Association of Diabetes Mellitus With Health Status Outcomes in Patients With Peripheral Artery Disease: Insights From the PORTRAIT Registry

Nonstandard Abbreviations and Acronyms

After smoking, diabetes mellitus (DM) is the strongest risk factor for peripheral artery disease (PAD).¹ DM is present in 20% to 30% of patients with PAD,², ³ and patients with DM have a 2‐ to 4‐fold greater risk of developing PAD.⁴, ⁵ Patients with DM and PAD present with more advanced disease and have an increased rate of disease progression, peripheral vascular complications, and amputations compared with patients with PAD without DM.⁶, ⁷, ⁸, ⁹, ¹⁰ These patients also have greater functional impairment from PAD in terms of shorter walking velocities and distance¹¹ and greater rates of cardiovascular events compared with those with PAD and no DM.², ³, ¹²

While there are small single‐center reports suggesting a worse quality of life in patients with PAD and DM,¹³, ¹⁴, ¹⁵ these studies are limited by their small sample size, cross‐sectional design, absence of a control group, and potential confounding factors that were not adjusted for. There are no current data regarding the effect of comorbid DM on PAD‐specific health status and how it changes with time, which is important given the growing prevalence of DM. There is also lack of data regarding how patients' health status outcomes are affected by PAD treatment, specifically revascularization, based on their DM status. To address this gap in knowledge, we used an international multicenter registry of patients with symptomatic PAD to: (1) examine whether health status differs by DM status when patients present with new or worsening symptoms of PAD; (2) examine the trajectories of health status changes in patients with PAD over the course of a year according to their DM status; and (3) examine the effect of invasive PAD treatment on health status according to patient DM status. Documenting health status differences in patients with PAD who have comorbid DM can enable providers to better inform patients of their prognosis and can help identify strategies to potentially improve their health status.

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request.

Study Population

The PORTRAIT (Patient‐Centered Outcomes Related to Treatment Practices in Peripheral Arterial Disease: Investigating Trajectories) study is a multicenter, international prospective registry that enrolled patients presenting with new or worsening PAD symptoms to 16 specialty clinics (Data S1) across the United States, the Netherlands, and Australia from June 2011 to December 2015. Study details have been previously described.¹⁶ Briefly, adults with new or worsening claudication and an abnormal resting ankle‐brachial index (≤0.90) or a significant drop in postexercise ankle pressure of ≥20 mm Hg were included. Patients with noncompressible ankle‐brachial index (≥1.3), a recent episode of critical limb ischemia, or recent peripheral revascularization and those who were incarcerated; hard of hearing; unable to speak English, Dutch, or Spanish; or unable to provide informed consent were excluded.

Patients with an established diagnosis of DM were identified through medical record review at the time of their initial visit. DM type, treatment with insulin or oral hypoglycemics, glycemic control as measured by fasting plasma glucose and glycated hemoglobin (HbA_1c), and quality‐of‐care measures related to DM were abstracted from the patient's medical record. Demographic, socioeconomic (insurance, marital status, finances, employment, cost of care), symptom status (typical versus atypical symptoms), and lifestyle (activity level and smoking and alcohol use) factors were collected through patient interviews at the initial visit. Psychosocial factors such as depression, anxiety, social support, and perceived stress were assessed through standardized patient questionnaires, which included the Patient Health Questionnaire‐2 (PHQ‐2),¹⁷ 2‐Item Generalized Anxiety Disorder Scale (GAD‐2),¹⁸ ENRICHD Social Support Inventory (ESSI),¹⁹ and Perceived Stress Scale (PSS)²⁰ at baseline visit and at 3, 6, and 12 months of follow‐up. Other comorbidities, laboratory, performance measure adherence, medication, and diagnostic test information was collected from the medical record by trained study personnel. For all patients, serial health status and lifestyle factors were collected at 3, 6, and 12 months by telephone conducted by trained interviewers. PAD treatment was defined as receipt of supervised exercise therapy or revascularization (peripheral endovascular intervention or surgery; symptom‐driven) or medical therapy (antiplatelet and statin) within the first 3 months of follow‐up.

All study participants provided either written or telephonic informed consent. The study protocol was approved by the institutional review boards of all participating sites.

