Risk of all-cause mortality according to the European Society of Cardiology risk categories in individuals with type 2 diabetes: the Renal Insufficiency And Cardiovascular Events (RIACE) Italian Multicenter Study

Aims The 2019 and 2021 European Society of Cardiology (ESC) classifications stratified patients with type 2 diabetes into three categories according to the 10-year risk of death from atherosclerotic cardiovascular disease (ASCVD). The very high-risk category included individuals with established ASCVD, target organ damage (TOD), and/or, in the 2019 classification only, ≥ 3 additional ASCVD risk factors. We assessed risk of all-cause mortality according to the two ESC classifications in the Renal Insufficiency And Cardiovascular Events cohort. Methods Participants (n = 15,773) were stratified based on the presence of ASCVD, TOD, and ASCVD risk factors at baseline (2006–2008). Vital status was retrieved in 2015. Results Less than 1% of participants fell in the moderate-risk category. According to the 2019 classification, ~ 1/3 fell in the high-risk and ~ 2/3 in the very high-risk category, whereas the opposite occurred with the 2021 classification. Mortality risk increased across categories according to both classifications. Among very high-risk patients, mortality was much lower in those with ≥ 3 additional ASCVD risk factors and almost equal in those with TOD and ASCVD ± TOD, using the 2019 classification, whereas it was much higher in those with ASCVD + TOD and, to a lesser extent, TOD only than in those with ASCVD only, using the 2021 classification. Conclusions The negligible number of moderate-risk patients suggests that these classifications might overestimate risk of ASCVD death. Downgrading patients with ≥ 3 additional ASCVD risk factors to the high-risk category is consistent with mortality data. Risk of death is very high in the presence of TOD irrespective of established ASCVD. Trial registration: ClinicalTrials.gov, NCT00715481. Supplementary Information The online version contains supplementary material available at 10.1007/s00592-022-01942-8.


Introduction
Mortality is approximately twice higher in diabetic than in non-diabetic individuals [1,2], mainly but not exclusively due to an excess risk of atherosclerotic cardiovascular disease (ASCVD) [3]. However, death rates have consistently decreased over time in people with diabetes [4], mainly due to reduction in ASCVD events such as myocardial infarction and stroke [5]. This is likely related to improved treatment of multiple ASCVD risk factors, including hyperglycemia, dyslipidemia, and hypertension [6]. However, the burden from ASCVD remains disproportionately high in diabetic patients, especially in those with type 2 diabetes, thus requiring a systematic estimation of ASCVD risk in order to initiate therapeutic strategies for risk reduction [7].
Several prediction tools have been developed to estimate the risk of ASCVD events and all-cause or ASCVD mortality; however, these algorithms, which are mainly based on ASCVD risk factors, have numerous limitations resulting in insufficient performance, especially in people with type 2 diabetes [8]. Thus, at variance with guidelines of other scientific societies [9,10], those of the European Society of Cardiology (ESC) [11,12] recommend the use of a newlydeveloped, non-validated risk stratification system based on the presence of ASCVD risk factors, target organ damage (TOD), and/or previous ASCVD event(s) instead of one of the existing ASCVD risk prediction tools. Specifically, the 2019 ESC guidelines [11] identified three categories according to the 10-year risk of death from ASCVD, i.e., moderate (< 5% risk), high (5-10% risk), and very high (> 10% risk). The highest risk category included patients with established ASCVD, those with severe TOD (including microangiopathy and left ventricular hypertrophy), and/or those with at least three additional ASCVD risk factors. In 2021, this classification has been revised, mainly by restricting the definition of very high-risk individuals to patients with ASCVD and/or severe TOD, which included only a broader definition of microvascular disease [12].
Here, we analyzed the large cohort of patients with type 2 diabetes from the Renal Insufficiency And Cardiovascular Events (RIACE) Italian Multicenter Study to assess the distribution of a real-world sample into the 2019 and 2021 ESC risk categories and the risk of all-cause mortality in each of these categories.

Design
The RIACE Italian Multicenter Study is an observational, prospective, cohort study on the impact of estimated glomerular filtration rate (eGFR) on morbidity and mortality in people with type 2 diabetes [13].

Patients
The study population included 15,773 Caucasian patients (after excluding 160 individuals with missing or implausible values), consecutively attending 19 hospital-based, tertiary referral outpatient Diabetes Clinics of the National Health Service throughout Italy in the years 2006-2008. Exclusion criteria were dialysis or renal transplantation [13].

All-cause mortality
The vital status of study participants on 31 October 2015 was verified by interrogating the Italian Health Card database (http:// siste mats1. sanita. finan ze. it/ wps/ portal/), which provides updated and reliable information on all current Italian residents [14].

