Aims/hypothesis In type 2 diabetes mellitus, heart failure is a frequent, potentially fatal and often forgotten complication. Glucose-lowering agents and adjuvant therapies modify the risk of heart failure. We recently reported that 7.8 years of intensified compared with conventional multifactorial intervention in individuals with type 2 diabetes and microalbuminuria in the Steno-2 study reduced the risk of cardiovascular disease and prolonged life over 21.2 years of follow-up. In this post hoc analysis, we examine the impact of intensified multifactorial intervention on the risk of hospitalisation for heart failure. Methods One hundred and sixty individuals were randomised to conventional or intensified multifactorial intervention, using sealed envelopes. The trial was conducted using the Prospective, Randomised, Open, Blinded Endpoints (PROBE) design. After 7.8 years, all individuals were offered intensified therapy and the study continued as an observational follow-up study for an additional 13.4 years. Heart-failure hospitalisations were adjudicated from patient records by an external expert committee blinded for treatment allocation. Event rates were compared using a Cox regression model adjusted for age and sex. Results Eighty patients were assigned to each treatment group. Ten patients undergoing intensive therapy vs 24 undergoing conventional therapy were hospitalised for heart failure during follow-up. The HR (95% CI) was 0.30 (0.14, 0.64), p =0.002 in the intensive-therapy group compared with the conventional-therapy group. Including death in the endpoint did not lead to an alternate overall outcome; HR 0.51 (0.34, 0.76), p = 0.001. In a pooled cohort analysis, an increase in plasma N-terminal pro-B- type natriuretic peptide (NT-proBNP) during the first two years of the trial was associated with incident heart failure. Conclusions/interpretation Intensified, multifactorial intervention for 7.8 years in type 2 diabetic individuals with microalbuminuria reduced the risk of hospitalisation for heart failure by 70% during a total of 21.2 years of observation. Trial registration: ClinicalTrials.gov NCT00320008. . . . . . Keywords Complications Heart failure Microalbuminuria Multifactorial intervention NT-proBNP Type 2 diabetes Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00125-018-4642-y) contains peer-reviewed but unedited supplementary material, which is available to authorised users. * Oluf Pedersen Faculty of Health, University of Aarhus, Aarhus, Denmark email@example.com Faculty of Health, University of Copenhagen, Copenhagen, Denmark Department of Cardiology, Rigshospitalet, Copenhagen, Denmark Department of Cardiology and Endocrinology, Slagelse Hospital, Slagelse, Denmark Department of Endocrinology, Rigshospitalet, Copenhagen, Denmark Institute for Regional Health Research, University of Southern Denmark, Odense, Denmark Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Blegdamsvej 3B, DK-2100 Kbh Steno Diabetes Center Copenhagen, Gentofte, Denmark Ø, Denmark Diabetologia (2018) 61:1724–1733 1725 Abbreviations Despite the increased risk and unfavourable prognosis, ARB Angiotensin II receptor blocker heart failure in type 2 diabetes traditionally has been sparsely CIF Cumulative incidence function reported in cardiovascular outcomes trials for glucose- CVD Cardiovascular disease lowering drugs [5, 6]. Recent trials, however, have an in- DPP-4 Dipeptidyl peptidase 4 creased focus on heart failure, but with no consensus between EF Ejection fraction studies or specific guidance from regulatory bodies [7, 8]in GLP-1 Glucagon-like peptide 1 the methodology used to define prevalent and incident heart MI Myocardial infarction failure, increasing the risk of bias. NT-proBNP N-terminal pro-B-type natriuretic peptide Congestive heart failure in diabetes mellitus may be divid- SGLT-2 Sodium–glucose cotransporter 2 ed into a primary form termed metabolic or diabetic cardio- SHR Sub-distribution hazard ratio myopathy, and a secondary form predominantly caused by u-AER Urinary albumin excretion rate coronary ischaemia [9–12]; however, distinct definitions that can be used to differentiate do not exist. Causal factors for the development of primary heart failure include hypertension, fluid overload and possibly substrate overload, causing accu- mulation of intracellular fat and subsequent reduced contrac- tility of cardiomyocytes . Introduction In the Steno-2 study, we compared conventional multifac- torial management of patients with type 2 diabetes and Individuals with type 2 diabetes mellitus are at a high risk of microalbuminuria (24 h urinary albumin excretion 30– developing congestive heart failure, having a relative risk at 300 mg) with intensified multifactorial intervention targeting least twice as high as individuals without diabetes, especially known modifiable risk factors with individualised lifestyle when urinary albumin excretion rate (u-AER) is elevated . intervention and tailored polypharmacy [13–15]ata Heart failure is a serious complication in type 2 diabetes, with specialised diabetes clinic. After 3.8 years of intervention, a median survival from diagnosis of 3.5 years and a 5-year patients who received intensified treatment had a reduction mortality rate of 75% . The prognosis of individuals with in the hazard of microvascular complications of around 50% type 2 diabetes and heart failure is worse than that of heart- ; after 7.8 years of intervention, a 53% reduction was seen in cardiovascular endpoints in the arm that was allocated to failure patients without diabetes mellitus . 