Introduction

Patients with coronary artery disease (CAD) who also have diabetes mellitus (DM) experience more cardiovascular events compared to non-diabetic patients1,2. Many studies have demonstrated a higher prevalence of DM in patients who underwent percutaneous coronary intervention3,4,5 (PCI) and those with DM were at a heightened risk of cardiovascular events6,7, such as restenosis8, stent thrombosis9, and higher rates of revascularization10. Furthermore, several researches have indicated that about 16–19% of patients undergoing PCI had newly diagnosed DM4,5,11. And newly diagnosed DM is shown to be associated with increased short-term and one-year major adverse cardiovascular and cerebral events (MACCE)5,12. However, there were limited data regarding the impact of newly diagnosed DM up to five years outcomes in elderly patients (age ≥ 65) who underwent PCI.

Thus, the objective of this study was to assess the prognostic implications of newly diagnosed DM in elderly patients undergoing PCI after a median follow-up of 5 years, in comparison to individuals with known DM or without DM.

Methods

Study population

Since 2004, all patients who underwent invasive PCI at Korea University Guro Hospital (KUGH) in Seoul, South Korea, were prospectively enrolled in the KUGH-PCI registry. It is a prospective study of all comers conducted at a single center that reflects ‘real world’ practice. All patients’ demographic data, clinical characteristics, description of PCI, medications on discharge, and blood tests including admission blood glucose, and glycated hemoglobin A1c (HbA1c) level were collected. The study exclusion criterion was defined as the exclusion of patients aged < 65 years, PCI procedural failure, no-drug-eluting stents (DES) or absent HbA1c level. For clinical diagnoses and clinical events, standard definitions were used for all variables related to patients and lesions. For all the participants or their legal guardians, a thorough literal and verbal explanation of study procedures was provided before their written informed consents were obtained. The Institutional Review Board of KUGH approved this study (#KUGH10045). All methods were performed in accordance with relevant guidelines and regulations.

Study definitions

In this study, newly diagnosed DM was defined as admission HbA1c ≥ 6.5%, without known DM history or current use of hypoglycemic agents. Multivessel disease was defined as at least two coronary vessels or left main coronary disease with more than 50% luminal narrowing. Diffuse long lesions referred to lesion length > 30 mm. Chronic total occlusion (CTO) was an obstruction of a native coronary artery that lasts for more than three months. Target lesion revascularisation (TLR) referred to the repeat revascularization in the previously implanted stent or 5 mm proximal or distal to the stented segment. Target vessel revascularisation (TVR) referred to the intervention in the same treated vessel. And non-TVR referred to a revascularization of any segment of the non-target coronary artery.

The primary clinical outcomes of interest were major adverse cardiovascular events (MACE) defined as the composite of cardiac death, non-fatal myocardial infarction (MI), and any repeat revascularization (including TLR, TVR, and non-TVR) among the three groups. Secondary outcomes were the individual components of MACE, all-cause mortality, TLR, TVR, non-TVR, stent thrombosis and stroke.

Statistical analysis

For continuous variables, data were expressed as mean ± standard deviations, or median (IQR) for variables that are not normally distributed. Differences among the three groups were evaluated by one-way ANOVA or Kruskal Wallis test and a post hoc test was applied: the Tukey’s Honestly Significant Difference test or Steel–Dwass test. For categorical variables, differences were expressed as counts and percentages and analyzed with χ2 or Fisher’s exact test between the groups as appropriate. Random forest method was used to impute missing values (using the randomForest package in the R language).

To adjust for potential confounders, univariable Cox proportional hazards model analysis was performed. We tested available 20 baseline clinical and angiographic variables that could be of relevance: sex, age, ST-segment elevation myocardial infarction (STEMI), cardiogenic shock, previous myocardial infarction (MI), previous PCI, previous coronary artery bypass grafting (CABG), hypertension, chronic kidney disease (CKD), atrial fibrillation (AF), smoking, multivessel disease, left main coronary lesion, bifurcation lesion, ca1cification lesion, diffuse long lesion, CTO and number of stents, total stent length. All variables with P < 0.05 or considered of clinical implications were included for further multivariable Cox analysis. A nomogram was constructed to visualize the risk factors for clinical outcomes with best discrimination and calibration. A two-tailed p-value of < 0.05 was considered to be statistically significant. The Bonferroni correction was used to adjust the P values for pairwise comparisons. Various 5-year clinical outcomes were estimated with Kaplan–Meier method, and differences between groups were compared with log-rank test. Data processing and analysis were performed using R version 4.3.0 (2023–04–21), along with Storm Statistical Platform (www.medsta.cn/software).

