Laboratory and echocardiography markers in sickle cell patients with leg ulcers

Abstract

Chronic leg ulcers are a debilitating complication of sickle cell disease, associated with increased morbidity and perhaps mortality that affect 8–50% of patients. We evaluated the characteristics of SCD patients with a history of leg ulceration, including hemolytic rate, estimated pulmonary artery systolic pressure, and other parameters in a cohort of 505 adults with SCD. Ninety-four subjects (18%) had either active ulcers at enrollment or history of leg ulceration. Patients affected were older and predominantly had homozygous SS, lower body mass index, and pulse oximetry, higher tricuspid regurgitation velocities, markers of hemolysis, serum uric acid and serum amino-terminal probrain type natriuretic peptide, when compared to subjects without such history. In this prospective cohort of adults with SCD, we confirm that leg ulcers are still frequent and are associated with elevated TRV and markers of hemolysis. We describe a novel association of leg ulcer with hyperuricemia and oxygen desaturation and suggest potential implications for uric acid as a marker of vascular dysfunction. Chronic leg ulcers are a debilitating complication of sickle cell disease, associated with increased morbidity and perhaps mortality that affect 8–50% of patients. We evaluated the characteristics of sickle cell disease (SCD) patients with a history of leg ulceration, including hemolytic rate, estimated pulmonary artery systolic pressure, and other parameters in a cohort of 505 adults with SCD. Ninety-four subjects (18%) had either active ulcers at enrollment or history of leg ulceration. Patients affected were older and predominantly had homozygous hemoglobin SS, lower body mass index, and pulse oximetry, higher tricuspid regurgitation velocities, markers of hemolysis, serum uric acid and serum amino-terminal probrain type natriuretic peptide, when compared to subjects without such history. In this prospective cohort of adults with SCD, we confirm that leg ulcers are still frequent and are associated with elevated tricuspid jet velocity (TRV) and markers of hemolysis. We describe a novel association of leg ulcer with hyperuricemia and oxygen desaturation and suggest potential implications for uric acid as a marker of vascular dysfunction. The prevalence of leg ulceration in sickle cell disease patients varies, being low before 10 years of age, and in genotypes other than SS, and it is influenced by geographical location, with an occurrence as high as 75% in Jamaica and 8–10% in North America [1-3]. Advances in understanding the pathophysiology and management of leg ulcers have been slow. Abnormal vascular tone and activated, adhesive endothelium have been proposed as a pathway of end organ damage in sickle cell disease. Vasculopathy has been implicated in the development of sickle leg ulcers and other complications and is associated with hemolytic severity [4]. Patients with leg ulceration are at risk to develop serious complications of sickle cell disease, such as pulmonary hypertension, priapism, and possibly, renal disease [5-8]. Our investigation was carried out in a large cohort of adult patients with SCD screened prospectively for pulmonary hypertension by echocardiogram. We aimed to corroborate that leg ulcers are a manifestation of the hemolysis/vasculopathy subphenotype and identify additional clinical and laboratory characteristics associated with leg ulcers. The characteristics of the subjects are reported in Table I. Ninety-four (18%) of the 505 subjects reported having an active ulcer (24 subjects) or a history of leg ulceration (70 subjects). Most of the affected subjects had homozygous HbS (85 or 90.4%), whereas nine had double heterozygosity for HbS and HbC (9.6%). Affected patients were older (median age of 39 years compared to 31 years in the nonleg ulcer group, P < 0.0006). The body mass index (BMI) of subjects with leg ulcers was significantly lower than in patients without such history, with over a third of them having a BMI < 20.5 (P = 0.01), implying that patients affected with ulcers are generally underweight. The overall prevalence of hydroxyurea use was 39%, specifically 40.5% of the subjects without a history of leg ulcers were taking hydroxyurea compared to 34% of the subjects with a history of leg ulcers (P = 0.2). Similar findings were obtained among SS only: 46% versus 36% (P = 0.14). Hemoglobin F percentage was not different in the two groups (P = 0.6). The protective effect of fetal hemoglobin for the occurrence of leg ulcers has been inconsistent among several studies [3, 6, 8, 9]. A potentially confounding factor in modern cohorts is the widespread use of hydroxyurea in contrast to earlier analyses [3, 9]. It is possible that hydroxyurea statistically unlinks leg ulceration and fetal hemoglobin levels. Our study design could not answer whether hydroxyurea use in SCD is useful or detrimental for leg ulcers, because we did not establish the temporal relationship between hydroxyurea and development of leg ulceration. Some authors have recommended transfusion therapy for the management of leg ulcers [10]. We did not show a correlation between levels of hemoglobin A, a marker of recent transfusion, and prevalence of leg ulcers. On the other hand, lower hemoglobin was associated with a history of leg ulcers in patients with and without α-thalassemia. Patients with a history of leg ulcers exhibited more abnormal levels of markers of