Prevalence and predictors of non-rheumatic valvular heart disease in patients with sickle cell disease: insights from the National In-Patient Database in 2016 and 2017

Introduction

In the United States, moderate-to-severe valvular heart disease (VHD) is reported to be as high as 2.5% among the general population and 17% among heart failure patients.^1–3 Moreover, its incidence increases as the population ages, with one out of eight patients aged 75 or older having some form of VHD. ¹ This fact explains the shift in etiologies of VHD in the developed world, with an abrupt decline in the prevalence of rheumatic heart disease and a significant increase in aging-related degenerative VHD. ⁴ Other recognized VHD etiologies include infectious endocarditis, inflammatory conditions, congenital heart disease, and sometimes under-recognized conditions such as radiation-induced and drug-induced VHD.^5–7

Sickle cell disease (SCD) life expectancy has dramatically improved in high-income countries, with patients approaching the fifth decade in the United States. ⁸ Aging in SCD has raised concerns about chronic organ damage and associated morbidity. ⁹ Adaptive and maladaptive cardiac remodeling in response to reduced oxygen-carrying capacity, chronic volume/iron overload, microvascular dysfunction, systemic arterial/pulmonary hypertension (PH), and kidney failure are potential etiologic mechanisms.^10–12 The reports linking VHD to SCD are scarce. Consequently, we aimed to assess the prevalence, predictors, and clinical outcomes in patients with non-rheumatic valvular heart disease (NRVHD) among patients with SCD using the United States National Inpatient Sample (NIS) database from 2016 to 2017.

Materials and methods

We present a retrospective analysis of hospital admissions for SCD patients in 2016 and 2017 from the National Inpatient Database. The NIS, part of the Healthcare Cost and Utilization Project (HCUP), is the largest publicly available all-payer inpatient healthcare database in the United States. It consists of a stratified systematic sample of discharges from all the HCUP hospitals, and when weighted, it yields national estimates of more than 35 million hospitalizations per year. This database provides clinical and nonclinical information and protects individual admission data from being known; consequently, institutional board review approval is exempted.

We identified 18 years old and older patients admitted for SCD using the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM). Using ICD-10 codes, we also identified and included NRVHD diagnosis in the analysis. We obtained patient-level factors (i.e., demographics and comorbidities) and hospital-level characteristics (i.e., geographic region, ownership, bed size, and teaching status) from the NIS database. Supplemental Table 1 lists the ICD-10-CM codes used. We extracted and analyzed data using the Statistical Package for the Social Sciences (SPSS Statistics version 22.0, IBM) using weighted data as recommended by the Agency for Healthcare Research and Quality. We reported the mean and standard deviation for continuous variables and performed univariate analysis using the ANOVA test. We described percentages for categorical variables and used Pearson’s chi-squared test for univariate analysis. Binomial logistic regression allowed us to calculate the adjusted odds ratios (aOR) for NRVHD and in-hospital mortality predictors. Multiple imputations addressed missing data for demographics and hospital characteristics variables as recommended by the HCUP.

Results

Between January 2016 and December 2017, we identified 192,460 weighted admissions for SCD; 2450 (1.3%) of them had NRVHD. Mitral insufficiency (MI) was the most common NRVHD found in 52.2% of the cases, while mitral valve prolapse (MVP) represented 12.4% of the NRVHD population. Aortic stenosis (AS) and aortic insufficiency (AI) were present in 10.8% and 12.7% of the NRVHD, respectively. Right-sided NRVHD had a lower prevalence, with 17.1% of patients having tricuspid insufficiency (TI) and 6.3% pulmonary insufficiency (PI). There were no cases of mitral, tricuspid, or pulmonary stenosis. In contrast, the most common valvulopathy in the population without SCD and NRVHD was AS in 71.9% of the cases, followed by AI in 17.9%, MI in 13.1%, and TI in 11.9% of the patients. Percentages of NRVHD and NRVHD rates per 100,000 patients are detailed in Supplemental Table 2 and Figure 1 for patients with and without SCD, respectively.

OPEN IN VIEWER

Figure 1.

Rates of non-VHD per 100,000 patients in SCD and non-SCD populations.

Within the SCD population, most of the patients that had NRVHD were women (61% vs. 39%, p < 0.001), older (44 vs. 34 years old, p < 0.001), and had a higher comorbidity burden (Elixhauser index ≥ 4 in 38.6% vs. 10.6%, p < 0.001) compared to those without NRVHD. The most common comorbidities in the NRVHD patients were hypertension (38.8%), secondary PH (36.1%), chronic kidney disease (CKD; 29.2%), electrolyte derangement (28.6%), chronic pulmonary disease (21.2%), and congestive heart failure (25.5%). Most of the admissions with NRVHD were in large-sized (59.6% vs. 14.3%, p = 0.005) urban teaching hospitals (79.8% vs. 4.1%, p < 0.086), located in the South of the US (55.7%) in which the most common primary expected payer was Medicare (49.8% vs. 32.6%, p < 0.001).