Baseline measurements
Baseline data were collected using a standardized protocol across participating centers [13].
Participants underwent a structured interview in order to collect the following information: age at the time of the interview, smoking status (never, former, current), known diabetes duration, severe co-morbidities (including chronic obstructive pulmonary disease, chronic liver disease, and cancer), and current glucose-, lipid-, and blood pressure (BP)-lowering treatments.
Body mass index (BMI) was calculated from weight and height, whereas waist circumference was estimated from log-transformed BMI values; BP was measured with a sphygmomanometer with the patients seated with the arm at the heart level.
Hemoglobin A 1c (HbA 1c ) was measured by HPLC using DCCT-aligned methods; triglycerides and total and HDL cholesterol were determined in fasting blood samples by colorimetric enzymatic methods; LDL cholesterol concentration was estimated using the Friedewald formula.
The presence of diabetic kidney disease (DKD) was assessed by measuring albuminuria and serum creatinine, as previously detailed [13,15]. Albumin excretion rate was obtained from 24-h urine collections or calculated from albumin-to-creatinine ratio in early morning, first voided urine samples; albumin concentration in urines was measured by immunonephelometry or immunoturbidimetry, in the absence of interfering clinical conditions. Serum (and urine) creatinine was measured by the modified Jaffe method, traceable to IDMS, and eGFR was calculated by the Chronic Kidney Disease Epidemiology Collaboration equation. Patients were then assigned to one of the following DKD phenotypes: no DKD, albuminuria alone (albuminuric DKD with preserved eGFR), reduced eGFR alone (non-albuminuric DKD), or both albuminuria and reduced eGFR (albuminuric DKD with reduced eGFR), as previously reported [16].
The presence of diabetic retinopathy (DR) was assessed in each center by an expert ophthalmologist by dilated fundoscopy [17]. Patients with mild or moderate non-proliferative DR were classified as having non-advanced DR, whereas those with severe non-proliferative DR, proliferative DR, or maculopathy were grouped into the advanced, sight-threatening DR category. DR grade was assigned based on the worse eye.
Previous major acute ASCVD events, including myocardial infarction; stroke; foot ulcer/gangrene/amputation; and coronary, carotid, and lower limb revascularization, were adjudicated based on hospital discharge records by an ad hoc committee in each center [18].

Statistical analysis
For the purpose of the current analysis, the RIACE participants were classified according to the 2019 or the 2021 ESC risk categories, with minor modifications due to the lack of information on left ventricular hypertrophy and neuropathy, respectively (Supplemental Table 1).
Data are expressed as mean ± SD for continuous variables, and number of cases and percentage for categorical variables. Comparisons among categories were performed by one-way ANOVA for continuous variables and Pearson's Chi-square test for categorical variables.
Crude mortality rates were described as events per 1000 patient-years, with 95% exact Poisson confidence intervals (CIs) and adjusted for age and sex by a Poisson regression model. Kaplan-Meier survival probabilities for all-cause mortality were estimated according to the 2019 or 2021 risk categories, and differences were analyzed using the log-rank statistic. The hazard ratios (HRs) and their 95% CIs were estimated by Cox proportional hazards regression, using the moderate risk category as reference. These analyses were sequentially adjusted for (a) parameters that were not used for patients' categorization, i.e., sex and severe comorbidities, including chronic obstructive pulmonary disease, chronic liver disease, and cancer (model 1); (b) parameters that were used only for defining the moderate risk category in the 2021 classification, i.e., glycemic control (HbA 1c and anti-hyperglycemic treatment) (model 2); and (c) additional ASCVD risk factors that were used for patients' categorization, i.e., age, smoking, diabetes duration, BMI, triglycerides, total and HDL cholesterol, systolic and diastolic BP, and lipid-lowering and anti-hypertensive treatment (model 3).
All p values were two-sided, and a p < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS version 13.0 (SPSS Inc., Chicago, IL, USA).

Results
The distribution and clinical features of the RIACE participants according to the 2019 and 2021 ESC risk categories and subcategories are shown in Tables 1 and 2, respectively.
Based on the 2019 classification, only 60 patients (0.4%) were assigned to the moderate-risk category, whereas approximately one third fell in the high-risk category and two thirds fell in the very high-risk category. Of the 63.8% participants assigned to the very high-risk category, 31.8% had ≥ 3 additional ASCVD risk factors; 8.9% had TOD, 3.8% with < 3 and 5.1% with ≥ 3 additional ASCVD risk factors; and 23.1% had previous ASCVD event(s), 8.5% with < 3 and 14.6% with ≥ 3 additional ASCVD risk factors, and 18.0% without and 5.1% with TOD.
Based on the 2021 classification, the number of patients assigned to the moderate-risk category increased but remained negligible (126, 0.8%), whereas the proportion of those falling in the high-risk and very high-risk categories was inverted (approximately two thirds and one third, respectively), compared with the 2019 classification. Of the 32.6% participants assigned to the very high-risk category, 9.5% had TOD only, 16.7% had previous ASCVD only, and 6.5% had both.
Differences between the two classifications (Supplemental Table 2) were due to reallocation of patients with ≥ 3 additional ASCVD risk factors to the high-risk category and a broader definition of TOD, which increased the number of individuals with TOD from 2184 to 2495.