1726 Diabetologia (2018) 61:1724–1733 intensified intervention . At this point, the formal aspirin therapy. In order to achieve targets, means of regula- randomisation was neutralised and all patients were offered tion were applied in a sequential manner starting with lifestyle intensified treatment as in the original intensive-therapy arm modification including weight loss, smoking cessation and and the trial continued as an observational follow-up study. increased physical activity with subsequent addition of which- Five years after the end of the trial (in total 13.3 years of ever pharmacological agent(s) were needed to achieve the follow-up), we reported a 46% reduction in total mortality of specific targets. patients in the arm originally allocated the intensive interven- tion . We have also recently, in a 21.2-year follow-up after Endpoint definitions and data trial initiation, demonstrated that the intensified multifactorial approach increased life length with a median of 7.9 years, a The primary endpoint of this current follow-up study was life gain that was matched by time free of incident ischaemic hospitalisation with congestive heart failure. The secondary heart disease of 8.1 years . During the same period, pro- endpoints were time-to-first-event of composites of heart fail- gression in nephropathy and loss of kidney function was di- ure or cardiovascular death and of heart failure or all-cause minished, which led to a reduction in the risk of end-stage death. Individuals were followed by in-trial study visits after renal disease that did not fulfil the pre-specified statistical approximately 2, 4, 8, 13 and 21 years after randomisation significance definition, however (p =0.061) . (Fig. 1). At these visits, patients had a comprehensive screen- In the present post hoc analysis, we report the 21.2-year ing for micro- and macrovascular complications performed as risk of developing heart failure in patients with type 2 diabetes described in . and microalbuminuria, who were allocated conventional mul- Heart failure was defined as described in electronic supple- tifactorial care or intensified multifactorial care for a trial pe- mentary material (ESM) Methods, including prespecified riod of 7.8 years. criteria for symptoms, signs and treatment initiated or intensi- fied. Outcome data were extracted from patient records and adjudicated by an external expert committee that was blinded Methods to treatment allocation. Mortality data were collected from the Danish Civil Registry. The detailed protocol for patient inclusion, randomisation and Baseline ejection fraction (EF) was calculated using the intervention, as well as for anthropometric, clinical and bio- modified Quinones formula . We were not able to correct chemical measurements, has been reported previously . for apical movement, resulting in a likely overestimation of The protocol for the follow-up trial was in accordance with baseline EF. No individuals had, however, any symptoms of the declaration of Helsinki and approved by the local ethics heart failure at baseline, thus we assume the effect to be equal committee (Ethics Committee, Capital Region of Denmark; in the two treatment groups of patients. protocol ID number: H-KA-99035-GS, add. 41104) and by the Danish Data Protection Agency (J.Nr. 2015-41-4042) Statistical methods and the trial was registered at ClinicalTrials.gov,number NCT00320008. All participants gave their informed consent Baseline characteristics were compared using a t test for data upon randomisation and confirmed that on follow-up visits. with Gaussian distribution and the Mann–Whitney U test for data with a non-Gaussian distribution. Time-to-event analyses Patients were conducted using a Cox proportional hazards model with treatment as a covariate and illustrated using cumulative inci- One hundred and sixty Danish patients with type 2 diabetes and dence function (CIF) curves. The primary analyses were ad- microalbuminuria were randomised from year 1993, using justed for age and sex. Furthermore, a model adjusted for sealed envelopes, to either conventional or intensified multifac- BMI, plasma N-terminal pro-B-type natriuretic peptide (NT- torial intervention. Individuals were followed for up to 21.2 years proBNP), EF, systolic blood pressure, HbA , u-AER and 1c (Fig. 1). The trial was conducted using the Prospective, GFR measured by Cr-EDTA clearance (corrected for body Randomised, Open, Blinded Endpoints (PROBE) design. surface area ) was applied with a stepwise backward elim- ination approach, with a threshold for staying in the model of Intervention α = 0.10. Proportional hazards were checked by visual inspec- tion of CIF curves. Stepwise backward elimination was cho- The treatment regimen in both randomisation groups was tar- sen to reduce the risk of overfitting the model. We performed get driven, with targets described in Table 1. In the conven- the Cox regressions without elimination of variables as sensi- tional arm, targets followed national recommendations at all tivity analyses. times. The intensive arm had stricter glycaemic, lipid and A number of exploratory analyses were carried out to con- blood pressure