When comparing the results to the general population, the prevalence of NRVHD in people without SCD was 2.1%. Most patients with NRVHD and SCD were females, 61%, while in the NRVHD without SCD group, the sex was represented almost equally (males, 49.0%, and females, 50.2%). Moreover, patients with NRVHD in the SCD population were younger than those without SCD. Regarding the comorbidities, hypertension was the predominant one in patients without SCD and NRVHD at 61%, followed by electrolyte derangements (30.3%) and CKD (28.5%). The presence of secondary PH was significantly lower in this group of patients (2.6%). The most common valvulopathy in the general population was AS (71.9%), while AI (17.6%), TI (11.9%), and MI (13.3%) were less prevalent. The general patient characteristics are shown in Supplemental Table 2.

The characteristics associated with the presence of NRVHD in patients with SCD after multivariate logistic regression (adjusting for sex, race, age, and comorbidities) were secondary PH (aOR: 4.92, 95% confidence interval (CI): 4.47–5.41), congestive heart failure (aOR: 2.41, 95% CI: 2.18–2.66, p < 0.001), CKD (aOR: 2.35, 95% CI: 2.10–2.63, p < 0.001), peripheral vascular disease (aOR: 1.91, 95% CI 1.54–2.36, p < 0.001), rheumatoid arthritis/collagen vascular diseases (aOR: 1.82, 95% CI 1.50–2.21, p < 0.001), weight loss (aOR: 1.35, 95% CI: 1.11–1.64, p < 0.001), coagulopathy (aOR: 1.25, 95% CI: 1.06–1.46), hypothyroidism (aOR: 1.27, 95% CI: 1.05–1.53, p < 0.001), female sex (aOR: 1.17, 95% CI: 1.08–1.28, p < 0.001), and low yearly median household income (aOR: 1.94, 95% CI: 1.01–1.39, p < 0.020). On the other hand, the black population (aOR: 0.57, 95% CI: 0.47–0.71, p < 0.001), Hispanic patients (aOR: 0.36, 95% CI: 0.26–0.51, p < 0.001), and those admitted in the west region of the USA (aOR: 0.61, 95% CI: 0.51–0.74) were associated with lower prevalence of NRVHD. Table 1 details the aOR for factors associated with the increased prevalence of NRVHD in patients with SCD.

OPEN IN VIEWER

Table 1.

Clinical and demographic characteristics of patients with SCD associated with a higher prevalence of NRVHD.

Variables associated	aOR	95% CI	p-value
Age (years)
18–39 (ref)	—	—	—
40–59	2.41	(2.18–2.66)	0
60–79	3.55	(3.07–4.10)	0
80 and older	4.74	(3.45–6.52)	0
Female	1.17	(1.08–1.28)	0
Race
White (ref)	—	—	—
Black	0.57	(0.47–0.71)	0
Hispanic	0.36	(0.26–0.51)	0
Other	0.87	(0.53–1.42)	0.581
Elixhauser comorbidity index
0 (ref)	—	—	—
1–3	8.90	(6.93–11.42)	0
≥4	23.63	(18.27–30.55)	0
Comorbidities
Secondary PH	4.92	(4.47–5.41)	0
Congestive heart failure	2.67	(2.40–2.97)	0
CKD	2.35	(2.10–2.63)	0
Peripheral vascular disease	1.91	(1.54–2.36)	0
Rheumatoid arthritis/collagen vascular disease	1.82	(1.50–2.21)	0
Lymphoma	1.39	(0.82–2.35)	0.226
Weight loss	1.35	(1.11–1.64)	0.003
Hypothyroidism	1.27	(1.05–1.53)	0.014
Coagulopathy	1.25	(1.06–1.46)	0.007
Alcohol use	1.23	(0.87–1.74)	0.24
Drug use	1.12	(0.98–1.29)	0.098
Chronic liver disease	1.10	(0.91–1.32)	0.335
Depression	1.07	(0.94–1.22)	0.324
Neurological diseases	1.03	(0.89–1.20)	0.68
Diabetes mellitus	1.01	(0.81–1.25)	0.963

aOR, adjusted odds ratio; CI, confidence interval; CKD, chronic kidney disease; NRVHD, non-rheumatic valvular heart disease; PH, pulmonary hypertension; SCD, sickle cell disease.

The presence of NRVHD in patients with SCD was associated with a higher in-hospital length of stay (6.4 days vs. 5.3 days, p < 0.001) and higher in-hospital mortality when compared to those without NRVHD (aOR: 1.59, 95% CI: 1.19–2.14, p = 0.002) when adjusted for comorbidities, sex, age, and race.

Discussion

In this retrospective analysis of 2450 admitted patients with SCD with NRVHD from the NIS database, we found an overall high prevalence of mitral regurgitation present in more than half of the patients (52.2%), followed by TI (17.1%), trailed closely by AI (12.7%), MVP (12.4%), and AS (10.8%). We did not identify cases of stenotic disease in the mitral, tricuspid, or pulmonary valves. Importantly, patients with SCD show a higher prevalence of MI (46,545 vs. 11,062 per 100,000 patient population) and TI (15,273 vs. 10,008 per 100,000 patient population) than those without SCD.