All-cause mortality according to the 2019 ESC risk categories and subcategories
Crude and sex-adjusted mortality rates increased significantly from the moderate-risk to the very high-risk category; when further adjusting for age, no differences were observed among categories and subcategories (Table 3).
Kaplan-Meier estimates (Fig. 1A) and unadjusted HRs (Table 4) showed the same trend, though mortality risk was not significantly different between the moderate-risk and high-risk categories. The HRs for mortality were not significantly affected when adjusting for sex, comorbidities, and glycemic control, but differences disappeared after further adjustment for additional ASCVD risk factors (Table 4). Within the very high-risk category, death rates (Table 3), Kaplan-Meier estimates (Fig. 1B), and unadjusted and adjusted HRs (Table 4) of patients with TOD were almost equal to those of patients with previous ASCVD event(s), the majority of whom had also TOD and much higher than those of individuals with ≥ 3 additional ASCVD risk factors.

All-cause mortality according to the 2021 ESC risk categories and subcategories
Crude and sex-adjusted mortality rates increased significantly from the moderate-risk to the very high-risk category; when further adjusting for age, differences persisted only between the high-risk and very high-risk categories (Table 3).
Kaplan-Meier estimates (Fig. 1C) and unadjusted HRs (Table 4) showed the same trend. The HRs for mortality were not significantly affected when adjusting for sex, comorbidities, and glycemic control, but risk remained significantly higher only in the very high-risk versus the moderate-risk category when accounting for additional ASCVD risk factors mortality (Table 4).
Within the very high-risk category, death rates (Table 3), Kaplan-Meier estimates (Fig. 1D), and unadjusted and adjusted HRs (Table 4) were much higher in the subcategories with previous ASCVD event(s) and TOD and, to a lesser extent, TOD only than in the subcategory with previous ASCVD event(s) only. Moreover, when further adjusting for additional ASCVD risk factors, the HRs for mortality remained significantly higher versus the moderate-risk category in patients with TOD only or previous ASCVD event(s) and TOD (Table 4).