targets, and received ACE inhibitors and firm the validity of the results: in order to investigate the Diabetologia (2018) 61:1724–1733 1727 Fig. 1 CONSORT diagram showing patient flow throughout Randomised (n=160) the trial. The first 7.8 years were the active intervention period, after which time the randomisation was neutralised Allocated to conventional treatment and continued as a post-trial Allocated to intensive treatment (n=80) (n=80) observational follow-up study with all remaining patients being offered the same treatment as in the original intensive-therapy group Assessed for microvascular complications Assessed for microvascular complications (n=76) (n=78) Lost to follow-up: n=4 Lost to follow-up: n=2 Died: 4 Died: 2 Assessed for macrovascular complications Assessed for macrovascular complications (n=67) (n=63) Lost to follow-up: n=13 Lost to follow-up: n=17 Died: 13 Died: 17 Interventional part of study ended; all patients offered intensive treatment in 2001 Post-trial follow-up assessment (n=56) Post-trial follow-up assessment (n=40) Lost to follow-up: n=24 Lost to follow-up: n=40 Died: 24 Died: 40 Long term follow-up assessment (n=42) Long term follow-up assessment (n=24) Lost to follow-up: n=38 Lost to follow-up: n=56 Died: 38 Died: 55 Emigrated: 1 Table 1 Treatment targets for treatment groups Intensive Conventional 1993–1999 2000–2001 1993–1999 2000–2001 Systolic blood pressure (mmHg) <140 <130 <160 <135 Diastolic blood pressure (mmHg) <85 <80 <95 <85 HbA (%) <6.5 <6.5 <7.5 <6.5 1c HbA (mmol/mol) <48 <48 <58 <48 1c Fasting serum total cholesterol (mmol/l) <4.9 <4.5 <6.5 <4.9 Fasting serum triacylglycerol (mmol/l) <1.7 <1.7 <2.2 <2.0 Treatment with ACE inhibitor irrespective of BP Yes Yes No Yes Aspirin therapy Known ischaemia Yes Yes Yes Yes Peripheral vascular disease Yes Yes No No No known vascular disease No Yes No No The conventional-therapy group was at all times treated with targets as least as strict as recommended in national guidelines. Aspirin treatment was initiated if any of the indication criteria were met. The mean intervention duration was 7.8 years and thereafter all patients were offered treatment similar to that of the original intensive-therapy group 1728 Diabetologia (2018) 61:1724–1733 difference in distribution of primary and secondary heart- anthropometrics, and to clinical and biochemical measures. failure events between groups, and to evaluate the impact of No individuals had symptoms or clinical signs (including interventions on each type, we used a competing-risks regres- echocardiographic evidence) of heart failure at baseline. sion model (Fine and Gray ) to assess competing risk from Mean levels of plasma NT-proBNP were in the upper range myocardial infarction (MI) and from death in an analysis of the normal interval. All patients had microalbuminuria, but where death was not included in the endpoint. We considered preserved renal function. the cases where MI occurred before a hospitalisation for heart Individuals were followed for up to 21.9 years after base- failure as secondary heart failure. Estimates were, in these line with mean follow-up of 21.2 years for those surviving to analyses, reported as a sub-distribution hazard ratio (SHR) the end of follow-up. with 95% CIs. In addition, we performed sensitivity analyses for the primary and secondary outcomes excluding baseline Primary and secondary outcomes EF from the further-adjusted Cox regressions of primary and secondary outcomes. These analyses were carried out because Primary outcome Over the entire study course, ten patients of the observation that baseline EF was missing for 15 indi- (13%) in intensive-therapy group developed heart failure vs viduals (9.4%) (five in the intensive-therapy group; ten in the 24 patients (30%) in the conventional-therapy group. All conventional-therapy group). We also investigated whether events complied with the formal definition of the primary the change in plasma NT-proBNP and u-AER values during outcome. The age- and sex-adjusted HR was 0.30 (0.14, the first two years of intervention was associated with the 0.64), p = 0.002 in the intensive-therapy group compared with primary and secondary outcomes by adding the delta value the conventional-therapy group (Fig. 2). The further-adjusted divided into tertiles to the further-adjusted Cox regression. (BMI, HbA , u-AER, GFR, EF, systolic blood pressure and 1c Numerical results are presented followed by 95% CIs in NT-proBNP) HR was 0.23 (0.10, 0.54), p = 0.001 with age brackets. Significance level was set at α = 0.05. Statistical (p =0.017), BMI (p =0.010) baseline EF (p =0.032) staying analyses were performed using STATA/IC version 15 in the final model as independent covariates along with treat- (StataCorp, College Station, TX, USA). ment allocation. Baseline GFR (p = 0.065), u-AER (p = 0.070) and HbA (p = 0.077) were all of borderline 1c significance. Results Secondary outcome The secondary outcomes are illustrated in Baseline patient characteristics are presented in Table 2. Fig. 3. Eighteen patients (23%) in the intensive-therapy group Individuals in the two groups were similar with regards to and 35 patients (44%) in the conventional-therapy group Table 2 Baseline clinical, an- Baseline 1993 thropometric and biochemical data Clinical variable (mean ± SD) Intensive (N = 80) Conventional (N =80) Age (years) 54.9 ± 7.2 55.2 ± 