Rawshani, S.F., B.E., A.-M.S., M.M., S.G.),
controls on the basis of age, sex, and county. We assessed patients with diabetes Gothenburg, and the Department of Pub-
according to age categories and according to the presence of five risk factors (ele- lic Health and Caring Sciences–Geriat-
rics, Uppsala University, and the Swedish
vated glycated hemoglobin level, elevated low-density lipoprotein cholesterol level, Medical Products Agency, Uppsala (B.Z.)
albuminuria, smoking, and elevated blood pressure). Cox regression was used to — all in Sweden; the Institute of Cardio-
study the excess risk of outcomes (death, acute myocardial infarction, stroke, and vascular and Medical Sciences, Universi-
ty of Glasgow, Glasgow, United Kingdom
hospitalization for heart failure) associated with smoking and the number of vari- (N.S.); and the Division of Cardiology,
ables outside target ranges. We also examined the relationship between various Department of Internal Medicine, Univer-
risk factors and cardiovascular outcomes. sity of Texas Southwestern Medical Cen-
ter, Dallas (D.K.M.). Address reprint re-
RESULTS quests to Dr. Aidin Rawshani at the
Swedish National Diabetes Register Re-
The median follow-up among all the study participants was 5.7 years, during which gion of Västra Götaland, Medicinaregatan
175,345 deaths occurred. Among patients with type 2 diabetes, the excess risk of 18G, Gothenburg 413 45, Sweden, or at
outcomes decreased stepwise for each risk-factor variable within the target range. ­aidin​.­rawshani@​­gu​.­se.

Among patients with diabetes who had all five variables within target ranges, the N Engl J Med 2018;379:633-44.
hazard ratio for death from any cause, as compared with controls, was 1.06 (95% DOI: 10.1056/NEJMoa1800256
Copyright © 2018 Massachusetts Medical Society.
confidence interval [CI], 1.00 to 1.12), the hazard ratio for acute myocardial in-
farction was 0.84 (95% CI, 0.75 to 0.93), and the hazard ratio for stroke was 0.95
(95% CI, 0.84 to 1.07). The risk of hospitalization for heart failure was consistently
higher among patients with diabetes than among controls (hazard ratio, 1.45; 95% CI,
1.34 to 1.57). In patients with type 2 diabetes, a glycated hemoglobin level outside
the target range was the strongest predictor of stroke and acute myocardial infarction;
smoking was the strongest predictor of death.
CONCLUSIONS
Patients with type 2 diabetes who had five risk-factor variables within the target
ranges appeared to have little or no excess risk of death, myocardial infarction, or
stroke, as compared with the general population. (Funded by the Swedish Associa-
tion of Local Authorities and Regions and others.)

n engl j med 379;7 nejm.org  August 16, 2018 633


The New England Journal of Medicine
Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

T
ype 2 diabetes is a complex disease Data Sources and Study Cohort
that leads to continuous medical care with The Swedish National Diabetes Register has been
comprehensive, multifactorial strategies for described previously.1,4 Type 2 diabetes was de-
reducing cardiovascular risk. Patients with type fined according to epidemiologic criteria — treat-
2 diabetes have risks of death and cardiovascular ment with diet, with or without the use of oral
events that are 2 to 4 times as great as the risks antihyperglycemic agents, or treatment with in-
in the general population.1 Results from random- sulin, with or without the use of oral antihyper-
ized trials support a range of interventions that glycemic agents. The latter category (insulin use)
target isolated risk factors such as elevated levels applied only to patients who were 40 years of age
of glycated hemoglobin, blood pressure, and cho- or older at the time of diagnosis. Patients with at
lesterol to prevent or postpone complications of least one entry in the register between January 1,
type 2 diabetes. The Steno-2 Study investigated 1998, and December 31, 2012, were included in
the effects of multifactorial risk-factor control by the study. At baseline (defined as the first entry
means of behavior modification and pharmaco- in the register), each patient with type 2 diabetes
logic therapy and showed long-lasting reductions was matched for age, sex, and county with five
in the risks of death and cardiovascular events controls without diabetes who were randomly
among patients in whom these risks were reduced, selected from the Swedish population register by
as compared with patients who had been ran- Statistics Sweden.
domly assigned to usual care.2,3 We constructed two cohorts of patients with
The extent to which the excess risk associated type 2 diabetes. One cohort excluded patients
with type 2 diabetes may be mitigated, or poten- with previous stroke, acute myocardial infarction,
tially eliminated, by contemporary evidence-based or amputation; those who had undergone dialy-
treatment and multifactorial risk-factor modifica- sis or renal transplantation; and those with a
tion is unclear. In a nationwide cohort, we evalu- body-mass index (the weight in kilograms divided
ated the association between the excess risks of by the square of the height in meters) of less than
death and cardiovascular outcomes among pa- 18.5. The second cohort of patients with type 2
tients with type 2 diabetes, according to the num- diabetes had exclusion criteria that were similar
ber of risk-factor variables within therapeutic to those of the first cohort but also excluded pa-
guideline levels, as compared with controls who tients with previous coronary heart disease, atrial
were matched for age, sex, and county in Sweden. fibrillation, or heart failure. Controls who met
Risk-factor data were not available for controls. any of these criteria were excluded without the
In ancillary analyses, we estimated the exclusion of their matched patient.
strength of the associations between various risk
factors and the incremental risks of death and Outcomes
cardiovascular outcomes associated with diabe- We assessed death from any cause, fatal or non-
tes. Moreover, we examined the association be- fatal acute myocardial infarction (henceforth re-
tween selected risk-factor variables such as levelsferred to as acute myocardial infarction), fatal or
of glycated hemoglobin, systolic blood pres- nonfatal stroke (henceforth referred to as stroke),
sure, and low-density lipoprotein (LDL) choles- and hospitalization for heart failure. Outcomes
terol within evidence-based target ranges and were identified in hospital discharge records with
these outcomes. use of codes in the International Classification of
Diseases, 9th Revision and 10th Revision. The specific
codes are listed in Table S1 in the Supplementary
Me thods
Appendix, available with the full text of this ar-
Study Design and Support ticle at NEJM.org. Patients were followed until an
The Regional Ethics Review Board of Gothen- event occurred or until December 31, 2013, for all
burg, Sweden, approved the study. All the patients the outcomes except for death from any cause, for
provided written informed consent before inclu- which follow-up ended on December 31, 2014.
sion in the Swedish National Diabetes Register.
The Swedish Association of Local Authorities and Statistical Analysis
Regions and other nonprofit agencies supported Crude incidence rates were calculated according
the study; no industry support was provided. to the number of risk-factor variables within tar-

