The ABANDIA study is a large prospective, multicentre trial comparing the outcome after elective infrarenal AAA repair in patients with or without DM, with comprehensive DM screening at baseline. The total prevalence of DM was 15%. Twenty-five percent of the DM patients were undiagnosed at baseline. No statistically significant association with increased risk of complications, mortality, or longer length of stay in the perioperative period (≤ 30 days) was found in patients with DM. However, post-hoc analyses indicate that patients with undiagnosed DM are at increased risk of organ-related complications compared to non-DM patients.

Retrospective European studies report a DM prevalence in surgically treated AAA ranging from 12% in the EUROSTAR registry to 24% in France reported by Raffort et al [13, 15, 26, 27]. In Norway, the DM prevalence in the general population is reported to be 5–12% when based on HbA1c values [28,29,30]. The highest DM prevalence is found in males, those with obesity, and with advancing age. Hence, the results from this study are comparable to that of the age-matched general Norwegian population.

Globally, it is estimated that 45% of all DM cases are undiagnosed [31]. In Europe, the estimate is 20–30% [31]. In comparison, the proportion of undiagnosed DM cases in the general Norwegian population has been reduced substantially in the last few years and is now estimated to be at 11% compared to 25% in this study [29]. The national Norwegian guidelines’ recommendation of HbA1c screening in high-risk individuals may account for this lower proportion of undiagnosed DM in the general population. Although AAA patients often have cardiovascular comorbidities, the vascular surgery guidelines in the USA and in Europe do not address screening for DM in AAA patients [21, 22, 32]. Thus, the high proportion of undiagnosed DM in this study compared to that in the general Norwegian population may be a result of a generally low focus on DM screening in AAA patients.

Mortality, reintervention, and complication rates were low in this study regardless of DM status. This is in accordance with a contemporary Swedish national study reporting a 30-day mortality rate of 0.9% [33]. Several explanations may account for the low mortality, reintervention, and complication rates in the present study. All primary interventions were performed in an elective setting, and elective AAA repair is reported to have lower mortality rates than acute repair [34]. In addition, only infrarenal AAAs of degenerative origin were included in this study. Studies that include AAA of complex origin or suprarenal aortic clamping during OSR report a higher 30-days mortality [33, 35, 36].

DM was not associated with an increased risk of 30-day mortality, reintervention, or in-hospital complications, contrary to what is generally reported [4, 5]. In AAA patients, studies report ambiguous findings; Leurs et al. reported increased 30-day mortality rates in patients with DM following planned EVAR [13]. This is in accordance with a meta-analysis from 2014 reporting a lower perioperative survival rate in AAA patients with DM [16]. However, Hughes et al. found no differences in mortality [14], and Lopez-de-Andres et al. found that elderly type 2 DM patients had significantly lower mortality following AAA repair (OSR and EVAR) than non-DM patients [27].

Perioperative mortality rates were higher around the year 2000 than they are today [13, 14]. In more recent studies, mortality rates following infrarenal AAA repair are generally low, in some instances less than 1% [33, 37]. An increased uptake of EVAR, technical developments, a higher proportion of patients on BMT, and improved DM diagnostics may have positively influenced the outcome following AAA repair. In this study, patients with DM were more likely to receive best medical treatment compared to non-DM patients which is in accordance with DM guidelines [38]. A more optimal medical treatment may be reflected in the findings of lower low-density lipoprotein (LDL) levels in the DM patients. Also, the preoperative mean HbA1c in the DM patients were 52 mmol/mol (6.9%), which is in accordance with general treatment goals according to the DM guidelines [39,40,41]. De Martino et al. showed that anti-platelet and statin use was associated with improved 5-year survival in patients having coronary risk factors and undergoing AAA repair compared to patients on neither medication [42]. The impact on short-term morbidity and mortality after AAA repair of increasing proportions of patients receiving best medical treatment is still unclear. In addition, with low mortality regardless of type of repair, high numbers of operations would be needed to be able to demonstrate a possible significant difference in mortality between patients with and without DM.

Furthermore, undiagnosed DM was not considered in the mortality and morbidity analyses of the above-mentioned studies [13, 14, 33, 37]. Hence, the significance of undiagnosed DM for postoperative morbidity and mortality was not investigated in those studies.

Results from this study indicate that patients with undiagnosed DM are at increased risk of organ-related complications following AAA repair compared to non-DM patients. To reduce the risk of diabetes-related cardiovascular complications, DM guidelines emphasise the importance of DM screening in high risk individuals, and the use of best medical treatment [38]. AAA patients may be considered high risk individuals due to their high burden of cardiovascular comorbidities, obesity, and high age.

The negative impact of undiagnosed DM has been reported in percutaneous coronary intervention [43] but has not been examined in patients with AAA. However, retrospective cohort studies suggest that hyperglycaemia in patients with and without DM following aortic and non-cardiac surgery may be an indicator of poor clinical outcome [44, 45]. This study was not designed to investigate which pathophysiological mechanisms might be involved in increased risk of peri-operative organ-related complications in patients having undiagnosed DM. However, one could speculate that a metabolic dysregulation and the lack of stabilising DM medication may be involved.

The strength of this study is its prospective multicentre design, with national-level data on elective AAA surgeries from 11 of 15 vascular surgery units in Norway. The study included both OSR and EVAR. In general, studies exploring and comparing outcome following AAA repair in patients with and without DM are often retrospective, registry-based and include a heterogenous study population. In this study, only patients admitted for elective, degenerative infrarenal AAA repair, were included. Hence, the study population was homogeneous and well defined.

The guidelines for DM recommend confirmatory testing in asymptomatic individuals if a screening test is positive [46]. A second, and confirmative, HbA1c measurement was made in 97% of the study population with screening-detected DM, minimizing the risk of misdiagnosing DM patients.

During the study period (December 2017–December 2020), a change of principal investigator due to terminal disease, employment-related changes in site staffing at several study sites, and lack of focus on recruitment during the Covid-19 pandemic situation and holiday processing, affected recruitment into the study. The enrolment has, for this reason, not been consecutive, leading to a reduced participation rate.

Even though the sample size is not small (n = 877), the confidence intervals for ORs and HRs in the regression models are wide enough to also include clinically significant effect size. The association of undiagnosed DM (n = 32) with organ-related complications was evaluated using a post-hoc modified logistic regression analysis. Hence, results should be seen as hypothesis generating and confirmed in future studies. Due to the low number of deaths, we had low statistical power (i.e., large risk of type II error) for this outcome. Hence the association between DM and 30-day mortality should be interpreted with caution, and further explored. Potential underreporting of complications cannot be excluded since complications were reported by the treating vascular surgeon.