In this study, we elucidated the long-term prognostic role of single and concomitant MVC in a cohort of middle-aged patients with long-standing T1D or T2D. Over a remarkably long follow-up period of 21 years, CKD, CAN, and DR, but not DPN, reduced life expectancy by 1.3 to 2.6 years, and CKD and CAN increased the adjusted all-cause mortality by 117% and 54%, respectively. Importantly, the accurate assessment of multiple concomitant MVC allowed us to establish the synergistic effect of their combination on all-cause mortality, which is of clinical relevance as diabetic MVC are frequently associated. The presence of both CKD and CAN markedly increased mortality risk by fivefold. Furthermore, their combination with DR was associated with an even greater reduction in survival and increased mortality risk, as compared with the presence of any isolated MVC. This novel evidence substantiates the clinical value of a comprehensive screening for isolated or combined MVC in patients with diabetes for better risk stratification.

Previous studies with indirect estimation of GFR (calculated eGFR) established the role of CKD on CVD-related and all-cause mortality risk, regardless of the presence of other complications, often over follow-up periods shorter than 5–10 years [17,18,19,20,21,22]. Although the estimation of GFR based on serum creatinine level is feasible and widely used in large-scale studies and clinical practice, numerous limitations are known in reflecting measured renal function. Indeed, eGFR may differ from mGFR by ± 30%, resulting in 30–60% misclassification of CKD stage [23]. Furthermore, creatinine-based eGFR formulae are adjusted for clinical factors that can directly affect mortality rate, including age, sex, and race, which would eventually lead to an over- or underestimation of the mortality risk associated with CKD. Our findings extend previous indirect evidence [17,18,19,20,21,22] confirming the prognostic value of GFR in diabetes using a direct measurement of GFR, based on 99Tc-DTPA renal dynamic scintigraphy, and exploiting a follow-up period of more than 20 years. This long observation period, approaching the life expectancy of a typical adult patient referred for complicated T2D, is apt to appreciate the full magnitude of the time-dependent, negative prognostic impact of MVC. This is particularly important for CKD and its components, whose effects on mortality risk become increasingly evident after 5 to 10 years of observation (Fig. 2A–C).

Diabetic neuropathy is a frequent but often overlooked MVC associated with CVD and mortality [24,25,26,27,28,29,30,31,32,33,34]. In a recent meta-analysis [35], CAN demonstrated a pooled relative risk of all-cause mortality of 3.17 [2.11–4.78], which was higher in T1D than T2D. In the 19 available studies identified, however, the follow-up length ranged widely (up to 21 years in DCCT/EDIC study participants with T1D [31]) and the number of abnormal autonomic function tests used to define CAN varied from 1 to 3, thus making results difficult to compare, especially for the T2D subgroup [35]. In our analyses, CAN was defined by applying newer diagnostic criteria [14], which reduce the risk of overestimating the incidence of CAN, and its negative prognostic role was reappraised in both T2D and T1D over an extended follow-up period. In contrast to previous observations, in our cohort CAN increased the mortality risk preferentially in T2D, even though patients with T1D showed similar CAN prevalence and severity. This finding should be interpreted with caution, as it may depend on several unaccounted factors, including different disease duration or therapeutic management in the two groups, and on the limited number of participants with T1D and CAN in our cohort. The same considerations, along with the low prevalence of complicated DPN (e.g. foot ulcers), may explain our negative result on the prognostic value of DPN [28, 32,33,34, 34].

It has been reported that individuals with DR have a greater incidence of CVD and lower survival rates than those without DR [36,37,38,39,40]. In the meta-analysis by Kramer et al. [39], the presence of DR was linked to a markedly increased risk for all-cause mortality in both T2D (OR 2.41 [1.87–3.10]) and T1D (OR 3.65 [1.05–12.66]). Our findings are partially in agreement with these results, as we observed an increase in the relative mortality risk associated with DR, which, however, lost statistical significance after accounting for potential confounders. This observation is consistent with two recent studies in large cohorts of patients with T2D [41] and T1D [42] showing that DR has a neutral effect on mortality independently of known confounders. Nonetheless, our negative findings can derive as well from peculiar participants’ characteristics and limited statistical power to assess the independent effect of DR in our cohort.

Multiple diseases with a common pathophysiology may compete for determining negative health outcomes, resulting in less-than-additive effects. In contrast, sparse evidence indicates that having two or more concomitant MVC may result in an additive or even synergistic effect on incident CVD and survival rates, compared with having a single isolated MVC, particularly in T1D [42,43,44,45]. In fact, CAN [30] and DR [46] were independent predictors of mortality only in T1D patients with CKD, but not in those without CKD, in two 5- to 10-year longitudinal studies. Consistently, we observed a steep increase in mortality rates when CAN and/or DR were associated with CKD, extending previous observations to patients with both T1D and T2D on a longer follow up.

Study participants presented with poor metabolic control, high prevalence of MCV (especially DR), and were naïve from any novel disease-modifier treatment. In fact, therapeutic targets of traditional risk factors in the late nineties were not as strict as today, and referral of patients with diabetes to a tertiary level clinic was often delayed. Also, the two classes of drugs with glucose-independent protective effects against diabetic MVC, namely glucagon-like peptide 1 (GLP-1) receptor agonists [47] and sodium-glucose co-transporter 2 (SGLT-2) inhibitors [48], were approved in T2D in 2005 and 2013, respectively, and became widely used only later. This partly justifies the high mortality rate in our cohort with respect to current life expectancy, being in line with studies performed in the same years [49]. It also underscores the significant ameliorations in quality of life and survival secondary to the recent advances in diabetes research and clinical practice, confirming that an early and vigorous control of glycemia and cardiovascular risk factors can deeply affect survival in diabetes, which would be otherwise poor when MVC arise.

There are some limitations to the present study, mostly related to its retrospective design. First, we could not retrieve reliable information on incident CVD events or the cause of death. Although MVC are stronger risk factors for CVD mortality than for non-CVD mortality [28], this precluded us to detail the prognostic role of each MVC on relevant clinical outcomes. Second, we could not evaluate the time course of MVC, clinical parameters including glycemic control, nor treatment changes after enrollment, which may influence study results. Thus, the prognostic role of the temporal association between different MVC and the potential effects of time-changing unaccounted factors should be assessed in future studies. Third, the number of subjects and events in some analyses was limited, especially when dealing with the T1D subgroup, which warrants caution in the interpretation of negative findings. Forth, the diagnosis of DR relied on a single operator-dependent assessment, which however was performed by a trained ophthalmologist to minimize the risk of misclassifications.

In conclusion, our study demonstrates the long-term, synergistic, negative effects of diabetes MVC on all-cause mortality in patients with both T2D and T1D. The substantial burden of MVC in diabetes fosters a comprehensive baseline screening to identify the subgroup of patients with higher mortality risk, who may benefit the most from targeted interventions to improve survival and quality of life.