634 n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

get ranges and are presented as events per 10,000 Relative importance provides an estimate of how
patient-years of observation with exact Poisson important each risk factor is in terms of predict-
confidence intervals of 95%. We constructed Cox ing the outcome. Several methods are available
regression models and included baseline values in for this task; we chose to calculate the relative
the models. All the models were adjusted for so- importance as measured by the R2 values of the
cioeconomic variables (income, marital status, models,7 and we tested the consistency of the re-
immigrant status, and educational level), with sults by calculating the explainable log-likelihood
stratification according to sex to allow for dif- that was attributable to each risk factor.
ferent underlying baseline hazards among men We assessed the level associated with the low-
and women, and age was used as the time scale. est risk, given the glycated hemoglobin level,
Some models were adjusted for preexisting con- systolic blood pressure, and LDL cholesterol level,
ditions. For death from any cause, we adjusted by modeling the association between each risk
for coronary heart disease and heart failure; for factor and the outcomes using restricted cubic
myocardial infarction, we adjusted for atrial fi- splines. The evidence-based target level was set
brillation and heart failure; for stroke, we adjust- as the reference for each risk factor; not smok-
ed for atrial fibrillation, heart failure, and coro- ing was set as the reference for smoking status.
nary heart disease; and for hospitalization for Missing data were imputed with the multi-
heart failure, we adjusted for atrial fibrillation variate imputation by chained equations (MICE)
and coronary heart disease. algorithm. We imputed five complete data sets;
For the main analyses, we estimated the risk a list of the variables that were used in the im-
of each outcome among patients with type 2 putation model is provided in Table S3 in the
diabetes, according to the number of risk-factor Supplementary Appendix. The estimates from each
variables within target ranges, as compared with imputed data set were combined into one overall
the controls matched for age, sex, and county. We estimate with the use of Rubin’s rule. Figures S4
defined whether the risk-factor variables were and S5 in the Supplementary Appendix show the
within target ranges on the basis of guideline- frequency of missing-data elements and the dis-
recommended target levels.5,6 The regression mod- tribution of each variable before and after the
els included the covariable “category,” which de- imputation. For the main analysis, we pooled the
noted the number of risk-factor variables that were results across all five imputed data sets; the re-
not within the target range (scale, none to five sults of the analyses with the imputed data sets
variables). The following five risk factors were were virtually identical to those in the cohort with
considered: the glycated hemoglobin level (cutoff complete cases and were thus consistent with a
value, ≥7.0% or ≥53 mmol per mole), systolic and complete-case analysis. For the ancillary analyses,
diastolic blood pressure (cutoff value, ≥140 mm Hg we used the first imputed data set. All the analy-
for systolic blood pressure or ≥80 for diastolic ses were performed with the use of RStudio soft-
blood pressure), albuminuria (the presence of ware, version 3.2.3.
microalbuminuria or macroalbuminuria), smok-
ing (being a current smoker at study entry), and R e sult s
the LDL cholesterol level (cutoff value, ≥2.5 mmol
per liter [97 mg per deciliter]). Study Population
We constructed a separate Cox model for each A total of 433,619 patients with type 2 diabetes and
age category, according to age at baseline: young- 2,168,095 controls were identified, and 271,174
er than 55 years of age, 55 years to younger than patients with type 2 diabetes and 1,355,870
65 years of age, 65 years to younger than 80 years matched controls were included in the study
of age, and 80 years of age or older. We adjusted (Fig. S1 in the Supplementary Appendix). The
for the duration of diabetes by assigning matched median follow-up among all the study partici-
controls a duration of 0 years, and patients with pants was 5.7 years, during which 175,345 deaths
type 2 diabetes had their duration of diabetes occurred. The baseline characteristics of the pa-
centralized around the grand mean (the mean tients with complete data on all five risk factors
duration among all the study participants). (96,673 patients with diabetes [35.6%]) and their
Among patients with type 2 diabetes, we ana- matched controls are presented in Table 1. The
lyzed the relative importance of the risk factors. number of patients in each risk-factor group in

n engl j med 379;7 nejm.org  August 16, 2018 635


The New England Journal of Medicine
Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