7.2 Proportion males (%) 79 70 Diabetes duration (years) median (range) 4 (0; 30) 6 (0; 29) Systolic BP (mmHg) 146 ± 11 149 ± 19 HbA – (mmol/mol) 68 ± 6 73 ± 5 1c HbA – (%) 8.4 ± 2.7 8.8 ± 2.6 1c BMI (kg/m ) (SD) 29.7 (3.8) 29.9 (4.9) GFR (ml/min/1.73m ) 116 ± 24 118 ± 25 u-AER (mg/24 h) median (IQR) 78 (61; 120) 69 (47; 113) Plasma NT-proBNP (pmol/l) median (IQR) 35 (12; 71) 32 (13; 67) Plasma NT-proBNP ≥ 100 pmol/l (N [%]) 14 (18) 16 (20) b c Left ventricle EF (SD) 67 (8) 67 (8) One patient with missing data Five patients (9%) with missing data Ten patients (12.5%) with missing data IQR, interquartile range Diabetologia (2018) 61:1724–1733 1729 0 4 8 12 16 20 Time (years) since randomisation 0 Number at risk 0 4 8 12 16 20 Intensive 80 75 65 56 51 40 Time (years) since randomisation Conventional 80 74 57 38 29 21 Number at risk Fig. 2 CIF plot of hospitalisation for heart failure. Dashed line, conven- Intensive 80 75 65 56 51 40 tional therapy; solid line, intensive therapy. The unadjusted relative haz- Conventional 80 74 57 38 29 21 ard reduction was 69% in the intensive-therapy group. Logrank p =0.001 reached the secondary endpoint of heart-failure hospitalisation or death from CVD, leading to an age- and sex-adjusted HR of 0.38 (0.22, 0.68), p = 0.001. Further adjusted, the HR was 0.31 (0.16, 0.58), p = 0.001. Age (p = 0.001), BMI (p = 0.023) and baseline HbA (p = 0.011), EF (p = 0.003) and 1c GFR (p = 0.011) stayed in the model after elimination and u- AER (p = 0.084) was of borderline significance. Forty-one patients (51%) originally allocated to intensified 0 4 8 12 16 20 therapy and 59 patients (74%) allocated to conventional ther- Time (years) since randomisation Number at risk apy reached the combined endpoint of heart failure or death Intensive 80 75 65 56 51 40 from all causes. The HR was 0.51 (0.34, 0.76), p =0.001. This Conventional 80 74 57 38 29 21 estimate was stable in the further-adjusted model, with age (p Fig. 3 CIF plots of the secondary outcomes. Dashed line, conventional < 0.001), GFR (p = 0.050) and plasma NT-proBNP (p = therapy; solid line, intensive therapy. (a) Heart failure or cardiovascular 0.033) staying in the model in addition to treatment allocation, death. The unadjusted relative hazard reduction was 61% in the intensive- therapy group. Logrank p < 0.001. (b) Heart failure or death from all and with EF (p =0.074) and HbA (p = 0.090) being of bor- 1c causes. The unadjusted relative hazard reduction was 48% in the inten- derline significance. sive-therapy group. Logrank p =0.001 Exploratory analyses the conventional-therapy group died during follow-up Competing risks from MI In total, nine individuals in the (Fig. 4). Median time from hospitalisation for heart failure intensive-therapy group and 23 in the conventional-therapy to death was 2.9 years and did not differ between the two group experienced an MI during the study. Of those, two in treatment groups. For MI, the corresponding numbers the intensive-therapy group and nine in the conventional- were six (67%) patients in the intensive-therapy and 18 therapy group experienced MI prior to hospitalisation for (75%) in the conventional-therapy group with median heart failure (Figs 4a, b). The SHR of heart failure controlled time from MI to death being 2.1 years (please note that for MI in the competing-risks regression was 0.37 (0.17, these numbers differ from the numbers in Figs 4a, b 0.78), p = 0.009, meaning that the observed difference in because there, patients may progress from, e.g., MI to hospitalisation for heart failure was not explained by the dif- heart failure and then die subsequently, which is not ference in prior MI between groups. A competing-risks re- possible to control for in the simple analysis described gression using all-cause death as the competing event to heart above). failure also confirmed the result with SHR of 0.36, p =0.008. Among individuals with incident heart failure, seven Sensitivity analyses excluding EF in the Cox regression (70%) in the intensive-therapygroup and20(83%) in Estimates of HR for the primary and the secondary endpoints Cumulative incidence of heart failure (%) Cumulative incidence of Cumulative incidence of heart failure or heart failure or death (%) cardiovascular death (%) 1730 Diabetologia (2018) 61:1724–1733 a b MI MI 8 18 1 2 5 1 5 13 80 DM2 Death DM2 Death 27 80 27 6 38 15 55 2 9 8 15 HF HF 3 4 100 100 HF HF MI 75 75 Death MI Death 50 Alive with no HF or MI Alive with no HF or MI 25 25 0 0 0 5 10 15 20 0 5 10 15 20 Time from randomisation (years) Time from randomisation (years) Fig. 4 Transition frequencies from entry to MI, HF and/or death. (a) patients in the intensive-therapy and 29 in the conventional-therapy group Intensive-therapy group and (b) conventional-therapy group: arrows ter- were classified as having died from non-CV causes and 12 vs 26 patients minate at the event and originate from the original state of the patients. died from CV causes, respectively. Thirty-seven patients in the intensive- The black number at the arrow end is the number with the given event therapy group and 20 patients in the conventional-therapy group ended coming from the state at arrow origin. The coloured number in the bottom follow-up alive and with no incident HF or MI during follow-up. (c) right corner is the number of patients not progressing from the given state. Intensive-therapy