Evaluation of the utility of cardiac biomarkers for risk stratification in patients with lower extremity artery disease: A retrospective study

TL;DR

This study examines whether cardiac biomarkers can help assess risk in patients with lower extremity artery disease, finding limited utility for most biomarkers except troponin for mortality.

Contribution

The study evaluates the effectiveness of specific cardiac biomarkers for risk stratification in lower extremity artery disease patients.

Findings

CLTI patients had higher levels of NT-proBNP, CK-MB, and myoglobin compared to non-CLTI patients.

Troponin was the only biomarker independently associated with mortality in multivariate analysis.

Most cardiac biomarkers did not independently predict CLTI status in multivariate analysis.

Abstract

Critical limb threatening ischemia (CLTI) is associated with a one-year mortality rate of up to 25% making prompt diagnosis essentially. This study aims to investigate if cardiac biomarkers may serve as an effective tool for risk stratification in patients with lower extremity artery disease (LEAD). For this cross-sectional retrospective analysis, 21712 patients with LEAD were screened for eligibility from 2004 to 2020. Out of these patients, 367 were included and subdivided into those with CLTI and those without CLTI. Cardiac biomarkers, including N-terminal prohormone of brain natriuretic peptide (NT-proBNP), troponin, NT-proBNP/troponin ratio, creatin kinase myocardial band (CK-MB) and myoglobin, were retrospectively analyzed. Fifty-nine patients had CLTI (16.1%) with higher rates of NT-proBNP, NT-proBNP/troponin ratio, CK-MB and myoglobin (all p < 0.05) compared to non-CLTI patients. In univariate analysis, NT-proBNP, NT-proBNP/troponin ratio, CK-MB, myoglobin, age, C-reactive protein and non-insulin dependent diabetes mellitus (NIDDM) were associated with CLTI (all p < 0.05). In multivariate analysis, age and NIDDM remained significant predictors (all p < 0.05) while cardiac biomarkers were not independently associated with CLTI. Troponin, NT-proBNP and myoglobin were associated with mortality in univariate analysis (all p < 0.05). In multivariate analysis, troponin only remains to be associated with mortality (p = 0.001). Selected cardiac biomarkers failed to demonstrate statistically significant differentiation between CLTI and non-CLTI patients with LEAD, while troponin may be potentially associated with mortality.

Figures

Univariate analysis of cardiac biomarkers regarding discrimination between CLTI and non-CLTI.

Full text

1. Introduction

Peripheral artery disease (PAD) encompasses a wide range of arterial disorders affecting both the upper and lower extremities, as well as arteries in the carotid, vertebral, mesenteric, and renal regions, primarily caused due to atherosclerosis [1]. One predominant manifestation of PAD is lower extremity artery disease (LEAD), which exhibits a prevalence ranging from 5–8% affecting estimated more than 200 million individuals worldwide [2]. LEAD is associated with a significant mortality and morbidity, particularly in advanced stages such as chronic limb-threatening ischemia (CLTI). CLTI is characterized by ischemic rest pain or by tissue damage including infections, ulcers and necrosis of the lower limb with a reported one-year mortality rate of 25% and a one-year amputation rate of 30% [1,3]. Consequently, early diagnosis and adequate treatment strategies for CLTI are essential to prevent further complications.

LEAD frequently coexists with other arterial disorders, including cerebrovascular (CVAD) and coronary artery disease (CAD). The prevalence of LEAD among patients with acute coronary syndrome (ACS) is estimated by 40% and LEAD is an independent predictor of subsequent cardiovascular events [4,5]. In the Reach Registries, more than 50% of the patients with LEAD have concomitant CAD and/or CVAD [6]. While cardiac biomarkers like N-terminal prohormone of brain natriuretic peptide (NT-pro-BNP) or troponin have traditionally been studied in the context of cardiac diseases, recent investigations have evaluated their relevance also in LEAD, potentially due to a high comorbidity of different atherosclerotic disorders in those patients. Troponin and NT-pro-BNP have emerged as promising prognostic indicators for LEAD in several studies. It has been shown that these two parameters are associated independently with the incidence of LEAD. Additionally, it has been demonstrated that elevated troponin T levels are associated with higher mortality and amputation rates [2,7–10]. Less commonly investigated cardiac biomarkers, like NT-proBNP/troponin ratio, have additionally been studied for differentiation of ACS and whether troponin elevation in the emergency department was cardiac-related or not. This ratio may be helpful to discriminate ACS from Takotsubo cardiomyopathy and if troponin elevation was caused from conditions other than ACS [4,11]. Since 34% of CLTI patients present with elevated troponin levels, we want to investigate if different cardiac biomarkers may serve as prognostic markers and mortality predictors in patients with LEAD, who underwent endovascular revascularization, and may be useful in the future for risk stratification in these patients [8].