Several studies have attempted to outline the imaging evidence of the association between cardiovascular disease and SCD, and only a few have focused on VHD.^13,14 In 1985, Lippman et al. first reported an unusually high incidence of MVP in 25% of 57 SCD patients. ¹⁵ Their findings faced criticism due to the lack of a control group, age disparity among the groups, and the use of M-mode echocardiography alone. In another study performed in Saudi Arabia, Hussain et al. reported no statistically significant difference in the prevalence of MVP in four age-matched groups of normal control subjects, SCD, sickle cell trait, and patients with other anemias. ¹⁶ The prevalence of MVP in this study ranged from 13.3% to 21.4%, the lowest of which was in the SCD group. Our study found similar results with MVP present in 12.4% of patients with NRVHD. Markiewicz et al., on the other hand, found a higher prevalence of 21% of MVP. ¹⁷ These differences may be explained by racial/ethnic, genetic, and overall geographic disparities between the included groups.

Ahmed et al. reported PH as the most common echocardiographic abnormality in 58% of the study population. They reported a significant incidence of TR in 50% and MR in 39% of the study population. ¹³ These results contrast with our findings as MR was present in 52% and TR in 17.1% of our study population. However, the prevalence of PH, though not the same, was also high in our study, with 36% of the patients with both SCD and NRVHD. One possible explanation for this disparity could be that the study by Ahmed et al. included a select group of hospitalized patients, so the reported high prevalence of PH may not truly represent the overall prevalence of PH in SCD patients, and the higher incidence of TR could be reflective of this fact. In our study, the secondary PH described encompasses the patients in the subgroup IV, which is clinically associated with miscellaneous diseases such as SCD. Our findings approximate three prospective studies in adults, in which echocardiography found that 20% of the participants had a mild elevation in estimated pulmonary artery pressures, and 9% had moderate-to-severe PH.^18–20

Our analysis found a greater propensity for NRVHD in patients with congestive heart failure, aOR 2.41. Cardiovascular morbidity in SCD has been attributed to chronic hemolytic anemia, which creates a constant state of high cardiac output leading to ventricular chamber dilation and increased ventricular wall stress. ²¹ Chronic recurrent vaso-occlusive episodes also subject cardiac myocytes to repeated ischemic insults even in the absence of atherosclerotic coronary artery disease.^22,23 These ischemic episodes are a possible mechanism of valvular pathology in SCD due to the involvement of papillary muscles leading to secondary valvular insufficiency.

We also found an increased association between CKD and the incidence of NRVHD in SCD patients. CKD manifests in SCD patients between 30 and 40 years old, similar to the age group in our study population. At the same time, around 5%–18% of SCD patients with CKD progress to end-stage renal disease. ²⁴ Therefore, our findings could reflect the more significant association between SCD and CKD and no particular pathological mechanism between CKD and NRVHD in SCD patients. Similarly, rheumatoid arthritis/collagen vascular diseases were also associated with increased odds of NRVHD in SCD patients (aOR: 1.82). Collagen vascular diseases are commonly implicated in cardiovascular events ranging from accelerated atherosclerosis to valvular abnormalities. ¹² Our study population’s findings likely share the mechanism of valvular abnormalities from collagen vascular disease in the general population.

Limitations must be accounted for in the result interpretation of our study. First, utilization of inpatient data, and not patient-level data, as the unit of analysis does not translate to generalizability for all patients with SCD. Second, this cross-sectional study did not evaluate the causal association between NRVHD and SCD. Moreover, the existence of NRVHD could have been influenced by other comorbidities and individual patient factors. Third, although we have included all reasonable ICD-10-CM codes for the variables studied, coding errors may have influenced our results since they tend to be more prevalent with billable codes and when administrative databases are used to investigate SCD and other rare diseases. Fourth, using an inpatient database limited the study to SCD patients admitted to the hospital, representing a small proportion of the SCD population.

Nevertheless, this study has provided an important and objective general overview of a large sample of hospitalized SCD patients. One of the significant strengths of our study is that we comprised a large number of admissions from the NIS database that has been used and validated in numerous medical research disciplines. The large sample size we obtained ensures that our study is highly powered to discern the association between SCD and NRVHD. Lastly, since we applied sampling weights to estimate inpatient outcomes, our results are generalizable to a much larger population than the sample studied. We recommend further investigating in the less appreciated relationship and prevalence of NRVHD and SCD.

NRVHDs, especially MI and TI, are comorbidities in SCD. Literature is scarce on this topic. The predictors found for its occurrence could help address modifiable factors that can positively affect patients with SCD who, due to the natural history of the disease, are at risk of developing VHD.

Supplemental Material

Supplemental Material