Discussion
This analysis of patients with type 2 diabetes from the RIACE cohort showed that very few of these individuals fell in the moderate-risk category according to both ESC classifications. Moreover, the proportion of participants in the very high-risk risk category was approximately twice higher than that of participants in the high-risk category according to the 2019 classification. These findings are consistent with previous reports assessing the distribution of patients from diabetes outpatient clinics according to the 2019 ESC classification [19,20], in which the proportion of individuals at moderate risk was similar and that of those at very high risk was even higher than in the RIACE cohort. However, using the 2021 classification, the proportion of participants in the very high-risk category became about half of that of participants in the high-risk category, due to the reallocation of patients with at least three additional ASCVD risk factors to the high-risk category. The finding that these patients showed a much lower mortality risk than those with TOD and/or previous ASCVD event(s) and closer to those in the high-risk category as in the 2019 classification is consistent with the revised risk stratification system proposed in the 2021 guidelines. Finally, within the very high-risk category, the mortality risk of individuals with TOD only was almost similar to that of patients with previous ASCVD event(s), with and without TOD, according to the 2019 classification, and higher than that of patients with previous ASCVD event(s) only, according to the 2021 classification. These findings might reflect the strong, independent association of microangiopathy, especially DKD, with measures of macroangiopathy such as left ventricular hypertrophy [21] and coronary calcification [22]. It is in fact plausible that, in individuals with TOD due to DKD and/or DR but without established ASCVD, the overall risk of death results from both overt microvascular and subclinical macrovascular diseases, thus indicating the need for noninvasive assessment of subclinical ASCVD and treatment with drugs providing protection from both ASCVD and DKD [23].  Taken together, the results of this analysis provide important insights into the risk stratification of patients with type 2 diabetes, though comparison between the ESC classification system and the existing prediction tools is not feasible as the former allows only a broad categorization of patients according to the risk of death from ASCVD, instead of quantifying the predicted risk. As stated above, prediction algorithms have several pitfalls that limit their performance in these individuals [8]. One limitation is that many of them have been derived from general population samples and not established (or validated) in people with type 2 diabetes. The ESC guidelines do in fact discourage to apply those from the general population to patients with diabetes [11,12], though comparisons with diabetes-specific algorithms have not univocally shown that the latter ones perform better [8,24]. Another limitation is the time-period when they were developed, as some of them date back to several years ago and, hence, do not consider the impact of recent therapeutic advances on ASCVD risk. Moreover, they estimate risk of different outcomes, including all-cause or ASCVD mortality and ASCVD events, with most of them being specific for myocardial infarction and stroke without considering other events, such as heart failure and peripheral artery disease. More importantly, they are mainly based on ASCVD risk factors, as only few of them consider previous ASCVD events [25,26], thus being applicable also to patients with established ASCVD, and/or the presence of TOD, which was found to be associated with an extremely elevated mortality risk among the RIACE participants. In fact, many of the most used algorithms, either derived from the general population [9,[27][28][29][30] or people with type 2 diabetes [31,32], do not include measures of TOD. In addition, the remaining algorithms consider only measure(s) of kidney damage, i.e., serum creatinine or eGFR [26,33], albuminuria [34], or both [25,35]. Therefore, it is not surprising that the Risk Equations for Complications of Type 2 Diabetes (RECODe), a tool for predicting complications and death that includes previous ASCVD events, serum creatinine, and albumin:creatinine ratio, was found to perform better than six of the above algorithms [25]. In this regard, the ESC classification system appears to be superior as it considers also TODs other than DKD, such as retinopathy and neuropathy. However, other complications are not taken into account, such as non-alcoholic fatty liver disease, which has been shown to independently predict fatal and non-fatal ASCVD events [36], and particularly measures of subclinical atherosclerosis. These measures include (a) coronary, carotid, or lower limb artery stenosis, as assessed by computed tomography angiography or ultrasound, which the American Diabetes Association guidelines consider an index of high risk if higher than 50% [37]; (b) functional imaging; (c) ankle-brachial index; and (d) coronary artery calcium scoring. All these are considered as risk modifiers, particularly the latter [12], which is recommended for coronary risk assessment in asymptomatic adults at intermediate 10-year risk (10% to 20%) or low-to-intermediate 10-year risk (6% to 10%), with calcium score driving reclassification of these individuals to the low-risk or high-risk category [38].
However, the ESC stratification systems do not appear to reflect the wide range of ASCVD risk observed in people with type 2 diabetes [39], as attested by the negligible number of patients in the moderate-risk category with both ESC classifications and the assignment of the majority of participants to the very high-risk category with the 2019 classification. This is likely because they do not take into account the degree of ASCVD risk factor control, at variance with all the existing risk prediction tools. In fact, optimal treatment of ASCVD risk factors in individuals with type 2 diabetes was found to significantly reduce or even eliminate the excess risk of death and ASCVD events compared to non-diabetic controls [40,41]. As a consequence, the ESC stratification systems might overestimate mortality risk in a great number of people with type 2 diabetes and, as acknowledged by the ESC Scientific Document Group [12], may not be appropriate to accurately quantify risk differences. This interpretation is in keeping with a systematic review and network  meta-analysis of randomized controlled trials [42] and a clinical practice guideline based on it [43], which stratified adults with type 2 diabetes into five risk categories using the RECODe prediction model [44] and assigned those with ≤ 3 and > 3 additional ASCVD risk factors to the very low-and low-risk category, respectively. Strength of our study includes the large sample size, the assessment of a wide range of clinical parameters, and the completeness of baseline and follow-up data. However, there are several limitations. First, the lack of information on the causes of death did not allow detecting ASCVD deaths, to which the ESC classification systems specifically refer. Second, results may have been affected by the lack of information on left ventricular hypertrophy and diabetic neuropathy, which were considered for defining TOD in the 2019 and 2021 classification, respectively [11,12]. Third, the study findings may not be applicable to the general ambulatory population, as only part of the individuals with type 2 diabetes attend Diabetes Clinics in Italy. Finally, the observational design makes causal interpretation impossible.
In conclusion, risk stratification of patients with type 2 diabetes from the RIACE cohort showed that only a few of them fell in the moderate-risk category according to both ESC classifications and that the majority of participants were assigned to the very high-risk category according to 2019 classification, due to the inclusion of those with at least three additional ASCVD risk factors. This suggests that the ESC stratification systems might overestimate mortality risk in patients with type 2 diabetes without TOD and ASCVD because they do not take into account the degree of ASCVD risk factor control. Reallocating individuals with at least three additional ASCVD risk factors to the highrisk category as in the 2021 classification was consistent with the observed all-cause mortality data. Mortality risk increased across categories according to both classifications, but differences among categories were more evident using the 2021 stratification criteria. Within the very high-risk category, risk of death was found to be particularly high in the presence TOD (namely microangiopathy), irrespective of established ASCVD, possibly due to the association with subclinical vascular disease.
were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.