Health Status Assessment

Disease‐specific health status was assessed using the Peripheral Artery Questionnaire (PAQ),²¹ and generic health status was assessed using the Euro‐Quality of Life 5 Dimension Questionnaire (EQ‐5D).²² Trained study personnel administered both questionnaires in‐person at the initial visit and via telephone at 3, 6, and 12 months of follow‐up. The PAQ is a 20‐item validated, multidimensional, PAD‐specific health status instrument that measures the following health domains in patients with PAD: physical limitation, symptoms, symptom stability, social limitation, treatment satisfaction, and quality of life.²¹ Scores for all subdomains range from 0 to 100 with higher scores correlating with better health status. A summary score is calculated by averaging PAQ physical limitation, symptom, social limitation, and quality‐of‐life subscales and also ranges from 0 to 100. A difference of 8 points in the PAQ summary score has been proposed to be clinically important.²³ The EQ‐5D is a generic health status instrument that assesses health status among 5 dimensions of mobility, self‐care, usual activities, pain or discomfort, and anxiety/depression and has a visual analog scale (VAS) component that rates an individual's perception of their overall health on a scale of 0 to 100, where higher scores indicate better health status.²²

Statistical Analysis

Baseline characteristics were compared between patients with PAD who did and did not have DM using Student t test or Wilcoxon rank sum test for continuous variables and chi‐square or Fisher exact tests for categorical variables. Health status was compared between patients with PAD with and without DM at baseline, 3‐, 6‐, and 12‐month time points using Student t tests, including an omnibus test to assess for any differences over all time points.

A hierarchical, multivariable, linear regression model, with a random effect for site, was used to assess the association between DM and health status at baseline. This model was performed in a stepwise fashion, with the first step only adjusting for DM to generate unadjusted effect estimates of patient health status. In the second step, the model was partially adjusted for differences in demographic characteristics of age, sex, race, and country. To account for other patient and treatment characteristics that could affect the association of DM with patients' health status, the following characteristics (Table S1) were adjusted for in the third step to generate fully adjusted estimates: (1) socioeconomic factors (education, current work for pay, insurance, avoidance of care because of cost); (2) PAD disease severity (ankle‐brachial index, proximal versus distal location, laterality [bilateral versus unilateral disease], exacerbation versus new diagnosis, duration of pain, history of ulcer, amputation, or prior peripheral intervention); (3) comorbidities (hypertension, dyslipidemia, cerebrovascular accident, coronary artery disease, chronic heart failure, chronic kidney disease, chronic lung disease, musculoskeletal problems, sleep apnea, obesity/body mass index); and (4) psychosocial factors (ESSI score, PSS score, GAD‐2 score, PHQ‐2). In each step, all variables listed above were entered at the same time (simultaneous forced entry). Effect of DM on PAQ outcomes were presented as mean estimates with 95% CIs. Covariates were chosen a priori based on previously published literature and clinical judgement and supplemented with a random forest approach.

Additionally, another hierarchical, multivariable, linear model was used to examine the independent association of DM status with health status over follow‐up (3, 6, and 12 months). A random effect for site was included along with a Cholesky‐structured covariance matrix to account for repeated measurements. DM status and follow‐up time in months were included as fixed effects and a 2‐way interaction term was tested for differences in DM effect over time on health status. This repeated measure analysis was also performed in a similar stepwise fashion as the baseline health status analysis previously presented. In addition to the variables described above, fully adjusted repeated measures models for follow‐up health status were also adjusted for antiplatelet and statin receipt, referral to supervised exercise therapy within the first 3 months, and receipt of revascularization (endovascular PAD intervention or surgical PAD treatment) within 3 months, to adjust for differences in treatment at follow‐up among patients with and without DM.

To evaluate whether any specific group of patient factors would most explain the difference in health status among patients with and without DM beyond demographic factors, sensitivity analyses were conducted, where mean estimates, 95% Cis, and change in R² were evaluated after addition of patient factors related to each covariate group noted above separately, into a multiple regression model with DM and demographic patient characteristics.

To estimate whether patients with DM and PAD responded differently to PAD revascularization in terms of their health status outcomes on follow‐up, an interaction between DM status and revascularization was entered into the fully adjusted follow‐up health status model. A statistically significant interaction of DM×revascularization in the models would suggest that follow‐up health status differs based on receipt of invasive PAD treatment among patients with PAD who do and do not have DM.