2. Materials and methods

2.1. Study design and patient cohort

This study is a retrospective cross-sectional analysis, wherein a blinded, pseudonymized dataset of patients with known LEAD were analyzed. A comprehensive screening of patients, who had been admitted to the outpatient clinic of Angiology of the Medical University of Graz due to LEAD from 2004 to 2020, was performed on 01.04.2022 via a fully electronic patient information system, called Medical Documentation and Communication network of Styria (MEDOCS). MEDOCS is installed in the province of Styria, Austria, to provide electronic health data from all public Styrian hospitals and hospital alliances [12]. Patient’s demographics, clinical parameters, comorbidities, and laboratory findings were recorded and analyzed. During the admission at the outpatient clinic of Angiology of the Medical University of Graz, each patient had undergone a detailed medical history including clinical symptoms and physical examination. Additionally, measurement of an ankle-brachial index (ABI), a duplex ultrasonography of the lower leg arteries, and obtainment of blood samples had been performed.

Inclusion criteria for this study were a diagnosed LEAD with the requirement of an endovascular intervention and available laboratory results, including cardiac biomarkers, which had been obtained within eight days prior to the endovascular intervention. Patients were excluded from this study if no endovascular intervention was necessary or if laboratory parameters were either not available or older than eight days prior to the endovascular intervention. All eligible patients were further subdivided into patients with and without CLTI. According to the recent guidelines, CLTI was defined as LEAD characterized by ischemic rest pain, with or without tissue loss or infection according to Fontaine classification stage III and IV [13].

Cardiac biomarkers were defined as troponin, NT-proBNP, NT-proBNP/troponin ratio, creatin kinase myocardial band (CK-MB) and myoglobin and were measured via a lithium heparin tube by routine laboratory work-up. For the detection of troponin and NT-proBNP, the fully automated Cobas 8000 test system (Roche Diagnostics) was used according the manufacturer’s instructions. Plasma samples were processed after centrifugation within 60–90min. Briefly, troponin was measured in plasma samples by using the Elecsys® Troponin-T high sensitive electrochemiluminescence immunoassay (Roche Diagnostics) with a detection limit of 3ng/L, a normal reference range from 3–14ng/L and a measuring range of 3–10.000ng/L. NT-proBNP was measured in plasma samples by using the Elecsys® NT-proBNP II electrochemiluminescence immunoassay (Roche Diagnostics), showing a limit of detection of 5pg/mL with a normal reference range from 5–125pg/mL and a measuring range of 5–30.000pg/mL. Both assays were routinely controlled using PreciControl Cardiac II for the various troponin and NT-proBNP concentration ranges; quality controls were run standardized at least once within 24 hours. There were changes in the assays over the 16-year study period. However, to ensure consistency throughout the entire duration of the study, we consistently applied the same thresholds for interpretation of the biomarker results.

The data regarding mortality, extracted from the MEDOCS system, represents all-cause mortality, which was consistently used throughout the 16-year study period.

2.2. Statistical analysis

Data are given as median and interquartile range (IQR) for continuous data, and as a frequency for categorical data. Univariate logistic regression analysis was performed to identify potential predictors for CLTI and mortality. If distribution of predictors did not lead to normally distributed residuals, the initial predictor was transformed (e.g., logarithmized). All variables showing an univariate p-value <0.05 were analyzed regarding multicollinearity and excluded if necessary. For the final set of potential predictors, multivariate logistic regression was performed using backwards selection to identify independent predictors. Odds ratio (OR) with 95% confidence intervals (95% CI) were calculated. A p-value less than 5% was considered significant. For data analysis, IBM SPSS Statistics 26 (IBM Corporation, Armonk, NY, USA) was used.

2.3. Ethical statement

The study was approved by the Institutional Review Board of the Medical University Graz, Austria (EK 34–006 ex 21/22). Due to the retrospective nature of the study, it was not necessary to obtain informed consent from the patients according to the Institutional Review Board.

3. Results

In total, 21712 patients with LEAD were screened for eligibility. Of those, 367 patients fulfilled all inclusion criteria, had available cardiac biomarkers and were considered for the final analysis. Fifty-nine patients (16.1%) had CLTI. Patients’ baseline characteristics are presented in Table 1.

Table 1: Patients’ baseline characteristics.

Between the subgroup with CLTI and without CLTI in univariate analysis, all cardiac biomarkers except troponin were significantly elevated in patients with CLTI (all p < 0.05) (Fig 1). Additionally, patients with CLTI were in univariate analysis significantly older (p < 0.001), were less commonly smokers (p = 0.018), had less commonly a non-insulin dependent diabetes mellitus (NIDDM) (p = 0.006) and a lower body mass index (BMI) (p = 0.002). Male/female ratio as well as other previous cardiovascular comorbidities and comedications were comparable between both groups. In multivariate analysis, age (OR: 1.06, 95%CI: 1.03–1.10) and NIDDM (OR: 0.22, 95%CI: 0.09–0.55) remained significant predictors for CLTI. Cardiac biomarkers, however, failed to be significant (Table 2).