Results

Population

A total of 6972 patients who underwent PCI were enrolled during January 2004 and June 2021 (Fig. 1). Of these, 3529 were elderly patients (≥ 65 years old), 287 (8.1%) patients were newly diagnosed with DM, 1127 (32.0%) patients had previously known DM, and 2115 (59.9%) patients did not have DM. The median follow-up was 1806 days (5.0 years, median 825–2740 days [2.3–7.6 years]). Eventually, patients aged < 65 years (n = 3443), PCI procedural failure or no-DES (n = 149), absent HbA1c level (n = 924) were excluded (Fig. 1).

Figure 1
figure 1

Flowchart of the study population and distribution of the population. DES = drug-eluting stents, HbA1c = hemoglobin A1c, KUGH = Korea University Guro Hospital, PCI = percutaneous coronary intervention.

Full size image

Considering the potential impact of stent generation and PCI procedural failure on clinical outcomes, a total of 2456 patients with DES, successful PCI revascularization and available record of blood glucose and HbA1c level were included in the final analysis. In this study population, 269 (10.9%) patients were newly diagnosed DM, 820 (33.4%) patients had previously known DM, and 1367 (55.7%) patients were non-DM. The median follow-up was 1975 days (5.4 years, median 882–2925 days [2.4–8.0 years]).

Baseline clinical and angiographic characteristics

The baseline clinical characteristics of patients categorized by glycemic status are detailed in Table 1. Newly diagnosed diabetic patients were younger (median age, 71 vs. 73 years old, p = 0.026), had higher rates of previous MI (2.23% vs. 0.24%, p = 0.005), previous PCI (4.46% vs. 0.00%, p < 0.001), previous CABG (1.86% vs. 0.24%, p = 0.007) and lower rates of hypertension (71.38% vs. 79.27%, p < 0.001), AF (3.35% vs. 7.07%, p = 0.004) and CKD (5.95% vs. 10.98%, p < 0.001) compared to those with previously known DM. Higher levels of HbA1c (median level, 7.1% vs. 7.0%, p < 0.001), total cholesterol (median level, 4.29 vs. 4.01 mmol/l, p < 0.001), total triglycerides (median level, 1.30 vs. 1.19 mmol/l, p < 0.001) and low-density lipoprotein cholesterol (median level, 2.79 mmol/l vs. 2.30 mmol/l, p < 0.001) were more common in newly diagnosed diabetes group and non-diabetes group compared to previously known diabetes group.

Table 1 Clinical characteristics and five-year outcomes of patients and comparison among groups.
Full size table

The angiographic characteristics and stent information across three groups are displayed in Table 1. Newly diagnosed DM and previously known DM patients had similar prevalence of multi-vessel disease (29.74% vs. 34.15%, p = 0.363), left main coronary disease (4.46% vs. 5.12%, p = 0.900), bifurcation lesions (34.94% vs. 27.69%, p = 0.067), ca1cification lesion (22.30% vs. 21.38%, p = 0.984), diffuse long lesions (54.65% vs. 48.54%, p = 0.185), and CTO (5.95% vs. 8.90%, p = 0.250). Thus, newly diagnosed DM and previously known DM patients had similar number of stents (1.69 ± 0.97 vs. 1.85 ± 1.11, p = 0.073) and total stent length (42.79 ± 27.10 mm vs. 43.70 ± 30.72 mm, p = 0.430). Over 98% of the patients was prescribed aspirin and clopidogrel after discharge. And there was no significant disparity between the utilization of antiplatelet therapy in all three groups (p > 0.100).

Five-year clinical outcomes

The newly diagnosed and previously known DM groups exhibited higher rates of total death (13.38% vs. 8.05%, p = 0.011; 12.32% vs. 8.05%, p = 0.004), cardiac death (7.81% vs. 3.88%, p = 0.022; 6.22% vs. 3.88%, p = 0.025) and MACE (25.28% vs. 15.44%, p < 0.001; 19.15% vs. 15.44%, p = 0.039) compared to the non-diabetic group (Table 1). Newly diagnosed DM seemed to have higher risk of MACE (25.28% vs. 19.15%, p = 0.039) and a trend towards higher MI (5.95% vs. 2.68%, p = 0.058) than previously known DM. However, there were no statistically significant differences in terms of TLR, TVR, non-TVR, and stent thrombosis across the three groups.