group and (d) conventional-therapy group: survival Example: ten intensive-therapy patients developed HF (pale orange box, frequencies without MI/HF. Both MI and HF were more frequent in the Fig. 4a). Eight had no previous MI; two developed HF after previous MI. conventional-therapy group (b) and the difference in HF was not driven One developed MI after HF, six died after HF and three ended the obser- primarily by increased MI risk. DM2, type 2 diabetes mellitus; HF, heart vation alive with HF. Twenty-seven patients died without prior MI or HF. failure In the primary analysis of data from the 21.2 years of follow-up, 26 were stable when removing baseline EF from the model (ESM Change in plasma NT-proBNP over the first two years of fol- Table 1). Using Cox regression without elimination of vari- low-up Change in plasma NT-proBNP during the first two ables confirmed the results, without any change of HR esti- years was significantly associated with the primary outcome mates (ESM Table 2). (in the period from year 2 to 21) in the adjusted Cox regression The inclusion of plasma NT-proBNP as a dichotomous model. Individuals (pooled cohort) with changes in the highest variable with a cut-off of 100 pmol/l in the further-adjusted tertile of delta plasma NT-proBNP (i.e. the largest increase) model instead of as a continuous variable showed a significant had a 2.7-fold ([1.19, 5.93], p = 0.018) increased risk of heart association with heart failure (HR 3.2 for high vs low NT- failure as compared with individuals in the lower two tertiles proBNP [p = 0.09]) and with heart failure plus all-cause death of change. Individuals (pooled cohort) in the second tertile did (HR for high vs low NT-proBNP 1.9 [p =0.015]). not differ significantly in risk from those in the lowest tertile Survived without HF or MI (%) Survived without HF or MI (%) Diabetologia (2018) 61:1724–1733 1731 (p = 0.14). Assessing an interaction between change in NT- follow-up, thus allowing more time for atherosclerosis pro- proBNP and treatment allocation revealed that this finding gression. It is unlikely that the risk reductions seen in the was driven by the conventional-therapy group. In the Steno-2 study are attributable to one single component of conventional-therapy group, individuals with changes in the the multifactorial treatment regimen, but rather to the combi- higher tertile had a threefold increased risk of heart failure nation of polypharmacy and lifestyle intervention. The drug compared with those in the two lower tertiles (p = 0.030), pattern was highly complex (ESM Figs 1, 2), but the use of whereas this was not the case for the intensive-therapy group ACE inhibitors, angiotensin II receptor blockers (ARBs), (p = 0.60). Change in plasma NT-proBNP was not associated statins and aspirin was more frequent in the intensive- with cardiovascular mortality combined with heart failure, but therapy group, which is probably reflected in our findings. had a positive correlation with all-cause mortality; HR = 1.86 In particular, we would anticipate the effect of more frequent (1.18, 2.92), p = 0.07 for the highest tertile compared with the use of ACE inhibitors and ARBs to be prominent. lower two tertiles. These findings were not affected by treat- Our findings suggest that diabetes caregivers should pay ment allocation, but significantly more patients in the conven- attention to early signs and symptoms of congestive heart tionally treated therapy group were in the upper tertile of plas- failure including left-ventricular dysfunction and, together ma NT-proBNP (32 vs 19; p =0.041). with cardiologists, take appropriate actions. Measuring plas- In similar analyses, change in albuminuria was not associ- ma NT-proBNP might be helpful in assessing and monitoring ated with any of the outcomes (ESM Table 3). risk of heart failure since high values or large increments were strongly associated with heart-failure outcomes and could be used to guide diabetologists to refer patients for evaluation by Discussion cardiologists. Our findings indicate that the higher risk of heart failure associated with an increase in NT-proBNP can be mit- Heart failure is a major health issue and a risk factor for early igated by intensified multifactorial intervention. However, ev- death and disability in type 2 diabetes. Until now, no trials of idence that intensive lifestyle improvement may increase NT- intensified multifactorial intervention in type 2 diabetes have proBNP and the fact that NT-proBNP levels are inversely investigated the potential beneficial effects of this treatment related to BMI warrants caution when interpreting our find- modality on congestive heart failure. In this post hoc analysis ings regarding NT-proBNP [25, 26]. of outcome data from the Steno-2 randomised trial, we dem- Glucose-lowering drugs, as well as adjuvant therapies for onstrate that intensified multifactorial treatment in patients diabetes comorbidities, modify the risk of developing heart with type 2 diabetes mellitus, microalbuminuria and preserved failure. In 2009, increased risk of heart failure in diabetic baseline EF for 7.8 years reduced the hazard of developing patients treated with rosiglitazone in the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of clinical heart failure by 70% at 21.2 years of follow-up. Twenty per cent of all patients developed heart failure, Glycaemia