Univariate analysis of cardiac biomarkers regarding discrimination between CLTI and non-CLTI.

Table 2: Uni- and multivariate analysis of clinical and laboratory parameters between patients with CLTI and without CLTI.

In univariate analysis regarding mortality, troponin, NT-proBNP and myoglobin were significant predictors (all p < 0.05). In multivariate analysis, only troponin remained a significant predictor for mortality (OR: 2.96, 95%CI: 1.59–5.5) (Table 3).

Table 3: Uni- and multivariate analysis of cardiac biomarkers regarding mortality.

4. Discussion

This retrospective analysis could demonstrate that cardiac biomarkers do not seem to be robust for the prediction of CLTI in patients with LEAD undergoing endovascular interventions. Although most patients’ characteristics were comparable to previous studies, there were also some differences between patients with CLTI and without CLTI in this study. The average age of our patient cohort was in line with findings from prior studies supporting the association between LEAD and advanced age by which a prevalence of about 20% in patients older than 70 years has been described [13,14]. Interestingly, there was a comparatively lower prevalence of CAD in the investigated cohort, with only 21% of patients having a documented history of CAD. It has been reported that 30% of patients undergoing peripheral vascular surgery have a concomitant CAD and that 52% of patients with symptomatic LEAD had concomitant CAD [15,16]. In our entire cohort, approximately 39% had a documented diabetes mellitus, with 12% of our patients having insulin-dependent diabetes mellitus (IDDM) and 27% having NIDDM. The literature suggests that about 20–30% of LEAD patients have diabetes mellitus as a comorbidity [17]. This number is slightly lower than in our patient population. Nearly half of our patients were smokers, which is again comparable to available literature [18]. Overall, our total patient cohort seems to be similar to previous studies.

Comparisons between our CLTI and non-CLTI group revealed some differences to existing data. In our cohort, about 16% of LEAD patients had CLTI, which is similar to the reported data of recent literature, where the prevalence of CLTI among LEAD patients is reported to be 11% [19]. Regarding gender distribution between CLTI and non-CLTI patients, more men were in both groups, although this difference was marginal and statistically not significant. Based on the available data, however, we might have expected to find more women in the CLTI group, as they tend to present later and more often with an advanced LEAD stage [20]. Despite similar gender distributions, the CLTI group exhibited a significantly higher average age and age was confirmed as a potential independent risk factor for CLTI in multivariate analysis confirming that age may play a pivotal role in disease severity and progression. This finding is consistent with the existing literature indicating that the likelihood of CLTI in patients over 80 years of age is 2–3 times higher than in younger patients [21]. Moreover, this emphasizes also the importance of age in predicting disease severity and the necessity for targeted interventions in older populations. The BMI in our cohort was statistically lower in the CLTI group compared to the non-CLTI group. This may be explained by the previously described ‘obesity paradox’. In this phenomenon, a higher BMI is associated with a more favorable prognosis regarding cardiovascular diseases [22]. The statistically significant lower prevalence of smokers in our CLTI group is, however, in complete contrast to the existing literature since smoking is the main risk factor for LEAD, and active nicotine abuse is associated with a 2–3-fold increased risk of LEAD [23]. Additionally, lower rates of concomitant NIDDM were observed in the group of CLTI patients, which is again contrary to previous studies as several data demonstrated that diabetes mellitus worsens the outcome for LEAD and is associated with higher amputation rates [24,25]. One explanation for the lower prevalence rates of smokers in the CLTI group would be that the recent analysis did not include ex-smokers. Due to the inclusion criteria of patients with cardiac biomarkers, we likely have a high proportion of patients with a cardiovascular history, who may have already quit smoking. From our cohort according to the guidelines, each patient is recommended to join a prevention program, including smoking cessation, before revascularization is indicated. However, even with intensive programs, the rates of nicotine abstinence are reported to be at maximum 21.3% [26]. Furthermore, one study described the paradox of reduced restenosis rates of after lower limb endovascular revascularization in people smoking 10 cigarettes and more [27]. Our study did not discriminate if patients had undergone a previous endovascular recanalization or not. Therefore, it may be possible that patients with an initial higher LEAD stage had undergone previously to study inclusion an endovascular recanalization, which had led to an amelioration of their symptoms and subsequently to a decrease of their LEAD stage. Those patients who continued smoking afterwards may develop reduced restenosis rates according t Schillinger et al [27] and it may be assumed also a lower symptomatic LEAD stage. The lower rates of NIDDM in patients with CLTI may be explained by the fact that those patients presenting with CLTI do not know about their concomitant NIDDM due to a neglected self-health responsibility. This finding may not accurately depict the health status of individuals with NIDDM. Rather, it reflects the contrasting severity of the condition between those with IDDM and those without diabetes mellitus. In the CLTI cohort of Darling et al [28], the number of NIDDM was lower than the number of IDDM or non-diabetic patients. Additionally, the long-term mortality and the adverse event rate was lower in the NIDDM group, like in our cohort. Moreover, it should be noted that our analysis only included selected patients who underwent endovascular interventions. Those who were primarily treated with vascular surgery or required immediate amputation due to complex morphological vascular changes were not included in the evaluation. It is important to note that in this retrospective study, diabetes mellitus was identified based on pre-existing diagnoses rather than on HbA1c values. This approach could have led to underdiagnosis of diabetes mellitus. Another difference in the prevalence of smoking and NIDDM in the CLTI group could be attributed to a selection bias as the study did not included the most severely affected patients with acute ischemia leading to acute amputation or other vascular surgery procedure.