To assess the effect of glycemic control on health status in patients with DM, we conducted a sensitivity analysis evaluating PAQ and EQ‐5D outcomes in a subgroup of patients with PAD and DM according to glycemic control at baseline (subdivided into groups by baseline HbA_1c of <6%, 6%–6.9%, 7%–7.9%, 8%–8.9%, and >9%).

Covariate data were largely complete with >90% of patients missing ≤1 covariate. The covariates with the largest number of missing data were body mass index (21.7%), duration of pain (14.5%), and treatment at 3 months (5.2%). Missing data were imputed using sequential regression imputation that included all of the variables from the multivariable model. All analyses were performed using SAS version 9.4 (SAS Institute). A 2‐tailed P<0.05 was considered statistically significant.

Results

Baseline Characteristics

Of the 1275 eligible patients with PAD enrolled in the PORTRAIT registry, we excluded 71 who were missing either all follow‐up interviews (n=70) or baseline PAQ assessments (n=1). Of the remaining 1204 patients, 398 (33.1%) had DM at the time of their baseline visit, a majority of whom had type 2 DM (n=375, 94.2%). The median HbA_1c of patients with DM was 6.9% (interquartile range, 6.1–7.9), 140 (35.4%) patients were on insulin therapy, 255 (64.6%) were on oral hypoglycemic therapy, and 76 (19.3%) were taking both oral hypoglycemics and insulin for treatment of their DM. A quarter of these patients (n=103, 25.9%) reported receiving DM education, 79 (19.8%) reported receiving diet counseling, and 24 (6%) reported receiving weight management counseling before baseline visit. Diabetic neuropathy was present in 40 patients (10.1%) and nephropathy was present in 10 patients (2.5%). Baseline characteristics of the study population are shown in Table 1. Patients with PAD and comorbid DM were more likely to be non‐White and unemployed. They were also more likely to present with worsening symptoms of PAD and have bilateral and distal PAD disease, nonhealing ulcers, and prior peripheral interventions. The patients with DM also had more cardiovascular comorbidities but were less likely to be current smokers. They were more likely to have greater levels of depression and perceived stress but similar levels of anxiety and social support.

Health Status According to DM Status on Initial Presentation

On initial presentation, patients with DM reported significantly worse PAD‐specific physical and social limitations, poorer quality of life, and higher symptom burden but similar general health status compared with patients without DM (Table 1). Figure 1 and Figure S1 and Table S2 show mean unadjusted PAQ and EQ‐5D scores at baseline in patients with and without DM. After adjusting for age, sex, race, and country, patients with DM had worse PAQ summary scores at baseline (partially adjusted mean difference, −3.39; 95% CI, −5.84 to −0.94 [P=0.006]). After additionally adjusting for socioeconomic, comorbidities, psychosocial characteristics, and PAD severity (fully adjusted), the effect of DM on PAQ summary score at baseline (adjusted mean difference, −0.65; 95% CI, −2.86 to 1.56 [P=0.56]) was no longer significant. A similar pattern was noted for PAQ physical limitation, symptom, social limitation and quality‐of‐life subscales (Figure 2). There was no difference in general health status at baseline, as measured by the EQ‐5D VAS, between patients with and without DM, in unadjusted or adjusted models.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Health Status Outcomes on Follow‐Up According to DM Status

In the unadjusted model, patients with DM had a significantly lower PAQ summary score over the follow‐up period of 1 year (average over 3‐, 6‐, and 12‐month time points) compared with those without DM, with a mean difference of −4.07 (95% CI, −6.59 to −1.54; P=0.002), which persisted after adjusting for difference in patient demographics (mean difference, −3.68; 95% CI, −6.19 to −1.18 [P=0.004]). Upon further adjustment for socioeconomic factors, PAD severity, comorbidities, psychosocial factors, quality‐of‐care measures, and treatment, the estimate for average PAD‐specific health status differences over a year as measured by PAQ summary score between patients with and without DM was marginal and nonsignificant, with a mean difference of −1.59 (95% CI, −4.06 to 0.88; P=0.21) (Figure 2). A similar pattern between patients with and without DM was noted for other PAQ subdomains on follow‐up (Figure 3).

• Download figure
• Download PowerPoint

In contrast to baseline, general health status as measured by EQ‐5D VAS was lower in patients with PAD who had DM compared with those without DM, on unadjusted, demographic‐adjusted, and fully adjusted models (mean difference, −2.19; 95% CI, −4.01 to −0.38 [P=0.02]) over the follow‐up (Figure 3).