in Diabetes (RECORD) trial  was reported equalling the number of patients experiencing MI (20%), and subsequently, the use of dipeptidyl peptidase 4 (DPP-4) stroke (22%), amputation (19%) or blindness (19%)  inhibitors also has been flagged as potentially increasing the and mortality seemed to be at least as high for patients with risk of heart failure . Although the latter findings have incident heart failure as for patients experiencing an MI. We been questioned [29, 30], the Food and Drug Administration also found that the reduction in heart failure was influenced (FDA) recently added heart-failure warnings to the Summary by, but not dependent on, reduction of ischaemic heart disease of medicinal Products Characteristics (SmPC) for all marketed risk. The results were unchanged over multiple sensitivity DPP-4-inhibitors . In contrast, diabetes trials with the so- analyses including testing for common confounders and com- dium–glucose cotransporter 2 (SGLT-2) inhibitors peting risk. Our results clearly underline the importance of empagliflozin and canagliflozin have demonstrated heart- applying an intensive, multifactorial approach to the manage- failure risk reduction [32–34], whereas diabetes treatment ment of type 2 diabetes. with GLP-1 receptor agonists seem to have a neutral effect In the following studies (all including patients with type 2 on the risk of heart failure [35–37]. In addition, the extensive diabetes and micro- or macroalbuminuria), heart failure was use of ACE inhibitors [24, 38], ARBs, beta blockers and more common than other cardiovascular complications: re- statins  in type 2 diabetes may modify the risk of heart duction of Endpoints in NIDDM with the Angiotensin II failure, although the evidence regarding the latter is weak. Antagonist Losartan (RENAAL) , the Irbesartan This post hoc analysis has limitations: the Steno-2 study Diabetic Nephropathy Trial (IDNT) , the Aliskiren Trial has a small sample size of 160 individuals and the type 2 in Type 2 Diabetes Using Cardiorenal Endpoints diabetes study population with microalbuminuria is selected (ALTITUDE)  and Microvascular outcomes in the Heart as a high-risk patient group. Therefore, the magnitude of risk Outcomes Prevention Evaluation Study (MICRO-HOPE) reduction we demonstrate here might not be attributable to a . However, the Steno-2 study had significantly longer population of lower risk. Furthermore, the sample size leads to 1732 Diabetologia (2018) 61:1724–1733 in study design or in the collection, analysis or interpretation of data and an inherent risk of committing type 1 statistical errors due to the present post hoc analysis did not receive any additional funding. The unaccountable bias. The fact that the results presented in this study sponsor was not involved in the design of the study; the collection, paper are in accordance with previously published results re- analysis, and interpretation of data; writing the report; or the decision to garding other manifestations of cardiovascular disease can be submit the report for publication. regarded both as a strength and as a weakness, the latter Duality of interest Since completion of the Steno-2 21.2 years follow-up underlined by the above statement of unmeasurable bias driv- data acquisition, data management and interpretation, JO has been ing the difference in outcomes between groups. employed by Novo Nordisk Scandinavia A/B, Region Denmark. PR We do not have complete echocardiographic data available reports having given lectures for Astra Zeneca, Bayer and Boehringer for the entire patient population and baseline EF was based on Ingelheim, has served as a consultant for AbbVie, Astra Zeneca, Bayer, Eli Lilly, Boehringer Ingelheim, Astellas, Janssen and Novo Nordisk (all a calculated estimate. Therefore, whether individuals had pre- fees given to the Steno Diabetes Center) and has equity interest in Novo served or reduced EF prior to and after hospitalisation is un- Nordisk. HHP has equity interest in Merck and receives honoraria from certain. However, recent evidence shows that the prognosis is AbbVie and Novartis. OP has equity interest in Novo Nordisk A/S. The independent of EF . Novo Nordisk Foundation Center for Basic Metabolic Research is an independent research centre at the University of Copenhagen partially In the Steno-2 trial and during trial follow-up, the use of funded by an unrestricted donation from the Novo Nordisk Foundation. sulfonylurea drugs was frequent and even between groups PG, RR and LK declare that there is no duality of interest associated with (ESM Fig. 1) and the use of ACE inhibitors and beta blockers this manuscript. was more frequent in the intensive-therapy group (ESM Fig. Contribution statement OP conceived and designed the original Steno-2 2), but the total exposure to glitazones, DPP-4 inhibitors, study, devised and supervised the 21.2 years follow-up, acquired all GLP-1 inhibitors and SGLT-2 inhibitors etc. was very limited funding throughout the study and provided key content to the manuscript. (all <5% of patients in each group at each follow-up point). HHP contributed to the conception and design of the original Steno-2 The effects on heart failure of the above-mentioned drugs that study and provided key content to the manuscript. PG acquired all data have been seen in recent large-scale clinical trials of cardio- up to the 13.3 