the imputed data sets is also shown in Table 1. 0.72; 95% CI, 0.49 to 1.07) (Fig. 1B). Overall,
The complete data regarding the characteristics patients with type 2 diabetes who had no risk-
of the participants at baseline are presented in factor variables outside the target ranges had a
Table S4 in the Supplementary Appendix. lower risk of acute myocardial infarction than
The mean age of the patients was 60.60 years, the matched controls (hazard ratio, 0.84; 95% CI,
and 47,777 of 96,673 patients (49.4%) were wom- 0.75 to 0.93). Corresponding estimates for the
en. Table S5 in the Supplementary Appendix shows excess risk of stroke are shown in Figure 1C. The
the baseline characteristics of the patients for overall hazard ratio for stroke among patients
whom data were missing for at least one risk fac- with no risk-factor variables outside the target
tor. Figure S6 in the Supplementary Appendix ranges, as compared with controls, was 0.95
shows the trends in risk factors over the period (95% CI, 0.84 to 1.07) (Table S2 in the Supple-
from 1998 through 2012, and Figure S7 in the mentary Appendix). Similar to the findings with
Supplementary Appendix shows how causes of acute myocardial infarction, there was a higher
death varied among the groups according to the incremental risk of stroke in the younger age
number of risk-factor variables in the target categories and for each variable that was not
ranges. within the target range. Estimates regarding hos-
pitalization for heart failure are shown in Fig-
Risk of Cardiovascular Events ure 1D. Patients with type 2 diabetes who were
A total of 37,825 patients with diabetes (13.9%) younger than 55 years of age and had all five
and 137,520 controls (10.1%) died during the risk-factor variables outside the target ranges
study period. The numbers of events, incidence had the highest excess risk of hospitalization for
rates, and hazard ratios for all the outcomes heart failure of all the outcomes assessed (haz-
among patients with diabetes, as compared with ard ratio vs. control, 11.35; 95% CI, 7.16 to 18.01).
controls, at increasing numbers of risk-factor vari- The overall hazard ratio for hospitalization for
ables, as well as the risks of death among men and heart failure among patients with no risk-factor
women, are presented in Table S2 in the Supple- variables outside the target ranges, as compared
mentary Appendix. with controls, was 1.45 (95% CI, 1.34 to 1.57)
Figure 1 shows the adjusted hazard ratios for (Table S2 in the Supplementary Appendix).
the outcomes, according to age category and the
number of risk-factor variables within target rang- Risk-Factor Strength and Levels
es, among patients with diabetes as compared in Patients with Type 2 Diabetes
with matched controls. The results show a step- Figure 2A shows the predictors with the apparent
wise increase in the hazard ratios for each ad- greatest importance with regard to death from
ditional variable that was not within the target any cause. The five strongest predictors regard-
range among patients with diabetes, and the incre- ing the risk of death among patients with type 2
mental risks of cardiovascular events and death diabetes were smoking, physical activity, marital
that were associated with diabetes decreased in status, glycated hemoglobin level, and use of
a stepwise fashion from younger to older age statins (lipid-lowering medication). The data shown
groups. Patients with diabetes who were 80 years in Figure 3A suggest that lower glycated hemo-
of age or older at baseline had the lowest incre- globin levels than are currently recommended in
mental risk of cardiovascular events and death, guidelines were associated with a lower risk of
as compared with controls. In the overall cohort, death.
patients with type 2 diabetes who had no risk- The strongest predictors regarding the risk of
factor variables outside the target ranges had a acute myocardial infarction were the glycated
marginally higher risk of death than the controls hemoglobin level, systolic blood pressure, LDL
(hazard ratio, 1.06; 95% confidence interval [CI], cholesterol level, physical activity, and smoking
1.00 to 1.12). (Fig. 2B). These risk factors showed a linear as-
Patients with diabetes who were 80 years of sociation with the risk of acute myocardial in-
age or older at baseline and had no risk factors farction (Fig. 3B).
outside target ranges had the lowest hazard ratio, The strongest predictors regarding the risk of
as compared with controls, for acute myocardial stroke were the glycated hemoglobin level, systolic
infarction across all the groups (hazard ratio, blood pressure, duration of diabetes, physical ac-

636 n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
Table 1. Baseline Characteristics of the Patients with Type 2 Diabetes and Matched Controls.*

Variable Matched Controls Patients with Diabetes with Complete Data on All Risk Factors
Overall No. of Risk-Factor Variables outside Target Range
0 1 2 3 4 5
No. of participants
Complete-case data set 483,365 96,673 4,852 22,584 39,673 24,341 4927 296
Imputed data set 1 1,355,870 271,174 11,612 57,000 107,840 76,325 17,227 1,170
Imputed data set 2 1,355,870 271,174 11,569 57,033 107,709 76,455 17,213 1,195
Imputed data set 3 1,355,870 271,174 11,685 57,164 107,521 76,517 17,102 1,185
Imputed data set 4 1,355,870 271,174 11,669 57,217 107,551 76,431 17,137 1,169
Imputed data set 5 1,355,870 271,174 11,562 57,246 107,551 76,392 17,219 1,204
Female sex — no. (%) 238,885 (49.4) 47,777 (49.4) 2,525 (52.0) 11,528 (51.0) 19,799 (49.9) 11,713 (48.1) 2,085 (42.3) 127 (42.9)
Age — yr 60.58±10.89 60.58±10.89 60.96±12.06 61.04±11.23 60.93±10.79 60.00±10.54 58.32±10.06 57.27±10.21
Duration of diabetes — yr — 4.53±5.75 4.09±5.28 4.08±5.35 4.29±5.61 5.19±6.18 5.51±6.41 6.09±6.60
Age at diagnosis of diabetes — yr — 56.09±11.11 56.84±12.20 56.97±11.44 56.69±11.02 54.87±10.64 52.76±10.20 51.34±11.11
Glycated hemoglobin†
Millimoles per mole — 53.22±13.96 44.31±5.14 46.75±8.67 51.10±12.32 61.58±15.25 66.28±15.01 70.51±15.45
Percent — 7.02±1.28 6.21±0.47 6.43±0.79 6.83±1.13 7.79±1.39 8.22±1.37 8.60±1.41
LDL cholesterol
Millimoles per liter‡ — 3.00±0.95 1.98±0.39 2.55±0.87 3.09±0.92 3.38±0.85 3.50±0.84 3.65±0.81
Milligrams per deciliter — 116.0±36.7 76.5±15.1 98.6±33.6 119.5±35.5 130.7±32.8 135.3±32.4 141.1±31.3
Total cholesterol — mmol/liter‡ — 5.11±1.05 4.04±0.58 4.64±0.95 5.20±1.01 5.51±0.97 5.67±0.98 5.88±0.99
Current smoker — no. (%) — 16,486 (17.1) 0 886 (3.9) 4,268 (10.8) 7,125 (29.3) 3,911 (79.4) 296 (100)

n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Body-mass index§ — 30.36±5.53 29.53±5.48 29.85±5.37 30.40±5.48 30.85±5.68 30.74±5.67 30.85±5.86
Blood pressure — mm Hg
Systolic — 137.94±17.01 123.02±9.38 131.58±15.38 139.37±16.63 142.96±16.63 144.84±17.17 147.79±18.71
Diastolic — 79.16±9.57 69.54±5.88 75.08±8.89 80.08±9.17 82.35±8.88 83.89±8.92 84.53±9.88
Macroalbuminuria — no. (%) — 4,695 (4.9) 0 177 (0.8) 908 (2.3) 1,802 (7.4) 1,512 (30.7) 296 (100)

Copyright © 2018 Massachusetts Medical Society. All rights reserved.


