Introduction

In this era of modern medicine, nearly three hundred and fifty million cases of diabetes have been diagnosed to date worldwide [1]. Mortality is a great concern among this patient population, especially in those with type 2 diabetes mellitus (T2DM) in co-existence with cardiovascular disease (CVD). Values calculated by the World Health Organization show that there are approximately three million deaths annually due to T2DM and its related complications [2]. Cardiovascular mortality, which was evaluated in the Second Cardiovascular Outcome Trial Summit of the Diabetes and Cardiovascular Disease (D&CVD) EASD Study Group [3], is also a major health risk factor in the population of patients with T2DM.

T2DM is an independent cause of mortality. Although data from the Korean National Health Insurance Service–National Sample Cohort showed that 78% of diabetes-related deaths could not be ascribed to diabetes [4], other studies have shown that in T2DM patients with CVD who were re-vascularized by percutaneous coronary intervention (PCI), insulin therapy significantly contributed to a high death rate [5].

We have therefore conducted a meta-analysis to compare cardiac- versus non-cardiac-related mortality following PCI in a sample of patients with insulin-treated type 2 diabetes mellitus (ITDM). To date, few studies have systematically assessed cardiac versus non-cardiac mortality in such patients following PCI.

Methods

Searched Databases and Key Words/Index Terms

We systematically and thoroughly searched the MEDLINE/PubMed, EMBASE, Cochrane library, and Google Scholar databases. the following key words/index terms were used to identify articles of possible interest:

  • T2DM + PCI, or

  • T2DM + coronary angioplasty, or

  • T2DM + PCI, or

  • ITDM + PCI, or

  • Cardiac death + ITDM + PCI.

Inclusion and Exclusion Criteria

Studies were included in the meta-analysis review if: (1) they were trials or cohort studies based on patients with T2DM following PCI; (2) they separately reported outcomes in patients with ITDM; (3) they reported cardiac death and non-cardiac death among their endpoints.

Studies were excluded from the systematic review if: (1) they did not involve patients with T2DM following PCI; (2) they did not separately report patients with ITDM; (3) they did not report cardiac death among their clinical outcomes; (4) they were repeated studies or duplicate studies.

Participants

All participants in the studies ultimately included in our systematic review were patients with ITDM. However, the extent of the coronary artery disease varied among studies and included ITDM patients with stable coronary artery disease, de novo coronary artery disease, acute myocardial infarction, single-vessel coronary artery disease, multi-vessel coronary artery disease (Table 1).

Table 1 Participants with insulin-treated type 2 diabetes mellitus and coronary artery disease participating in the studies included in the systematic review
Full size table

Definition of Endpoints

In this analysis, cardiac death was compared with non-cardiac death in diabetic patients who received insulin treatment. Therefore, the main focus was on: (1) cardiac mortality: death due to cardiac causes; (2) non-cardiac mortality: death which was not related to cardiac causes.

Data Extraction and Review

The following data were extracted by three reviewers independently of each other: (1) patients with ITDM; (2) the number of events corresponding to cardiac death; (3) the number of events corresponding to non-cardiac death; (4) baseline features; (5) methodological features of each study; (5) type of study.

The methodological qualities for the randomized controlled trials were assessed by using the guidelines set down in the Cochrane Handbook for Systematic Reviews of Interventions [6]. The Newcastle–Ottawa Scale (NOS) [7] was used to assess the methodological qualities for the observational studies.

Statistical Analysis

The computer program RevMan version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) was used as analytical software. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Heterogeneity was assessed by two meta-analytical methods: (1) The Cochrane Q statistic test (a P value of ≤ 0.05  indicates a statistically significant result); (2) the I2 statistic test (the lower the value, the lower the heterogeneity).

A fixed (I2 < 50%) effects model or a random (I2 > 50%) model was used based on the value of I2 that was obtained.

Compliance with Ethics Guidelines

This meta-analysis is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.

Results

Searched Outcomes

We followed the PRISMA guideline for this analysis [8].

A total of 1368 publications were identified from the database search using the chosen key words/index terms. Publications were excluded and eliminated based on the following criteria:

  • They were not related to the aim of this meta-analysis (n = 1282).

  • They did not report cardiovascular death, but instead reported total death among their clinical outcomes (n = 12).

  • They were meta-analyses or review articles themselves (n = 5).

  • They did not separately report patients with ITDM (n = 21).

  • They were duplicates of the same study (n = 36).

Ultimately, a total number of 12 articles (6 randomized controlled trials [RCTs] and 6 observational cohorts) [9,10,11,12,13,14,15,16,17,18,19,20] were included in this meta-analysis, as shown in Fig. 1.

Fig. 1
figure 1

Flow diagram of study selection. ITDM Insulin-treated type 2 diabetes mellitus

Full size image

General and Baseline Features of the Participants

A total number of 4072 patients with ITDM who participated in 12 observational studies/RCTs were included in this meta-analysis. Of these, 1658 participants were extracted from RCTs and 2414 participants were extracted from observational studies. Two studies had a follow-up period of < 1 year, four studies had a follow-up period of 1 year, and six studies had a follow-up period of > 1 [range 3–5] year). One study reported patients who were treated with a bare metal stent, whereas all of the other studies involved patients who were treated with drug-eluting stents DES, specifically everolimus-eluting stents (EES), paclitaxel-eluting stents (PES), sirolimus-eluting stents (SES), and zotarolimus-eluting stents (ZES), as shown in Table 2.

Table 2 Total number of events and other features of the studies included in the meta-analysis
Full size table

The baseline features of the participants with ITDM are given in Table 3. Based on the features which are listed, there was no significant difference between those patients who died due to a cardiac cause and those who died due to a non-cardiac cause.