years examination, handled patient care during the inter- vention period, planned the 21.2 years follow-up and supervised data vascular outcomes occur on top of the standard of care; in acquisition and processing in the present follow-up. JO coordinated and most current recommendations, the standard of care resembles performed the 21.2 years follow-up, acquired all data for the 21.2 years the treatment targets for the patients originally allocated to follow-up, processed data and performed all statistical work including intensified multifactorial care in the Steno-2 trial. preparation of figures. PR supervised and facilitated all work performed at the Steno Diabetes Center (i.e. the patient assessments for the long-term The significant and meaningful risk reductions reported follow-up), was responsible for quality assurance/quality control of the here should be seen as a benefit of intervention against tradi- performed measures and provided key intellectual content with regards to tional risk factors; the introduction of novel glucose-lowering results interpretation and discussion. RR provided assistance with inter- drugs with pleiotropic effects, on top of improved glucose pretation of statistical calculations, interpretation of echocardiographic data and has contributed significantly to the introduction and results sec- control, that reduce the risk of cardiovascular morbidity and tions. LK defined heart failure outcomes, assisted in data interpretation mortality by other, not presently fully elucidated, mechanisms and presentation and provided significant content to the introduction and should lead to optimism in the field of diabetes care. discussion sections. The manuscript was drafted by JO with significant contributions from all other authors. All authors gave final approval for In conclusion, our study demonstrated hospitalisation for the paper to be published. OP, HHP and JO are guarantors of this work. heart failure to be a frequent and fatal complication in patients with type 2 diabetes and microalbuminuria. Intensified multi- factorial intervention significantly reduced the occurrence of Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// this outcome. creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appro- Acknowledgements We wish to thank all the individuals who participat- priate credit to the original author(s) and the source, provide a link to the ed in the trial as well as their relatives. We also wish to thank P. Creative Commons license, and indicate if changes were made. Hildebrandt, DMSc (Frederiksberg Heart Clinic, Copenhagen, Denmark) and J. Faber, DMSc (Herlev Hospital, Copenhagen, Denmark) for serving in the endpoint assessment committee throughout the entire follow-up. Furthermore, we wish to thank I. Holstein (Steno Diabetes Center Copenhagen, Denmark) and G. Lademann (Center for References Basic Metabolic Research, Copenhagen, Denmark) for essential manage- rial assistance. Some of the data presented in this paper were presented as an abstract at the 53rd EASD Annual Meeting in 2017. 1. Gerstein HC, Mann JF, Yi Q et al (2001) Albuminuria and risk of cardiovascular events, death, and heart failure in diabetic and non- Data availability Data is available upon reasonable request to the cor- diabetic individuals. JAMA 286:421–426 responding author. 2. Johansson I, Edner M, Dahlstrom U, Nasman P, Ryden L, Norhammar A (2014) Is the prognosis in patients with diabetes Funding The Steno-2 study, including the 21.2 years of follow-up, and heart failure a matter of unsatisfactory management? An obser- was supported by unrestricted grants from Novo Nordisk A/S, vational study from the Swedish heart failure registry. Eur J Heart Bagsværd, Denmark. Novo Nordisk A/S were not in any way involved Fail 16:409–418 Diabetologia (2018) 61:1724–1733 1733 3. Shah KS, Xu H, Matsouaka RA et al (2017) Heart failure with 21. Brenner BM, Cooper ME, de Zeeuw D et al (2001) Effects of losartan on renal and cardiovascular outcomes in patients with type preserved, borderline, and reduced ejection fraction: 5-year out- comes. J Am Coll Cardiol 70:2476–2486 2 diabetes and nephropathy. N Engl J Med 345:861–869 4. van den Berge JC, Constantinescu AA, Boiten HJ, van Domburg 22. Berl T, Hunsicker LG, Lewis JB et al (2003) Cardiovascular out- RT, Deckers JW, Akkerhuis KM (2018) Short- and long-term prog- comes in the Irbesartan diabetic nephropathy trial of patients with nosis of patients with acute heart failure with and without diabetes: type 2 diabetes and overt nephropathy. Ann Intern Med 138:542– changes over the last three decades. Diabetes Care 41:143–149 549 5. Gilbert RE, Krum H (2015) Heart failure in diabetes: effects of anti- 23. Parving HH, Brenner BM, McMurray JJ et al (2012) Cardiorenal hyperglycaemic drug therapy. Lancet 385:2107–2117 end points in a trial of aliskiren for type 2 diabetes. N Engl J Med 6. McMurray JJ, Gerstein HC, Holman RR, Pfeffer MA (2014) Heart 367:2204–2213 failure: a cardiovascular outcome in diabetes that can no longer be 24. Heart Outcomes Prevention Evaluation (HOPE) study investigators ignored. Lancet Diabetes Endocrinol 2:843–851 (2000) Effects of ramipril on cardiovascular and microvascular out- 7. US Department of Health and Human Services, Food and Drug comes in people with diabetes mellitus: results of the HOPE study Administration, Center for Drug Evaluation and