Estimated GFR — ml/min/1.73 m2¶ — 84.19±21.52 82.29±20.71 83.01±20.59 83.43±20.94 85.90±22.39 88.76±24.72 91.93±29.16
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

Treatment — no. (%)‖


Statin — 57,945 (61.5) 2,907 (61.5) 13,312 (60.4) 24,206 (62.6) 14,449 (61.0) 2,862 (59.8) 209 (73.1)
Antihypertensive agent — 40,553 (42.4) 2,934 (61.0) 11,035 (49.3) 15,839 (40.4) 8,809 (36.7) 1,823 (37.6) 113 (38.4)

Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
* Plus–minus values are means ±SD. Controls were matched for age, sex, and county, on the basis of data from Statistics Sweden. All information regarding patients with type 2 diabetes was
obtained from the National Diabetes Register. Missing data were imputed with the multivariate imputation by chained equations (MICE) algorithm. We imputed five complete data sets (for
details, see Table S3 and Figs. S4 and S5 in the Supplementary Appendix). Percentages in this table are based on the complete-case data set. LDL denotes low-density lipoprotein.
† Concentrations of glycated hemoglobin were based on values from the International Federation of Clinical Chemistry and Laboratory Medicine. Percentages for the glycated hemoglobin
level were based on values from the National Glycohemoglobin Standardization Program.
‡ To convert values for cholesterol to milligrams per deciliter, divide by 0.02586.
§ The body-mass index is the weight in kilograms divided by the square of the height in meters.
¶ The glomerular filtration rate (GFR) was estimated with the use of the Modification of Diet in Renal Disease equation.
‖ Treatment was assessed only in patients with diabetes. Overall, data on statin use were missing for 1106 patients, and data on the use of an antihypertensive agent were missing for

637
2445. Percentages were calculated on the basis of patients with available data.
The n e w e ng l a n d j o u r na l of m e dic i n e

A Excess Mortality in Relation to Range of Risk-Factor Control B Excess Acute Myocardial Infarction in Relation to Range of
Hazard Ratio (95% CI) Risk-Factor Control Hazard Ratio (95% CI)
Control Control
≥80 yr Reference ≥80 yr Reference
≥65 to <80 yr Reference ≥65 to <80 yr Reference
≥55 to <65 yr Reference ≥55 to <65 yr Reference
<55 yr Reference <55 yr Reference
No risk factors No risk factors
≥80 yr 0.99 (0.84–1.17) ≥80 yr 0.72 (0.49–1.07)
≥65 to <80 yr 1.01 (0.92–1.12) ≥65 to <80 yr 0.80 (0.69–0.93)
≥55 to <65 yr 1.15 (1.00–1.34) ≥55 to <65 yr 0.93 (0.73–1.18)
<55 yr 1.29 (0.94–1.77) <55 yr 0.91 (0.62–1.35)
1 Risk factor 1 Risk factor
≥80 yr 0.94 (0.88–1.00) ≥80 yr 1.05 (0.93–1.19)
≥65 to <80 yr 1.05 (1.02–1.09) ≥65 to <80 yr 1.05 (0.97–1.14)
≥55 to <65 yr 1.23 (1.16–1.31) ≥55 to <65 yr 1.14 (1.04–1.25)
<55 yr 1.56 (1.34–1.81) <55 yr 1.46 (1.26–1.69)
2 Risk factors 2 Risk factors
≥80 yr 0.99 (0.94–1.04) ≥80 yr 1.38 (1.27–1.49)
≥65 to <80 yr 1.17 (1.13–1.20) ≥65 to <80 yr 1.44 (1.39–1.50)
≥55 to <65 yr 1.32 (1.27–1.38) ≥55 to <65 yr 1.54 (1.44–1.65)
<55 yr 1.68 (1.56–1.80) <55 yr 2.08 (1.90–2.27)
3 Risk factors 3 Risk factors
≥80 yr 1.13 (1.06–1.21) ≥80 yr 1.78 (1.60–1.98)
≥65 to <80 yr 1.46 (1.42–1.50) ≥65 to <80 yr 2.11 (2.02–2.20)
≥55 to <65 yr 1.63 (1.55–1.71) ≥55 to <65 yr 2.16 (2.02–2.31)
<55 yr 2.21 (2.05–2.37) <55 yr 3.02 (2.80–3.27)
4 Risk factors 4 Risk factors
≥80 yr 1.47 (1.28–1.70) ≥80 yr 2.32 (1.78–3.01)
≥65 to <80 yr 2.10 (1.96–2.26) ≥65 to <80 yr 2.87 (2.62–3.14)
≥55 to <65 yr 2.53 (2.37–2.70) ≥55 to <65 yr 3.32 (3.02–3.66)
<55 yr 2.80 (2.51–3.13) <55 yr 4.56 (4.01–5.18)
5 Risk factors 5 Risk factors
≥80 yr 1.39 (0.51–3.80) ≥80 yr 3.19 (1.23–8.28)
≥65 to <80 yr 3.10 (2.53–3.80) ≥65 to <80 yr 4.60 (3.37–6.29)
≥55 to <65 yr 3.88 (3.07–4.92) ≥55 to <65 yr 4.84 (3.78–6.21)
<55 yr 4.99 (3.43–7.27) <55 yr 7.69 (5.02–11.77)
1 2 3 4 6 8 1 2 3 4 6 8 10