Regarding cardiac biomarkers, NT-proBNP levels were significantly elevated in the CLTI group in univariate analysis but failed to be an independent predictor for CLTI in multivariate analysis. This is contrary to the study by Kumakura et al. [29], in which over 800 patients were included. This study reported that elevated NT-proBNP levels were independently associated with CLTI. Potential reasons for the discrepancy between our results and the result of Kumakura et al. [29] may be explained by the larger sample size, the prospective study design and routine collection of NT-proBNP. This could explain why NT-proBNP cannot be used to differentiate independently between CLTI and non-CLTI in our cohort. Similarly, NT-proBNP/troponin ratio, CK-MB and myoglobin achieved significance only in the univariate analysis, but failed in multivariate analysis suggesting a nuanced role in predicting CLTI. Additionally, troponin failed to demonstrate a statistically significant difference between the groups in univariate analysis, prompting considerations about the choice of troponin assay and its implications for diagnostic accuracy. Hicks et al. [30] were also unable to find a strong association of troponin and LEAD among their cohort. Overall, the respective cardiac parameters did not appear to be predict robustly and independently CLTI in patients with LEAD. However, compared to other studies, our study differentiated between patients with CLTI and non-CLTI. Most previous studies used the ankle brachial index (ABI) for the diagnosis of LEAD, while we have included only patients who had undergone an endovascular recanalization due to their symptomatic LEAD. Therefore, we did not include early stage of LEAD. Whether these parameters, in combination with others, can help identify LEAD patients who require urgent revascularization and aggressive conservative management remains to be clarified. It may be worthwhile to consider differentiation based on earlier stages (Fontaine stages I and IIa) as well.

In univariate analysis, troponin, NT-proBNP and myoglobin were statistically significant associated with mortality. This finding is comparable to the existing literature, wherein troponin and NT-proBNP is associated with increased mortality in patients with LEAD [10,31]. In multivariate analysis, only troponin remains to be associated with mortality, which is consistent with the study by Linenmann et al. [7] in which troponin was also related with higher mortality and amputation rates. On the other side, our multivariate models did not include additional confounding factors, which may influence the association results including arterial hypertension, renal insufficiency. Due to the fact that only a small number of patients were included and therefore the multivariate models may be at risk of overfitting, larger cohorts are needed to validate our results on mortality and also on CLTI.

Limitations of our study are the retrospective design and reliance on data exclusively from medical records. The use of medical records as the primary data source may lead to potential limitations to the study. The study only includes patients who underwent endovascular interventions, excluding those who received surgical treatment or immediate amputation. This represents a selection bias and may limit the applicability of the findings to the broader LEAD population. Additionally NT-proBNP and troponin were not routinely collected in all patients, potentially skewing the results. This selective inclusion may have biased the findings toward patients with more severe cardiovascular comorbidities. Due to inclusion of only patients with available cardiac biomarker data and those undergoing endovascular interventions, the results of this study cannot be reproduced on the general LEAD population, as those LEAD patients with a less severe disease or those treated surgically were excluded. Furthermore, our small sample size of patients with CLTI reduces also the statistical power and limits the robustness of conclusion. Additional selection bias are that no information about the proportion of former smokers were recorded.

5. Conclusion

In conclusion, while cardiac biomarkers tend to be elevated in patients with CLTI, our study did not find them to be significantly indicative of CLTI presence in our cohort. This underscores on the one hand the complex nature of risk factors in LEAD and, on the other hand, these findings may be accompanied with methodological limitations. Future research should focus on larger sample sizes and more diverse cohorts to thoroughly evaluate the utility of these biomarkers for risk stratification in LEAD patients.

Supporting information

S1 DataMin data set.(XLSX)