hemolysis, such as significantly higher serum lactate dehydrogenase (LDH) levels (P < 0.0001), absolute reticulocyte counts (P < 0.005), aspartate transaminase AST (P = 0.0004), and lower hemoglobin concentration (P < 0.0001). They also had 20% higher serum uric acid concentrations (P < 0.0001) alkaline phosphatase, and slightly lower serum albumin (P < 0.0001). Median ferritin level was higher in the leg ulcer group (528 vs. 388 mg/dL), which might be indicative of chronic inflammation and/or a population that is more frequently transfused because of more severe disease. The proportion of subjects that reported history of chronic transfusion was similar in the two groups, 16% in the “no ulcer” and 20% in the “yes ulcer,” leading to speculation that high ferritin is at least in part because of chronic inflammation. Higher iron burden may also contribute to vascular dysfunction. The role of α-thalassemia trait as a modulator for leg ulcer risk has been proposed by some investigators [3, 9]. Alpha thalassemia status was known in a subset of subjects, 294/505 (58%) and did not affect the prevalence of leg ulcers. The relationship of α-thalassemia trait to elevated estimated pulmonary arterial pressure has been similarly inconsistent [11-13]. These apparently conflicting results could be due to the small magnitude of the effect of α-thalassemia, below the detection of some studies. Moreover, α-thalassemia determination was available on only in ∼60% of the total population studied in our study, limiting our statistical power. After adjusting for age, hemoglobin, LDH, AST, albumin, uric acid, reticulocyte count, direct bilirubin, and amino-terminal probrain type natriuretic peptide (NT-proBNP) were still significantly associated with leg ulcer history. Multivariate logistic regression analysis, presented in Table II, identified older age (P < 0.001), lower hemoglobin concentration (P < 0.001), and higher AST (P = 0.06) as independent correlates of leg ulcer history in our cohort. We also evaluated leg ulcer history as a potential predictor of elevated TRV. After adjustment for other significant predictors of elevated TRV in our cohort, leg ulcer history was marginally associated with TRV ≥ 2.5 m/sec (P = 0.06), but not with TRV ≥ 3.0 m/sec (P = 0.4). Pulse oximetry readings were lower in subjects with leg ulcers. An oxygen saturation level <93% was three times more prevalent among subjects with a history of leg ulcers (22.4%) compared to subjects without a history of leg ulcers (7.6%, P = 0.0003). Recent publications report that, in SCD patients, small decreases of steady-state oxygen saturation correlate to the degree of anemia and indices of hemolysis, TRV, 6-min walk distance and risk for stroke or other central nervous system events [14-19]. The mechanisms by which a slight desaturation contributes to the risk of leg ulceration are not clear. However, steady-state desaturation is associated with more prominent desaturation during sleep which in turn causes endothelial, leukocyte, and erythrocyte activation and adhesion [16, 18]. These events could contribute to vasculopathy and decreased blood flow analogous to other conditions known to be associated with vasculopathy and leg ulcers, such as diabetes mellitus. Hemoglobin desaturation decreases arterial oxygen content and could compromise oxygen delivery to the skin. A measurable TRV, a noninvasive estimate of pulmonary arterial pressure, was obtained in 478 subjects. SCD subjects with leg ulcer history had a significantly higher prevalence of elevated TRV: TRV ≥ 2.5 m/sec was present in 59% of the subjects in the leg ulcer group and in 43% of the group without leg ulcer history (P = 0.006), and a TRV ≥ 3.0 m/sec, in 20% versus 13% (P = 0.004), for subjects with and without leg ulcers history, respectively. NT-proBNP is a widely used clinical laboratory marker of left ventricular stress that is elevated in adults with right ventricular stress because of pulmonary hypertension associated with SCD and other disorders [20, 21] and is associated with early mortality in adults with SCD. In this study, NT proBNP was significantly higher in subjects with leg ulcers (P < 0.0001). An increase in NT-proBNP has been associated with markers of hemolysis in SCD and other chronic anemias, with improvement after therapy [22], therefore it is possible that the higher NT-proBNP in the leg ulcer group reflects higher hemolysis. We also plan to investigate the possibility that diastolic-associated venous congestion contributes to leg ulceration. We evaluated the frequency of self-reported lifelong clinical complications in SCD subjects with leg ulcers. Clinical parameters included history of stroke or transient ischemic attack, headaches, avascular necrosis of bone, asthma, acute chest syndrome, cardiovascular dysfunction, priapism, kidney disease, and frequency of hospitalization for pain. None of them were significantly associated with leg ulceration history. Interestingly, the results of a prior cooperative study for sickle cell disease leg ulcer study show concordance with our results obtained using self-reported history [3]. The qualitatively very similar results support the self-reported approach. Of the 505 subjects in our cohort, 448 who enrolled on or before 12/31/2008 were followed for mortality (89%). Sixty-four of these subjects (14 %) had died by the time of analysis. Mean follow up was 54.3 months (25th and 75th percentiles: 24 and 85 months, respectively). Kaplan–Meier survival curves did not show