C Excess Stroke in Relation to Range of Risk-Factor Control D Excess Heart Failure in Relation to Range of Risk-Factor Control
Hazard Ratio (95% CI) Hazard Ratio (95% CI)
Control Control
≥80 yr Reference ≥80 yr Reference
≥65 to <80 yr Reference ≥65 to <80 yr Reference
≥55 to <65 yr Reference ≥55 to <65 yr Reference
<55 yr Reference <55 yr Reference
No risk factors No risk factors
≥80 yr 0.95 (0.74–1.22) ≥80 yr 1.12 (0.89–1.41)
≥65 to <80 yr 0.90 (0.76–1.06) ≥65 to <80 yr 1.42 (1.28–1.58)
≥55 to <65 yr 0.94 (0.72–1.23) ≥55 to <65 yr 1.61 (1.31–1.97)
<55 yr 1.22 (0.70–2.13) <55 yr 2.40 (1.63–3.54)
1 Risk factor 1 Risk factor
≥80 yr 1.06 (0.95–1.18) ≥80 yr 1.17 (1.08–1.27)
≥65 to <80 yr 1.11 (1.04–1.18) ≥65 to <80 yr 1.46 (1.39–1.53)
≥55 to <65 yr 1.27 (1.14–1.41) ≥55 to <65 yr 1.80 (1.63–1.98)
<55 yr 1.55 (1.23–1.95) <55 yr 2.37 (1.99–2.82)
2 Risk factors 2 Risk factors
≥80 yr 1.13 (1.04–1.24) ≥80 yr 1.23 (1.15–1.32)
≥65 to <80 yr 1.32 (1.26–1.38) ≥65 to <80 yr 1.62 (1.56–1.68)
≥55 to <65 yr 1.59 (1.50–1.69) ≥55 to <65 yr 2.11 (1.98–2.26)
<55 yr 2.04 (1.76–2.36) <55 yr 2.71 (2.40–3.05)
3 Risk factors 3 Risk factors
≥80 yr 1.35 (1.21–1.51) ≥80 yr 1.42 (1.31–1.54)
≥65 to <80 yr 1.73 (1.65–1.82) ≥65 to <80 yr 2.01 (1.92–2.10)
≥55 to <65 yr 2.13 (2.01–2.27) ≥55 to <65 yr 2.82 (2.63–3.02)
<55 yr 2.78 (2.46–3.16) <55 yr 3.93 (3.50–4.42)
4 Risk factors 4 Risk factors
≥80 yr 1.54 (1.12–2.11) ≥80 yr 1.81 (1.42–2.30)
≥65 to <80 yr 2.31 (2.09–2.55) ≥65 to <80 yr 2.88 (2.64–3.14)
≥55 to <65 yr 2.66 (2.30–3.08) ≥55 to <65 yr 3.85 (3.47–4.26)
<55 yr 3.34 (2.72–4.10) <55 yr 5.70 (4.84–6.71)
5 Risk factors 5 Risk factors
≥80 yr 2.65 (0.96–7.30) ≥80 yr 2.76 (0.82–9.25)
≥65 to <80 yr 3.54 (2.36–5.31) ≥65 to <80 yr 3.93 (2.75–5.60)
≥55 to <65 yr 2.79 (1.88–4.14) ≥55 to <65 yr 6.54 (4.85–8.81)
<55 yr 6.23 (3.22–12.05) <55 yr 11.35 (7.16–18.01)
1 2 3 4 6 8 10 1 2 3 4 5 7 9

638 n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

Figure 1 (facing page). Adjusted Hazard Ratios for Discussion


Outcomes, According to Age Category and Number
of Risk-Factor Variables outside Target Ranges, among Our analysis of Swedish nationwide registry data
Patients with Type 2 Diabetes, as Compared with from 1998 through 2012 showed that patients
Matched Controls. with type 2 diabetes and five selected risk-factor
Hazard ratios show the excess risk of each outcome variables within target range had, at most, mar-
among patients with type 2 diabetes, as compared
with matched controls from the general population,
ginally higher risks of death, stroke, and myocar-
according to age categories and to the number of risk- dial infarction than the general population. The
factor variables (scale, none to five) that were outside study indicates that having all five risk-factor vari-
target ranges currently recommended in guidelines. ables within the target ranges could theoretically
The analysis included patients with type 2 diabetes eliminate the excess risk of acute myocardial in-
and controls matched for age, sex, and county in Swe-
den. We constructed a Cox hazards model for each
farction. However, there was a substantial excess
age category, and these models were adjusted for the risk of hospitalization for heart failure among
covariable “category”; this covariable denotes the patients who had all the variables within target
number of risk-factor variables that were within target ranges. We identified a monotonic relationship
ranges. These Cox model analyses were performed on among younger age, increasing number of vari-
five imputed data sets for each age category, and haz-
ard ratios were pooled from all the data sets with the
ables not within target ranges, and a higher rela-
use of Rubin’s rule. tive risk of adverse cardiovascular outcomes. The
results suggest that there may be greater poten-
tial gains from more aggressive treatment in
younger patients with diabetes.
tivity, and atrial fibrillation (Fig. 2C). Levels be- The following risk factors were considered to
low the guideline target levels for glycated hemo- be the strongest predictors for cardiovascular out-
globin and systolic blood pressure were associated comes and death: low physical activity, smoking,
with lower risks of stroke (Fig. 3C). and glycated hemoglobin, systolic blood-pressure,
Hospitalization for heart failure was predict- and LDL cholesterol levels outside the target rang-
ed primarily by atrial fibrillation and a body-mass es. Using real-world data, we found that levels of
index outside the target range; a low estimated glycated hemoglobin, systolic blood pressure, and
glomerular filtration rate and high glycated he- LDL cholesterol that were lower than target levels
moglobin level were also strong predictors of were associated with lower risks of acute myocar-
this outcome (Fig. 2D). The risk of hospitaliza- dial infarction and stroke.
tion for heart failure was marginally lower at Randomized trials investigating the effect of
glycated hemoglobin levels of less than 53 mmol multifactorial cardiovascular risk-factor interven-
per mole (Fig. 3D). tion in patients with type 2 diabetes are scarce,
The glycated hemoglobin level was the stron- and contemporary studies were designed to mea-
gest or the second strongest predictor regarding sure the cumulative incidence of cardiovascular
the risk of the outcomes in five of the eight mod- events among patients with various risk factors
els (Fig. 2). Smoking was the strongest predictor (e.g., hyperglycemia, hypertension, dyslipidemia,
of death. and microalbuminuria) who received intensive
therapy, as compared with those who received
Relative Risk-Factor Strength and Explained conventional therapy.2,3,8,9 Observational studies
Log Likelihood and randomized trials have shown inconsistent
Figures S2 and S3 in the Supplementary Appen- evidence of effects of glycated hemoglobin levels
dix show the relative strength of the associations below contemporary guideline levels (<7.0%) with
for predictors of cardiovascular outcomes in pa- regard to cardiovascular events and death.10-15
tients with type 2 diabetes, with or without pre- In the present analyses, a glycated hemoglo-
existing conditions, with the use of explained bin level outside the target range was a strong
log-likelihood. The results were broadly consis- predictor for all outcomes, especially for athero-
tent with the results obtained with the use of thrombotic events, which shows the importance
explained relative risk (R2) models. of dysglycemia with regard to these complications.