Table 3 Baseline features of the participants
Full size table

The methodological qualities of the studies were also assessed. RCTs were assessed with the recommended features of the Cochrane collaboration guidelines [6]. Grades were given to define the limit of bias (low, low to moderate, moderate, and high). For the observational studies, NOS scores [7] were given, with a maximum number of nine points (a higher score indicates better quality study), as shown in Table 4.

Table 4 Assessment of bias risk
Full size table

Death Due to Cardiac Versus Non-cardiac Causes Following PCI in Patients with ITDM

Analysis of the combined data extracted from the included RCTs and observational studies revealed that death due to cardiac causes was significantly higher in patients with ITDM than death due to non-cardiac causes(OR 2.16, 95% CI 1.79–2.59; P = 0.00001; I2 = 0%) when (Fig. 2).

Fig. 2
figure 2

Cardiac versus non-cardiac death following percutaneous coronary intervention (PCI) in patients with ITDM. CI Confidence interval, M–H Mantel–Haenszel test

Full size image

However, data from RCTs and observational studies were also analyzed separately. When we considered only data obtained from RCTs in the analysis, death from cardiac causes was still significantly higher in patients with ITDM (OR 2.20, 95% CI 1.54–3.14; P = 0.0001; I2 = 14%) (Fig. 3). Similarly, when we considered only data obtained from observational cohorts, death due to cardiac causes was significantly higher in the ITDM patients (OR 2.14, 95% CI 1.73–2.66; P = 0.00001; I2 = 0%) (Fig. 4).

Fig. 3
figure 3

Cardiac versus non-cardiac death following PCI in patients with ITDM based only on data obtained from randomized controlled trials

Full size image
Fig. 4
figure 4

Cardiac versus non-cardiac death following PCI in ITDM based only on data obtained from observational cohorts

Full size image

When all the studies with a follow-up period of 1 year were analyzed together, cardiac death was still significantly higher in patients with ITDM compared to non-cardiac death (OR 2.39, 95% CI 1.47–3.88; P = 0.0004; I2 = 0%) (Fig. 5).

Fig. 5
figure 5

Cardiac versus non-cardiac death following PCI in ITDM during a follow-up period of 1 year

Full size image

When studies with longer follow-up periods were considered (range 3–5 years), mortality due to cardiac causes was still significantly higher in these patients with ITDM (OR 2.09, 95% CI 1.70–2.56; P = 0.00001; I2 = 15%) (Fig. 6).

Fig. 6
figure 6

Cardiac versus non-cardiac death following PCI in ITDM during a longer follow-up period (range 3–5 years)

Full size image

When the participants were analyzed based on the type of stents, death due to cardiac causes was still significantly higher in those patients having an EES (OR 2.31, 95% CI 1.26–4.26; P = 0.007, I2 = 0%) compared to those a PES (OR 2.36, 95% CI 1.63–3.39; P = 0.00001; I2 = 0%), SES (OR 2.11, 95% CI 1.67–2.67; P = 0.00001; I2 = 21%), or ZES (OR 2.12, 95% CI 1.11–4.05; P = 0.02), as shown in Fig. 7.

Fig. 7
figure 7

Cardiac versus non-cardiac death following PCI in ITDM according to drug-eluting stents

Full size image

Discussion

Cardiovascular death is a major concern among patients with T2DM who are treated by PCI. The results of our meta-analysis show that cardiovascular death in patients with ITDM who have undergone PCI was significantly higher that death due to non-cardiac causes. This result remained consistent even when data from the RCTs and observational studies were analyzed separately.

To assess the effect of differences in follow-up periods on the cause of death in this patient group, we also separately analyzed the data from all of the studies included in the meta-analysis which reported a follow-up period of 1 year. As reported in the "Results" section, cardiac death was still significantly higher in this subpopulation of patients with ITDM. When a longer follow-up period (3–5 years) was considered, the major cause of death remained cardiovascular.

We also assessed the impact of coronary stents on the results by analyzed the data from all studies based on the types of coronary stents which were implanted (EES, PES, SES, and ZES). However, cardiovascular cause of death was still significantly higher in the patients with ITDM.

An 11-year retrospective analysis of death certificates in Shanghai also showed an increasing occurrence of CVD among Chinese patients who had previously developed diabetes mellitus [21], with 29.9% of deaths among those diabetic patients due to cardiovascular causes; in comparison, other causes represented only small percentages. However, causes of death based specifically on patients with ITDM were not analyzed in that study.

Finally, even though research has shown diabetes mellitus to be independently associated with death due to CVD, other studies have shown that insulin therapy also makes a major contribution to such an outcome [5, 22]. In addition, female gender and higher co-morbidities have also been suggested to further contribute to such outcomes [23]. These factors should further be investigated in future studies.

There are a few limitations to our analysis. First, the total number of patients was relatively small, especially for the analysis on impact of types of coronary stents on cause of death in patients with ITDM treated by PCI. Second, data from different categories of participants (those with stable coronary artery disease, multi-vessel coronary artery disease, single-vessel coronary artery disease, left main coronary artery disease, and acute myocardial infarction) were combined and analyzed. Third, the anti-platelet agents which were used post-PCI were not taken into consideration and this might also have had an impact on the mortality rate.

Conclusions

In patients with ITDM, mortality due to cardiac causes was significantly higher than that due to non-cardiac causes following PCI. The same conclusion was reached when different lengths of follow-up periods were assessed, when different data sets were used (total data set, data from the observational cohort or RCTs separately), and when types of coronary stents which were implanted were assessed.