Sensitivity Analyses

Sensitivity analyses exploring patient factors that most explain the differences between health status (PAQ summary score) among patients with and without DM at baseline (Table 2) showed a greater burden of psychosocial factors and comorbidities in patients with DM and that adjusting for this contributed most to the attenuation of the difference in PAQ summary scores between the 2 groups. For follow‐up health status over 3 to 12 months (Table 3), greater burden of comorbidities in patients with DM partially attenuated the difference in the follow‐up PAQ summary score.

A sensitivity analysis in a subset of patients with DM who had HbA_1c available at baseline (226 of 398, 56.8%) showed no difference in PAD‐specific or general health status at baseline or follow‐up among subgroups divided by glycemic control at baseline, except that patients with poor glycemic control ≥8% had worse PAQ symptom stability scores at baseline compared with those with HbA_1c <8% (Table S3).

Effect of Treatment on Health Status Outcomes According to Patient DM Status

Receipt of PAD revascularization within the first 3 months postbaseline was associated with a significant improvement in PAQ summary score and EQ‐5D for all patients (mean difference in PAQ summary score with 3‐month revascularization, 5.75 [95% CI, 3.01–8.49] P<0.0001; mean difference in EQ‐5D score, 3.03 [95% CI, 1.02–5.04] P=0.003), but there was no difference between patients with and without DM (interaction for DM×revascularization for PAQ summary score: P=0.69; EQ‐5D VAS, P=0.35). Similar results were obtained for other PAD subdomains (Table 4).

Discussion

In addition to improving survival, a key goal of PAD management is to improve symptoms, function, and quality of life of patients. In this large multicenter registry of patients presenting with symptoms related to PAD, presence of DM with PAD was associated with worse disease‐specific health status at presentation and in the year thereafter, even after accounting for demographic differences. This difference could potentially be explained by greater psychosocial and comorbidity burden in patients with DM, as the differences in health status were attenuated and no longer significant after accounting for those. Patients with DM had similar improvements in health status over a year with revascularization procedures as compared with their counterparts without DM.

Coexisting DM in patients with PAD is associated with significant clinical and economic morbidity¹⁰, ²⁴; however, its association with health‐related quality of life has not been previously delineated. Comorbid DM presents some unique challenges in PAD management, which could potentially affect health‐related quality of life. Patients with DM often present later in the disease course with more advanced disease and atypical symptoms.², ³, ²⁴ These patients often have diffuse, small‐vessel, and distal disease with significant calcium burden in their atherosclerotic plaques making intervention technically difficult.² These patients have greater rates of restenosis postrevascularization compared with their counterparts without DM but with PAD.², ²⁵, ²⁶, ²⁷, ²⁸ DM is also associated with increased rates of postintervention complications, including infections, amputations, and major adverse limb events.⁸, ²⁵, ²⁹ This is not only associated with a longer length of stay and higher hospital costs,³⁰ but can potentially have a significant effect on patient health status and quality of life.

Few studies have attempted to examine the effect of coexisting DM on quality of life in patients with PAD. In a small single‐center analysis of 92 patients with PAD, those with DM had shorter walking distance and walking speed and poorer general health status and quality of life compared with patients with PAD who did not have DM.¹³ A study by Amer et al¹⁴ also reported poorer general health status in patients with DM and PAD compared with patients with DM alone. In 920 patients with intermittent claudication, over half of whom had DM, Lozano et al¹⁵ reported slightly lower Walking Impairment Questionnaire scores and EQ‐5D scores in those with DM. All of these studies, however, were cross‐sectional in nature, had few patients with symptomatic PAD and DM,¹³, ¹⁴ and did not account for multiple coexisting comorbidities, socioeconomic and psychosocial factors, and PAD severity.