n engl j med 379;7 nejm.org  August 16, 2018 639


The New England Journal of Medicine
Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

A Death from Any Cause


Patients without Coexisting
All Patients Conditions at Baseline
Smoking Smoking
Physical activity Physical activity
Marital status Marital status
Glycated hemoglobin Glycated hemoglobin
Lipid-lowering medication Estimated GFR
Estimated GFR Lipid-lowering medication
Body-mass index Body-mass index
Income Income
Heart failure Albuminuria
Coronary heart disease Education
Systolic blood pressure Systolic blood pressure
Albuminuria Immigrant
Education Duration of diabetes
Duration of diabetes LDL cholesterol
Immigrant Diastolic blood pressure
LDL cholesterol Blood-pressure medication
Diastolic blood pressure
Blood-pressure medication
R2 R2
0.000 0.005 0.010 0.015 0.020 0.000 0.005 0.010 0.015 0.020 0.025

Increasing Importance Increasing Importance

B Acute Myocardial Infarction


Patients without Coexisting
All Patients Conditions at Baseline
Glycated hemoglobin Glycated hemoglobin
Systolic blood pressure LDL cholesterol
LDL cholesterol Systolic blood pressure
Physical activity Smoking
Smoking Physical activity
Duration of diabetes Estimated GFR
Estimated GFR Duration of diabetes
Income Income
Diastolic blood pressure Diastolic blood pressure
Heart failure Marital status
Blood-pressure medication Education
Marital status Blood-pressure medication
Education Albuminuria
Albuminuria Immigrant
Lipid-lowering medication Lipid-lowering medication
Immigrant Body-mass index
Atrial fibrillation
Body-mass index
R2 R2
0.000 0.005 0.010 0.015 0.020 0.000 0.005 0.010 0.015 0.020

Increasing Importance Increasing Importance

Figure 2. Relative Importance of Risk Factors for Predicting Death from Any Cause, Acute Myocardial Infarction, Stroke,
and Hospitalization for Heart Failure among Patients with Type 2 Diabetes, with or without Preexisting Conditions.
The estimated explained relative risk (i.e., relative importance) shows the strength of the association for various
risk-factor variables (with values outside the target ranges) for predicting death (Panel A), acute myocardial infarc-
tion (Panel B), stroke (Panel C, facing page), and hospitalization for heart failure (Panel D, facing page) among pa-
tients with type 2 diabetes. Results were obtained from the first imputed data set; there were no significant differ-
ences between the sets. The analysis was restricted to patients with type 2 diabetes. We constructed a Cox hazard
model for each outcome, which included every predictor. We then constructed a separate Cox model for each pre-
dictor and permutated covariables from each of these Cox models to estimate the explained relative risk (R 2).
R 2 was generated by developed applications for the Cox model and is bounded between 0 and 1. Risk factors show-
ing a clear and substantial R 2 measure, as compared with other adjacent predictors, are considered to be relevant.
Full definitions of the risk factors and the values that were considered to be outside the target ranges are provided
in the Supplementary Appendix. The body-mass index is the weight in kilograms divided by the square of the height
in meters. LDL denotes low-density lipoprotein, and GFR glomerular filtration rate.

640 n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

C Stroke
Patients without Coexisting
All Patients Conditions at Baseline
Glycated hemoglobin Glycated hemoglobin
Systolic blood pressure Systolic blood pressure
Duration of diabetes Physical activity
Physical activity Duration of diabetes
Atrial fibrillation Income
Income Smoking
Marital status Marital status
Smoking Lipid-lowering medication
Estimated GFR Estimated GFR
Lipid-lowering medication Blood-pressure medication
Blood-pressure medication Diastolic blood pressure
LDL cholesterol LDL cholesterol
Diastolic blood pressure Education
Body-mass index Albuminuria
Heart failure Body-mass index
Albuminuria Immigrant
Education
Immigrant
R2 R2
0.000 0.005 0.010 0.015 0.000 0.005 0.010 0.015

Increasing Importance Increasing Importance

D Heart Failure
Patients without Coexisting
All Patients Conditions at Baseline
Atrial fibrillation Body-mass index
Body-mass index Glycated hemoglobin
Physical activity Physical activity
Estimated GFR Smoking
Glycated hemoglobin Duration of diabetes
Coronary heart disease Income
Income Marital status
Duration of diabetes Estimated GFR
Blood-pressure medication Systolic blood pressure
Systolic blood pressure Blood-pressure medication
Diastolic blood pressure Diastolic blood pressure
Marital status Albuminuria
Smoking Education
Education Lipid-lowering medication
Lipid-lowering medication LDL cholesterol
Albuminuria Immigrant
LDL cholesterol
Immigrant
R2 R2
0.000 0.010 0.020 0.030 0.00 0.01 0.02 0.03 0.04