Kaplan–Meier survival analysis demonstrated that the 5-year cumulative incidences of MACE were significantly higher in the newly diagnosed diabetes and previously known diabetes, compared to non-diabetic group (pairwise log-rank p < 0.001; pairwise log-rank p = 0.003; Fig. 2D), but the difference between newly diagnosed and previously known diabetes was not significant (pairwise log-rank p = 0.135). The components of MACE in different glycemic status are shown in Fig. 2A–C. The 5-year cumulative incidences of cardiac death and repeat revascularization were similar as MACE. Only 5-year cumulative incidence of no-fatal MI in newly diagnosed diabetes was significantly higher than that in previously known diabetes (pairwise log-rank p = 0.031).

Figure 2
figure 2

Kaplan–Meier survival curves for five-year cardiac death (A), no-fatal myocardial infarction (B), repeat revascularization (C) and major adverse cardiovascular events (D). Newly detected diabetes (blue), previously known diabetes (green), and non-diabetes (red).

Full size image

Multivariable Cox regression model

The detailed results could be found in Table 2 and more univariate Cox analysis information in Supplementary Material. Multivariate Cox regression showed that newly diagnosed diabetes was an independent predictor of MACE [hazard ratio (HR) 1.64, 95% confidence internal (CI) 1.24–2.17, p < 0.001], cardiac death (HR 2.15, 95% CI 1.29–3.59, p = 0.003) and repeat revascularization (HR 1.52, 95% CI 1.09–2.11, p = 0.013), except for the non-fatal MI (HR 1.66, 95% CI 0.94–2.12, p = 0.081). While previously known DM was only independently associated with five-year MACE (HR 1.24, 95% CI 1.01–1.54, p = 0.044). Besides, cardiogenic shock was the strongest predictor of MACE (HR 4.54, 95% CI 3.28–6.30, p < 0.001), cardiac death (HR 9.85, 95%CI 6.41–15.14, p < 0.001), non-fatal MI (HR 3.66, 95% CI 1.68–7.97, p = 0.001) and repeat revascularization (HR 1.88, 95%CI 1.07–3.29, p = 0.027). Additionally, CKD remained a significant predictor of MACE (HR 1.71, 95% CI 1.23–2.37, p = 0.001), cardiac death (HR 3.24, 95% CI 1.99–5.28, p < 0.001), and non-fatal MI (HR 2.24, 95% CI 1.11–4.51, p = 0.024).

Table 2 Independent predictors of five-year clinical outcomes after multivariable Cox adjustment for baseline clinical and angiographic characteristics.
Full size table

To assess the performance of the models described above, concordance index (C-index) was calculated, and calibration curve was determined for the best regression model. The multivariable Cox regression model for five-year cardiac death had the best discrimination with C-index of 0.796 (95% CI 0.756–0.836). The subsequent nomogram and calibration curve are showed in Fig. 3.

Figure 3
figure 3

The figure (A) is a nomogram predicting the risk of five-year cardiac death in elderly patients underwent PCI. A score was given to the value of each variable on the point scale axis. By adding each score, a total score could be calculated and projected to the lower total score scale, so we could estimate five-year cardiac-death free survival probability. The figure (B) is the calibration curves for the nomogram. The x-axis represents the nomogram-predicted probability and y-axis represents the actual probability of cardiac death. Perfect prediction would correspond to the 45° gray dashed line. The red dotted line represents the entire cohort (n = 2456), and the blue solid line is bias-corrected by bootstrapping (B = 500 repetitions), indicating observed nomogram performance. CKD = chronic kidney disease, STEMI = ST-segment elevation myocardial infarction.

Full size image

Discussion

In this study, newly diagnosed DM was presented in approximately 8.1–10.9% of elderly patients who underwent PCI. Moreover, newly diagnosed DM was independently associated with increased risk of five-year MACE, and was a stronger risk factor compared with previously known DM.

Several studies have analysed the prevalence of newly diagnosed DM in specific CAD patients undergoing PCI and subsequent clinical outcomes. De la Hera et al.4 reported 16.2% of studied patients (stable angina or NSTEMI, n = 580) who underwent PCI was newly detected diabetes, based on the oral glucose tolerance teat (OGTT) after discharge. And newly diagnosed DM was not an independent predictor of 1-year outcomes. Tailakh et al.5 demonstrated that 19% of total 1151 CAD patients had newly diagnosed DM based on the HbA1c level. Newly diagnosed DM was independently associated with increased MACCE at one-month but not at one-year. Subgroup analysis showed male and patients younger than 75 years appeared higher one-year MACCE risk. Jimenez-Navarro et al.13 discovered that 21.4% of 374 CAD patients who underwent PCI were newly detected diabetes and previously known diabetes remained the only independent predictor of cardiovascular events in a mean follow-up of 35.8 ± 13.43 months. Tsuchida et al.14 reported 17% of the 298 patients studied had newly diagnosed DM, which was not a predictor of up to 10-year cardiovascular disease risk. Other studies focused on only STEMI patients15,16. Therefore, newly diagnosed DM seemed to have greater short-term impact and less long-term impact according to the above studies. But there is limited data to evaluate the impact of newly diagnosed DM in elderly patients.

As far as we know, our study focused specifically on elderly population and for the first time, investigating whether or not newly diagnosed DM impacted on five-year clinical outcomes in elderly patients undergoing PCI. The prevalence of newly diagnosed DM in elderly patients undergoing PCI was similar to that in prior studies across all age groups. All patients with DM, no matter newly or previously diagnosed, had more cardiovascular events than patients without DM, which has been investigated in numerous studies12,17,18,19. Our results are in agreement with previous findings, newly diagnosed diabetic patients were younger and had less comorbidities5,12. Furthermore, we reported more detailed information about angiographic characteristics of newly diagnosed diabetes than former studies. Despite having similar baseline to patients without DM, patients with newly diagnosed diabetes had comparable complicated coronary lesions with previous DM patients. Notably, newly diagnosed diabetic patients had more history of previous MI, PCI and CABG, along with higher HbA1c level, which may account for its poor prognosis5,19,20,21. A possible reason is that those patients with newly diagnosed DM may have been unaware of asymptomatic and uncontrolled DM for a long time. A recent study22 has shown similar impact of DM on increased risk of 5-year cardiovascular outcomes after PCI.

In the present study, we focused on elderly patients with newly diagnosed diabetes, a distinctive cohort that warrants special attention. It might be assumed that newly diagnosed diabetes has a lesser impact on elderly individuals. However, it is essential to recognize that older adults often face a unique set of challenges, including a higher risk of physical and mental frailty 23,24, a vulnerable immune system25, an increased burden of diabetes-related complications26, poorer treatment adherence27, and a potentially more conservative approach to blood glucose control due to concerns about hypoglycemia28. As a result, these factors may collectively contribute to the elevated incidence of adverse outcomes in the elderly population, which underscores the importance of our study's focus on this particular demographic. By acknowledging these factors and their implications, we can better appreciate why there is a need for a more targeted and nuanced approach to managing diabetes in elderly patients.

There are some limitations should be considered. Firstly, this was a single-center study, it is important to evaluate the findings in a larger population involving multiple centers. Despite a monocenter study, we included more than two thousand elderly patients underwent PCI to preliminarily explore the long-term outcomes of newly diagnosed DM. Secondly, there was no information available to evaluate the status of diabetes controlled, the treatment modalities, changes on medication details and adherence, which may affect the outcomes of patients undergoing PCI. There should be more specific work on those elements in the future. Thirdly, due to the absence of available variables, a mount of cases were excluded from the analysis, which could have biased our findings. Besides, it is possible that multivariable models may contain incomplete adjustments or unknown confounders that have not been consolidated.

Conclusions

In conclusion, our study indicated that newly diagnosed DM was associated with an increased risk of five-year MACE compared with non-DM and previously diagnosed DM in elderly patients who underwent PCI. Our findings support the importance of identifying and treating diabetes at early stage for elderly CAD patients undergoing PCI. And it is necessary to monitor on the blood glucose, and perform HbA1c screening among high-risk elderly patients who undergoing PCI. It is hoped that more attention and care should be given to the older population.