Research and MICRO-HOPE substudy. Lancet 355:253–259 (CDER) (2008) Guidance for industry: diabetes mellitus—evaluat- 25. Madamanchi C, Alhosaini H, Sumida A, Runge MS (2014) Obesity ing cardiovascular risk in new antidiabetic therapies to treat type 2 and natriuretic peptides, BNP and NT-proBNP: mechanisms and diabetes. Available from www.fda.gov/downloads/Drugs/ diagnostic implications for heart failure. Int J Cardiol 176:611–617 Guidances/ucm071627.pdf 26. Fedele D, Bicchiega V, Collo A et al (2017) Short term variation in 8. European Medicines Agency (2012) Guideline on clinical investi- NTproBNP after lifestyle intervention in severe obesity. PLoS One gation of medicinal products in the treatment or prevention of dia- 12:e0181212 betes mellitus. Available from www.ema.europa.eu/docs/en_GB/ 27. Home PD, Pocock SJ, Beck-Nielsen H et al (2009) Rosiglitazone document_library/Scientific_guideline/2012/06/WC500129256. evaluated for cardiovascular outcomes in oral agent combination pdf therapy for type 2 diabetes (RECORD): a multicentre, randomised, 9. Lehrke M, Marx N (2017) Diabetes mellitus and heart failure. Am J open-label trial. Lancet 373:2125–2135 Med 130:S40–S50 28. Li L, Li S, Deng K et al (2016) Dipeptidyl peptidase-4 inhibitors 10. Aroor AR, Mandavia CH, Sowers JR (2012) Insulin resistance and and risk of heart failure in type 2 diabetes: systematic review and heart failure: molecular mechanisms. Heart Fail Clin 8:609–617 meta-analysis of randomised and observational studies. BMJ 352: 11. Larghat AM, Swoboda PP, Biglands JD, Kearney MT, Greenwood i610 JP, Plein S (2014) The microvascular effects of insulin resistance 29. Ummarino D (2016) Diabetes: no increased risk of heart failure and diabetes on cardiac structure, function, and perfusion: a cardio- with incretin-based drugs. Nat Rev Cardiol 13:248 vascular magnetic resonance study. Eur Heart J Cardiovasc Imaging 30. Filion KB, Azoulay L, Platt RW et al (2016) A multicenter obser- 15:1368–1376 vational study of incretin-based drugs and heart failure. N Engl J 12. Nolan CJ, Ruderman NB, Kahn SE, Pedersen O, Prentki M (2015) Med 374:1145–1154 Insulin resistance as a physiological defense against metabolic 31. Ryden L, Van de Werf F, Armstrong PW et al (2017) Corrections stress: implications for the management of subsets of type 2 diabe- needed to 2016 ESC and AHA guidelines on heart failure. Lancet tes. Diabetes 64:673–686 Diabetes Endocrinol 5:325–326 13. Gæde P, Vedel P, Parving HH, Pedersen O (1999) Intensified mul- 32. Fitchett D, Zinman B, Wanner C et al (2016) Heart failure outcomes tifactorial intervention in patients with type 2 diabetes mellitus and with empagliflozin in patients with type 2 diabetes at high cardio- microalbuminuria: the Steno type 2 randomised study. Lancet 353: vascular risk: results of the EMPA-REG OUTCOME(R) trial. Eur 617–622 Heart J 37:1526–1534 14. Gæde P, Vedel P, Larsen N, Jensen GV, Parving HH, Pedersen O 33. Raz I, Cahn A (2016) Heart failure: SGLT2 inhibitors and heart (2003) Multifactorial intervention and cardiovascular disease in pa- failure—clinical implications. Nat Rev Cardiol 13:185–186 tients with type 2 diabetes. N Engl J Med 348:383–393 34. Neal B, Perkovic V, Mahaffey KW et al (2017) Canagliflozin and 15. Gæde P, Lund-Andersen H, Parving HH, Pedersen O (2008) Effect cardiovascular and renal events in type 2 diabetes. N Engl J Med of a multifactorial intervention on mortality in type 2 diabetes. N 377:644–657 Engl J Med 358:580–591 35. Holman RR, Bethel MA, Mentz RJ et al (2017) Effects of 16. Gæde P, Oellgaard J, Carstensen B et al (2016) Years of life gained once-weekly exenatide on cardiovascular outcomes in type 2 by multifactorial intervention in patients with type 2 diabetes diabetes. N Engl J Med 377:1228–1239 mellitus and microalbuminuria: 21 years follow-up on the Steno-2 36. Marso SP, Daniels GH, Brown-Frandsen K et al (2016) Liraglutide randomised trial. Diabetologia 59:2298–2307 and cardiovascular outcomes in type 2 diabetes. N Engl J Med 375: 17. Oellgaard J, Gæde P, Rossing P, Persson F, Parving HH, Pedersen O 311–322 (2017) Intensified multifactorial intervention in type 2 diabetics 37. Marso SP, Bain SC, Consoli A et al (2016) Semaglutide and car- with microalbuminuria leads to long-term renal benefits. Kidney diovascular outcomes in patients with type 2 diabetes. N Engl J Int 91:982–988 Med 375:1834–1844 18. Quinones MA, Waggoner AD, Reduto LA et al (1981) A new, 38. Heart Outcomes Prevention Evaluation Study Investigators, Yusuf simplified and accurate method for determining ejection fraction S, Sleight P et al (2000) Effects of an angiotensin-converting- with two-dimensional echocardiography. Circulation 64:744–753 enzyme inhibitor, ramipril, on cardiovascular events in high-risk 19. Brochner-Mortensen J (1972) A simple method for the determina- patients. N Engl J Med 342:145–153 tion of glomerular filtration rate. Scand J Clin Lab Invest 30:271– 39. Preiss D, Campbell RT, Murray HM et al (2015) The effect of statin 274 therapy on heart failure events: a collaborative meta-analysis of 20. Fine JP, Gray RJ (1999) A proportional hazards model for the unpublished data from major randomized trials. Eur Heart J 36: subdistribution of a competing risk. J Am Stat Assoc 94:496 1536–1546
Diabetologia – Springer Journals
Published: May 30, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera