Impact of systemic lupus erythematosus on in-hospital outcomes of peripheral artery disease—insight from the National Inpatient Sample database

Abstract

Chronic inflammatory disorders like systemic lupus erythematosus (SLE) and rheumatoid arthritis are associated with worse outcomes in ischemic heart disease. However, there is a paucity of data regarding outcomes in patients with peripheral arterial disease (PAD) with concomitant SLE. The purpose of this study was to compare clinical features and in-hospital outcomes of PAD in patients with and without SLE from the general population using the Healthcare Cost and Utilization Project National Inpatient Sample database. We performed a cross-sectional analysis on 520,665 patients diagnosed with PAD from quarter 4 of 2015 to 2017. The primary endpoint was risk-adjusted in-hospital mortality. Of the total patient population, 3080 patients (0.6%) had SLE compared with 517,585 controls (99.4%). The observed in-hospital mortality was higher in patients with SLE (6.3% vs. 4.6%, P < 0.001). To the best of our knowledge, this is the largest population-based study investigating the impact of SLE in patients with PAD. Our analysis showed higher in-hospital mortality in SLE patients than in those without SLE. Early diagnosis and aggressive management of SLE and its complications in these patients have the potential to improve overall outcomes.

Systemic lupus erythematosus (SLE) is a systemic autoimmune disorder with a wide range of clinical presentations. The disease has varying prevalence geographically and has high overall morbidity and mortality.¹ Cardiovascular diseases are prevalent in SLE patients and are a major cause of morbidity and mortality, mainly related to atherosclerosis.² The inflammation and circulating autoantibodies associated with SLE make it an independent risk factor for the development of premature atherosclerosis, even in the absence of traditional risk factors of hypertension, hyperlipidemia, or renal disease.^3,4 This manifestation of atherosclerosis in the lower limbs’ vasculature leads to a low ankle-brachial index and increased prevalence of peripheral arterial disease (PAD), a coronary artery disease risk equivalent.⁵ This article compares the in-hospital outcomes among PAD patients with and without SLE.

METHODS

We utilized National Inpatient Sample (NIS) data from the Agency for Healthcare Research and Quality from quarter 4 of 2015 to 2017. Their database is the Healthcare Cost and Utilization Project administrative longitudinal database, which contains encounter-level information on inpatient stays, emergency department visits, and ambulatory surgery in all US hospitals. As the database contains deidentified patient samples, its use is deemed exempt from institutional review board approval.

All patients ≥18 years diagnosed with PAD in the hospital were included in the study. The data were further classified based on the presence or absence of coexisting SLE. To retrieve patient samples and related procedures, we used ICD-9 Clinical Modification codes and ICD-9 procedure codes: PAD, 443.9; SLE, 710.0; heart failure, 428.0; valvular heart disease, 396.9; hypertension, 401.9; diabetes mellitus, 250.00; hypothyroidism, 244.9; renal failure, 586; liver disease, 571; peptic ulcer disease, 533.9; coagulopathy, 286.9; obesity, 278.00. The clinical classification software developed by the Agency for Healthcare Research and Quality was used to obtain codes for certain comorbidities. Patients with missing demographic data were excluded from the study.

The NIS data sample contains data regarding in-hospital outcomes, procedures, and other discharge-related information. Variables included patient characteristics (age, race, sex, comorbidities, disposition) and illness severity (length of stay, mortality, associated comorbidities). The primary outcome of our study was the trend in in-hospital mortality among PAD patients with and without SLE. The secondary outcome was the impact of the presence of SLE on length of stay and hospitalization cost.

Statistical analysis was performed with statistical software R 2.9.2 and the epiDisplay package. Trends were calculated utilizing the trend weights provided with the data sample, and rates were expressed as the number of patients for that calendar year. The Cochran-Mantel-Haenszel test was used to assess the significance of trends. The chi-square test was used to compare categorical variables, and linear regression was used to compare continuous variables such as age, length of stay, and hospitalization cost. Multivariate logistic regression analysis was used to analyze the impact of SLE on in-hospital mortality. Multivariate analysis was adjusted for age, race, hospital characteristics, and comorbidities. Predictive margins were used to account for changing demographics, age with time, and impact on length of stay. A two-tailed P value of 0.05 was considered statistically significant.

RESULTS

In our study period spanning quarter 4 of 2015 to 2017, a total of 520,665 patients who were diagnosed with PAD were identified, 3080 (0.6%) with SLE and 517,585 (99.4%) without SLE. The in-hospital mortality in patients with concomitant PAD and SLE was 6.3%, while that in patients with PAD and without SLE was 4.6% (P < 0.001). On multivariate logistic regression analysis, the odds ratio for in-hospital mortality for PAD patients with and without SLE, adjusted for age, race, comorbidities, primary payer, and hospital characteristics, was calculated to be 1.90 (P < 0.001).

The SLE group had a higher proportion of patients with heart failure (HF), valvular disease, renal failure, and coagulopathy. The patients with PAD but without SLE had a higher percentage of patients with hypertension, diabetes mellitus, and obesity. Groups with and without SLE had a similar percentage of patients with liver disease. A similar percentage was also observed for peptic ulcer disease (0.6% vs 0.9% respectively, n.s.).

The length of stay for patients diagnosed with PAD with concomitant SLE was longer (3–11 days with a median of 6 days) compared to patients with PAD alone (2–8 days with a median of 4 days). Similar results were observed for the total hospitalization cost, which was higher for the group with SLE when compared to the group without SLE (Table 1).

Table 1.

Patient-level characteristics of PAD with SLE vs PAD without SLE in 520,665 patients from 2015 quarter 4 to 2017

Characteristic	PAD with SLE	PAD without SLE	P value
N (%)	3080 (0.6%)	517,585 (99.4%)
Age (years): mean + SD	60.7 ± 16.4	72.2 ± 13.3	<0.001
Gendera			<0.001
Male	510 (16.6%)	257,785 (49.8%)
Female	2570 (83.4%)	259,570 (50.2%)
Age groups (years)			<0.001
18–34	275 (8.9%)	7720 (1.5%)
35–49	470 (15.3%)	23,075 (4.5%)
50–64	880 (28.6%)	96,800 (18.7%)
65–79	1075 (34.9%)	216,925 (41.9%)
≥80	380 (12.3%)	173,065 (33.4%)
Raceb			<0.001
White	1535 (49.8%)	336,865 (65.1%)
Black	740 (24%)	65,085 (12.6%)
Others	805 (26.1%)	115,630 (22.3%)
Elixhauser comorbidities
Heart failure	640 (20.8%)	91,955 (17.8%)	<0.001
Valvular disease	275 (8.9%)	34,380 (6.6%)	<0.001
Hypertension	1825 (59.2%)	330,190 (63.8%)	<0.001
Diabetes mellitus	580 (18.8%)	182,285 (35.2%)	<0.001
Hypothyroidism	465 (15.1%)	64,260 (12.4%)	<0.001
Renal failure	995 (32.3%)	145535 (28.1%)	<0.001
Liver disease	155 (5%)	27,025 (5.2%)	<0.001
Peptic ulcer disease	20 (0.6%)	5125 (0.9%)	0.05
Coagulopathy	435 (14.1%)	50,410 (9.7%)	<0.001
Obesity	320 (10.4%)	71,435 (13.8%)	<0.001
Outcomes
In-hospital mortalityc	195 (6.3%)	24,060 (4.6%)	<0.001
Adjusted odds ratiod	1.90 (1.59–2.27)		<0.001
Length of stay (days): median (IQR)	6 (3–11)	4 (2–8)	<0.001
Total hospitalization cost: median (IQR)	$15,670 ($9,096–$31,868)	$14,202 ($7,989–$26,803)	0.01

^aMissing 230.

^bMissing 5.

^cMissing 220.

^dAdjusted for age, race, gender, Elixhauser comorbidities, primary payer, hospital teaching status.

IQR indicates interquartile range; PAD, peripheral arterial disease; SD, standard deviation; SLE, systemic lupus erythematosus.

DISCUSSION

SLE is an autoimmune disorder that primarily affects young women. Premature atherosclerosis due to SLE-related inflammation and traditional risk factors are the main underlying pathologies that can manifest as cerebrovascular disease, cardiovascular disease, and PAD. This contributes to increased mortality and morbidity in these patients.^6,7 Furthermore, prior studies have revealed worse 1-year outcomes following coronary intervention in SLE patients compared to non-SLE patients.⁸ While outcomes of cerebrovascular and cardiovascular diseases are well studied in this population, there is a paucity of data on outcomes of PAD. We utilized the NIS data to analyze these outcomes and recognize associated risk factors that may impact them.

Our study showed higher in-hospital mortality in PAD patients with SLE, with an odds ratio of 1.90. Prior studies have shown that SLE is a significant risk factor for coronary artery disease and requires aggressive risk factor modification.⁷ They have also demonstrated increased mortality especially related to cardiovascular disease and cardiovascular disease event hazards in SLE patients compared to the general population.^9–11 Outcomes of PAD, which is also considered a cardiovascular disease equivalent, would be expected to be similarly impacted by the presence of SLE. Hence, it becomes imperative to appropriately and timely diagnose PAD in these patients.

The mean age of patients with concomitant SLE and PAD was 60.7 years, approximately 12 years younger than patients without SLE. This is consistent with prior studies that revealed increased atherosclerotic cardiovascular disease risk in young patients with SLE.¹² Approximately 83.4% of patients in the SLE group affected by PAD were women, while men and women without SLE were almost equally affected by PAD. It is well known that the prevalence of SLE is higher in women than men, with a striking ratio ranging from 6:1 to 10:1.^13,14 Recent literature has also revealed relatively worse surgical and endovascular revascularization outcomes in women, including graft failure, wound infection, bleeding, limb loss, mortality, technical access, and embolic events.¹⁵ This would also explain the relatively poorer outcomes of PAD in patients with SLE, most of whom were women.

A relatively higher percentage of patients in the SLE group belonged to the nonwhite population and had worse outcomes. This is consistent with prior studies that revealed that SLE is more frequent, is more severe with higher disease activity, and causes more damage accrual in nonwhite populations than in whites¹⁶ and hence would contribute to worse outcomes in them.

Our analysis of baseline characteristics revealed a higher prevalence of HF, valvular heart disease, renal failure, and coagulopathy in those with SLE. Although pericarditis is the most common cardiac manifestation of SLE, HF and valvular abnormalities are well-known complications that significantly impact cardiovascular mortality and morbidity.¹⁷ Kim et al’s retrospective cohort analysis showed a significantly higher incidence of HF in SLE patients as compared to the general population with a worse overall cardiovascular profile.¹⁸ The prevalence of HF is higher in PAD patients and is associated with increased perioperative complications in PAD patients undergoing revascularization.¹⁹ This would contribute to worse outcomes in patients with PAD, as these conditions share a similar risk factor profile and overlapping pathophysiological phenomenon.

Similarly, there was a higher incidence of coagulopathy in the SLE group. This could be related to hypercoagulability and decreased fibrinolysis in patients with SLE,²⁰ which is also implicated in the atherosclerotic process contributing to PAD development. Proper management with prophylactic aspirin or anticoagulation in appropriate cases of SLE-related coagulopathy would hence prevent PAD and associated outcomes.²¹ Lupus nephritis is a commonly known complication of SLE, impacting almost 40% of patients diagnosed with it, especially young women.²² A study of lupus nephritis revealed a high prevalence of stage IV disease and its association with a worse prognosis.²³ The chronic inflammation and common atherosclerotic risk factors increase the risk of PAD in patients with chronic kidney disease.²⁴ Additionally, a study on the impact of kidney disease on peripheral arterial interventions revealed an increased risk of amputations and early mortality in patients with chronic kidney disease or end-stage renal disease undergoing intervention for PAD.²⁵ On the contrary, the non-SLE patient group with PAD alone had a higher prevalence of the traditional modifiable risk factors implicated in PAD development. These include hypertension, diabetes mellitus, and obesity.²⁶ In accordance with the higher mortality and morbidity of PAD in SLE patients, these patients had a longer length of stay and higher hospitalization cost compared to patients without SLE.

The retrospective observational study design and administrative data have inherent limitations. As hospitalizations and not individual patients are represented in the data, there is a potential for overestimation of the number of patients. The mean age of patients with SLE and PAD was 60.7 years, and some of these patients could have had other comorbidities that could have independently impacted outcomes. The number of PAD patients with SLE in our study cohort was much smaller than the number without SLE, which could have potentially affected the significance of certain comparisons due to low power. No studies have examined the validity of ICD-9 codes used for PAD or SLE; however, these codes have been used by several investigators in the past to accurately identify these patients from the Healthcare Cost and Utilization Project database. An inpatient diagnosis of SLE without proper rheumatology consultation might have limited the study, but it’s hard to obtain that kind of detail from the available data. It is difficult to determine if data were obtained from planned or unplanned hospital admissions, which affects overall mortality, as emergent procedures usually carry higher mortality. Hemodynamic data such as vital signs, laboratory data, and imaging parameters were not available in this data. Furthermore, information on medications and their compliance, readmissions, and long-term follow-up were not available. However, most of the limitations are attenuated by the considerable sample size obtained from these large databases.

To the best of our knowledge, this is the largest population-based study investigating the impact of SLE in patients with PAD. Our analysis showed higher in-hospital mortality in patients with SLE compared with those without. Study results show a higher association of HF, valvular heart disease, renal disease, and coagulopathy in patients with worse PAD outcomes (in patients with SLE) but cannot implicate causality. Further research is warranted to better understand the correlation of these risk factors with outcomes of PAD in SLE patients.