Increasing Importance Increasing Importance

Low physical activity was also a strong predictor renal mechanisms may contribute to the develop-
of cardiovascular outcomes and death, but ran- ment of heart failure in patients with type 2 dia-
domized trials have not shown long-lasting ben- betes. A high body-mass index was a stronger
eficial effects from increased physical activity in risk factor for heart failure than for other out-
patients with diabetes.16-18 comes, which may explain why the risks associated
With regard to hospitalization for heart fail- with this outcome may continue to be higher
ure, the present analyses showed that the pres- among patients with type 2 diabetes than among
ence of atrial fibrillation, a high body-mass index, controls, since patients with diabetes are, on
and a glycated hemoglobin level and renal func- average, heavier than compared controls.
tion outside the target ranges were the strongest Our study shows, in accordance with previous
predictors. These findings indicate that cardio- studies, that lower systolic blood pressure is asso-

n engl j med 379;7 nejm.org  August 16, 2018 641


The New England Journal of Medicine
Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
642
A Death from Any Cause B Acute Myocardial Infarction

3.0
2.0
2.5

1.6 2.0

Hazard Ratio
Hazard Ratio
1.5
1.2
The

1.0 1.0

0.8 0.8
Glycated Hemoglobin (mmol/mol) Glycated Hemoglobin (mmol/mol)
50 75 100 50 100 150
Systolic Blood Pressure (mm Hg) Systolic Blood Pressure (mm Hg)
100 150 200 100 150 200
LDL Cholesterol (mmol/liter) LDL Cholesterol (mmol/liter)

n engl j med 379;7


0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5

C Stroke D Heart Failure


4.0

nejm.org
n e w e ng l a n d j o u r na l

2.0

The New England Journal of Medicine


of

3.0

1.6

August 16, 2018


2.0

Copyright © 2018 Massachusetts Medical Society. All rights reserved.


m e dic i n e

Hazard Ratio
Hazard Ratio

1.2

1.0 1.0

0.8
Glycated Hemoglobin (mmol/mol) Glycated Hemoglobin (mmol/mol)

Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
50 100 150 50 100 150
Systolic Blood Pressure (mm Hg) Systolic Blood Pressure (mm Hg)
100 150 200 100 150 200
LDL Cholesterol (mmol/liter) LDL Cholesterol (mmol/liter)
0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

Figure 3 (facing page). Association between Levels of


specific trials of blood-pressure reduction to dif-
Glycated Hemoglobin, Systolic Blood Pressure, and ferential targets in patients with type 2 diabetes
LDL Cholesterol and Death from Any Cause, Acute may be warranted.20,22
Myocardial Infarction, Stroke, and Heart Failure in Our observational study has several strengths
­Patients with Type 2 Diabetes. but also some notable limitations. Almost all the
We constructed a Cox model for each outcome and patients with type 2 diabetes in Sweden were
applied a prediction function to assess the relation-
ship between glycated hemoglobin, systolic blood-
included. The epidemiologic definitions of type
pressure, and LDL cholesterol levels and the risks of 2 diabetes and the outcomes are well validated.
death from any cause (Panel A), acute myocardial in- We did not consider changes in the risk-factor
farction (Panel B), stroke (Panel C), and hospitaliza- variables during follow-up, and although this
tion for heart failure (Panel D). Reference values were would have some advantages, the approach we
the contemporary guideline levels: 53 mmol per mole
for the glycated hemoglobin level, 140 mm Hg for the
used minimizes the risk of reverse causation in
systolic blood pressure, and 2.5 mmol per liter (97 mg the interpretation of the results. In addition, we
per deciliter) for the LDL cholesterol level. The dark did not distinguish between patients with all or
lines indicate the hazard function, and the shaded ar- some variables within target range without any
eas 95% confidence intervals. Continuous variables specific intervention and patients who had been
were modeled with restricted cubic splines, whereas
all the categorical variables were stratified. This analy-
medically treated to attain the observed risk-fac-
sis was restricted to patients with type 2 diabetes. To tor levels. We also acknowledge that residual
convert values for cholesterol to milligrams per decili- confounding and reverse causation are impossible
ter, divide by 0.02586. to overcome fully. Finally, given the observational
nature of this work, this cannot be a complete
comparison of the effects of treating risk factors;
ciated with lower risks of cardiovascular outcomes rather, because some patients may have had risk-
and death.19 The Systolic Blood Pressure Interven- factor variables in the target ranges without treat-
tion Trial (SPRINT) showed that systolic blood- ment, the findings represent the prognostic im-
pressure targets below guideline levels in patients portance of such risk factors for persons with
without diabetes were associated with a lower risk diabetes.
of cardiovascular outcomes and death.20 However, In conclusion, patients with type 2 diabetes
the Action to Control Cardiovascular Risk in Dia- who had five risk-factor variables within target
betes (ACCORD) trial examined the same systolic ranges appeared to have little or no excess risks of
blood-pressure targets in patients with type 2 death, myocardial infarction, and stroke as com-
diabetes (<120 mm Hg vs. <140 mm Hg) and did pared with the general population.
not show a significant effect on cardiovascular The views expressed in this article are those of the authors
mortality.15,21 Our analysis implies that systolic and are not necessarily the views of the Swedish Medical Prod-
ucts Agency.
blood pressure is a central factor for virtually all Supported by grants from the Swedish Association of Local
outcomes in patients with diabetes, and lower lev- Authorities and Regions, the Swedish State under the agreement
els of systolic blood pressure are associated with concerning research and education of doctors, Region Västra
Götaland, the Swedish Diabetes Foundation, the Swedish Heart
significantly lower risks of acute myocardial in- and Lung Foundation, the Swedish Research Council (2013-5187
farction and stroke among patients with diabetes. SIMSAM), the Wilhelm and Martina Lundgren Science Founda-
The assessment of systolic blood pressure and its tion, and the Swedish Council for Health, Working Life, and
Welfare.
relation to death and heart failure is more diffi- Disclosure forms provided by the authors are available with
cult, owing to potential reverse causality. More- the full text of this article at NEJM.org.