Our study is the first to report longitudinal health status outcomes over the course of a year in patients with PAD who did and did not have DM in a large multicenter cohort. Patients with DM had worse PAD‐specific health status at baseline and at 12 months of follow‐up. Our study suggests potential mechanisms that could explain this difference in health‐related quality of life. Sensitivity analyses suggest that greater prevalence of psychosocial factors such as depression and stress and other cardiac and noncardiac comorbidities in patients with PAD and DM compared with those without DM is a possible mechanism explaining the worse PAD‐specific health status noted in these patients. The health status differences between patients with and without DM were not significant after adjustment for these characteristics. Recognizing and adequately controlling coexisting comorbidities, especially psychosocial factors such as depression and stress, should be tested as potential interventions to improve the quality of life of patients with PAD and DM. Importantly, while initial PAD revascularization by itself was associated with significant improvements in PAD‐specific and general health status measures, there was no evidence of a differential effect of treatment with PAD interventions on the health status after a year of follow‐up in patients with DM. This suggests that both patients with and without DM should be offered similar treatment options for their PAD, including revascularization, as both groups experience similar gains in health status with treatment over time. DM status should not be a barrier to adopting an aggressive management approach in patients with PAD. Whether this pattern of improvement in health status is sustained longer than a year should be studied. A noteworthy finding of our study was that while patients with PAD with and without DM did not have any difference in generic health status at the time of their initial visit, patients with DM had significantly worse generic health status as measured by EQ‐5D VAS at 12 months compared with those without DM, even after accounting for many other patient and treatment characteristics that could affect the association of DM with patients' health status. This could be reflective of poor clinical outcomes following procedural treatment such as increased complications or readmissions. This could not be evaluated in the present study, but needs to be explored further in the future.

Our study results should be interpreted in the context of the following potential limitations. First, DM status was ascertained by medical record review and patient self‐report and not confirmed with laboratory testing. Second, given the observational nature of the study, there might be residual or unmeasured confounding in our results, even though we adjusted extensively for sociodemographic, clinical, psychosocial, and treatment factors. Third, we also could not assess whether health status outcomes differed with duration of DM. Since a majority of patients in our study had type 2 DM, whether patients with type 1 DM who have much more long‐standing DM and are on long‐term insulin also have similar health status outcomes is unknown. As follow‐up was limited to 1 year after the initial visit to a PAD provider, we could not determine the long‐term effect of DM status on PAD‐specific health status outcomes beyond a year. We were not able to assess the effect of glycemic control postbaseline on follow‐up PAD health status outcomes for all patients. However, sensitivity analysis showed no difference in health status based on glycemic control in a subset of patients with DM who had HbA_1c available at baseline. Finally, the health status perspectives of patients with PAD and comorbid DM were restricted to those with compressible ankle‐brachial index and those without critical limb ischemia, and our findings are not to be extended to patients with more severe disease.

Conclusions

We found that patients with DM and PAD have worse health status compared with those without DM when they present with symptoms of PAD and throughout a year of follow‐up thereafter. This may be explained by the differences in their psychosocial characteristics such as greater prevalence of psychosocial and cardiovascular comorbidities in patients with DM and PAD. However, they experience similar improvement in health status with revascularization as their counterparts without DM. Our results can help inform patients with PAD and DM and their providers regarding how their symptoms, function, and health‐related quality of life would be affected by their DM, how they can expect it to change over time, and how it is affected by PAD revascularization.

Disclosures

Dr Spertus owns copyright for the Peripheral Artery Questionnaire. He serves as a consultant to United Healthcare, Bayer, Janssen, AstraZeneca, and Novartis. He is the principal investigator of an analytic center for the American College of Cardiology. He has an equity interest in Health Outcomes Sciences. Dr Mena‐Hurtado serves as a consultant for Abbott, Cardinal Health, Boston Sci, Cook, Medtronic, and Bard BD. Dr Shishehbor is a consultant and serves on the global advisory board of Medtronic, Abbott Vascular, Terumo, Boston Scientific, and Philips. Dr Creager is supported by the American Heart Association Strategically Focused Research Network in Vascular Disease under award number 18SFRN339008. Dr Smolderen receives grant support from Boston Scientific, Abbott Vascular, and Terumo.

Sources of Funding

Research reported in this article was partially funded through a Patient‐Centered Outcomes Research Institute (PCORI) Award (IP2 PI000753‐01; CE‐1304‐6677), the Netherlands Organization for Scientific Research (VENI grant number 916.11.179), and an unrestricted grant from W. L. Gore & Associates, Inc (Flagstaff, AZ). The statements in this article are solely the responsibility of the authors and do not necessarily represent the views of the PCORI, its Board of Governors, or Methodology Committee. All articles for the PORTRAIT trial are prepared by independent authors who are not governed by the funding sponsors and are reviewed by an academic publications committee before submission. Drs Patel and Peri‐Okonny are supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health (NIH) under award number T32HL110837. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The funding organizations and sponsors of the study had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the article; or decision to submit the article for publication.

Footnotes