References
1. Rawshani A, Rawshani A, Franzén S, torial intervention and cardiovascular (version 2012): the Fifth Joint Task Force
et al. Mortality and cardiovascular disease disease in patients with type 2 diabetes. of the European Society of Cardiology and
in type 1 and type 2 diabetes. N Engl J N Engl J Med 2003;​348:​383-93. Other Societies on Cardiovascular Dis-
Med 2017;​376:​1407-18. 4. Eliasson B, Gudbjörnsdottir S. Diabe- ease Prevention in Clinical Practice (con-
2. Gaede P, Lund-Andersen H, Parving tes care — improvement through mea- stituted by representatives of nine societ-
H-H, Pedersen O. Effect of a multifacto- surement. Diabet Res Clin Pract 2014;​106:​ ies and by invited experts). Eur Heart J
rial intervention on mortality in type 2 Suppl 2::​S291-S294. 2012;​33:​1635-701.
diabetes. N Engl J Med 2008;​358:​580-91. 5. Perk J, De Backer G, Gohlke H, et al. 6. Inzucchi SE, Bergenstal RM, Buse JB,
3. Gaede P, Vedel P, Larsen N, Jensen European Guidelines on cardiovascular et al. Management of hyperglycemia in
GVH, Parving H-H, Pedersen O. Multifac- disease prevention in clinical practice type 2 diabetes, 2015: a patient-centered

n engl j med 379;7 nejm.org  August 16, 2018 643


The New England Journal of Medicine
Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
Risk Factors and Cardiovascular Outcomes in Type 2 Diabetes

approach: update to a position statement Effect of intensive control of glucose on 17. Zethelius B, Gudbjörnsdottir S, Elias-
of the American Diabetes Association and cardiovascular outcomes and death in pa- son B, Eeg-Olofsson K, Cederholm J.
the European Association for the Study of tients with diabetes mellitus: a meta- Level of physical activity associated with
Diabetes. Diabetes Care 2015;​38:​140-9. analysis of randomised controlled trials. risk of cardiovascular diseases and mor-
7. Heller G. A measure of explained risk Lancet 2009;​373:​1765-72. tality in patients with type-2 diabetes: re-
in the proportional hazards model. Bio- 13. Stratton IM, Cull CA, Adler AI, Mat- port from the Swedish National Diabetes
statistics 2012;​13:​315-25. thews DR, Neil HAW, Holman RR. Addi- Register. Eur J Prev Cardiol 2014;​21:​244-
8. Wan EYF, Fung CSC, Yu EYT, et al. Ef- tive effects of glycaemia and blood pres- 51.
fect of multifactorial treatment targets sure exposure on risk of complications in 18. Wing RR, Reboussin D, Lewis CE. In-
and relative importance of hemoglobin type 2 diabetes: a prospective observa- tensive lifestyle intervention in type 2 dia-
A1c, blood pressure, and low-density lipo- tional study (UKPDS 75). Diabetologia betes. N Engl J Med 2013;​369:​2358-9.
protein-cholesterol on cardiovascular dis- 2006;​49:​1761-9. 19. Stevens RJ, Kothari V, Adler AI, Strat-
eases in Chinese primary care patients 14. Patel A, MacMahon S, Chalmers J, et ton IM. The UKPDS risk engine: a model
with type 2 diabetes mellitus: a popula- al. Effects of a fixed combination of per- for the risk of coronary heart disease in
tion-based retrospective cohort study. indopril and indapamide on macrovascu- Type II diabetes (UKPDS 56). Clin Sci
J Am Heart Assoc 2017;​ 6(8):​
e006400- lar and microvascular outcomes in pa- Lond 2001;​101:​671-9.
e006414. tients with type 2 diabetes mellitus (the 20. The SPRINT Research Group. A ran-
9. Gæde P, Oellgaard J, Carstensen B, et ADVANCE trial): a randomised controlled domized trial of intensive versus standard
al. Years of life gained by multifactorial trial. Lancet 2007;​370:​829-40. blood-pressure control. N Engl J Med
intervention in patients with type 2 diabetes 15. Ismail-Beigi F, Craven T, Banerji MA, 2015;​373:​2103-16.
mellitus and microalbuminuria: 21 years et al. Effect of intensive treatment of hy- 21. Margolis KL, O’Connor PJ, Morgan
follow-up on the Steno-2 randomised trial. perglycaemia on microvascular outcomes TM, et al. Outcomes of combined cardio-
Diabetologia 2016;​59:​2298-307. in type 2 diabetes: an analysis of the vascular risk factor management strate-
10. Holman RR, Paul SK, Bethel MA, ACCORD randomised trial. Lancet 2010;​ gies in type 2 diabetes: the ACCORD ran-
Matthews DR, Neil HAW. 10-Year follow- 376:​419-30. domized trial. Diabetes Care 2014;​ 37:​
up of intensive glucose control in type 2 16. The Look AHEAD Research Group. 1721-8.
diabetes. N Engl J Med 2008;​359:​1577-89. Association of the magnitude of weight 22. Buckley LF, Dixon DL, Wohlford GF
11. Stratton IM, Adler AI, Neil HA, et al. loss and changes in physical fitness with IV, Wijessinghe DS, Baker WL, Van Tas-
Association of glycaemia with macrovas- long-term cardiovascular disease out- sell BW. Intensive versus standard blood
cular and microvascular complications of comes in overweight or obese people with pressure control in SPRINT-eligible par-
type 2 diabetes (UKPDS 35): prospective type 2 diabetes: a post-hoc analysis of the ticipants of the ACCORD-BP Trial. Diabe-
observational study. BMJ 2000;​321:​405-12. Look AHEAD randomised clinical trial. tes Care 2017;​40:​1733-8.
12. Ray KK, Seshasai SR, Wijesuriya S, et al. Lancet Diabetes Endocrinol 2016;​4:​913-21. Copyright © 2018 Massachusetts Medical Society.

clinical trial registration


The Journal requires investigators to register their clinical trials
in a public trials registry. The members of the International Committee
of Medical Journal Editors (ICMJE) will consider most reports of clinical
trials for publication only if the trials have been registered.
Current information on requirements and appropriate registries
is available at www.icmje.org/about-icmje/faqs/.

644 n engl j med 379;7 nejm.org  August 16, 2018

The New England Journal of Medicine


Downloaded from nejm.org on February 25, 2019. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.

Vous aimerez peut-être aussi