a difference in survival between groups. Age, TRV, and serum ferritin were the independent predictors of death in our population, Table III. In conclusion, we demonstrate that leg ulcers are still a relatively common and serious complication of sickle cell disease, with 21% of HbSS patients reporting an ulcer in the past or having an active ulcer, similar to the 25% incidence reported by Koshy over two decades ago [3]. Leg ulceration in patients with SCD is associated with markers of hemolytic severity, cardiovascular risk, and surrogate markers of pulmonary hypertension. These epidemiologic data support the model of a “leg ulcer phenotype” originally proposed by Ballas, revised by Alexander et al. and subsequently by our group suggesting an overlap of the multifactorial pathobiological mechanisms of leg ulcer and vasculopathy [4, 23, 24]. A new observation that emerged in our analysis is the association of increased levels of uric acid with history of chronic leg ulcers. This association was considerably stronger than the one observed with creatinine or blood urea nitrogen (BUN), thus suggesting that it was not merely a nonspecific reflection of renal dysfunction. It is known that in patients without SCD, uric acid is a possible cause of hypertension and a marker for cardiovascular disease, pulmonary hypertension, and early mortality [25-28]. High uric acid may be a consequence of higher hemolysis because of the high hematopoietic turnover. The evolving evidence for uric acid as a risk factor for vasculopathy is consistent with our association between uric acid and leg ulceration as part of a vasculopathic complex in SCD. Therapies of limited efficacy abound for chronic sickle cell leg ulcers, but more uniformly effective therapy is needed. Five hundred and twenty-five SCD adult subjects were prospectively enrolled in a National Heart, Lung and Blood Institute (NHLBI)-approved protocol (Clinical Trial.gov no. NCT00011648) from February 2001 to March 2010. All patients provided written informed consent before enrollment. Our analysis included 505 (96%) of these subjects based on the availability of either a history of leg ulcers or an assessment of active leg ulcers at enrollment. A preliminary analysis of the first 325 subjects was presented in part at the Fourth Annual Sickle Cell Disease Research and Educational Symposium & Annual National Sickle Cell Disease Scientific Meeting (Hollywood, FL) and published in its meeting report [29]. This is an updated and extended analysis of an expanded cohort. Data prospectively collected included, a detailed past medical history and physical examinations, echocardiographic imaging, laboratory analyses, and pulse oximetry measurements. Echocardiography was performed as previously described [30]. Laboratory analysis included complete blood count, chemistries, serum LDH, and NT-proBNP. Subjects were evaluated while at steady state, at least 2 weeks after any type of acute exacerbation of SCD. Patients were followed for a median of 52.3 months (25th and 75th percentiles: 24 and 84 months). The diagnosis of sickle cell disease genotype was made by high pressure liquid chromatography (HPLC) and confirmed by DNA sequencing in the first 270 subjects, where it was questionable [8]. The presence of coincident α-thalassemia (α3.7 deletion) was determined in 294 patients as previously described [17]. Statistical Analysis. Clinical and laboratory characteristics were compared between subjects with active ulcers or a history of leg ulceration and subjects without a history of ulcers using the nonparametric Wilcoxon rank sum test for continuous variables and Pearson's Chi-square statistic for categorical variables. P values less than 0.05 were considered significant. TRV was examined by categorizing subjects into groups based on a cut-off of two standard deviations above the mean (≥2.5 m/sec) and a cut-off of approximately three standard deviations above the mean (≥3.0 m/sec). Logistic regression models were used to investigate associations of a variety of patient characteristics with active or past ulcers. Covariates were log-transformed as necessary to normalize their distributions and to reduce the influence of outlying values. Significance of individual predictors was determined using the Wald Chi-square statistic, retaining as significant variables with P values less than 0.05. Associations of leg ulcer history with mortality were examined using data from 448 subjects for whom follow-up information was available and whose enrollment date was on or before 12/31/2008. Subjects, who were not known to be dead, were censored at the date of last contact with study staff. Kaplan–Meier survival curves were calculated to examine mortality over time among patients with a history of leg ulcers compared with patients without a history of leg ulcers, using the log-rank test and a P value of 0.05 to determine significant differences in mortality between the two groups. Cox Proportional Hazards regression models were used to examine the association of leg ulcer history with mortality while adjusting for significant covariates as determined in prior mortality analyses of this cohort. The likelihood ratio test was used to determine the significance of individual regression coefficients. All analyses were performed using SAS version 9.1.3 (SAS Institute, Cary, NC) and Stata version 9.0 (StatCorp LP, College Station, TX). Caterina P. Minniti wrote the manuscript, designed the research and analyzed the data. James G. Taylor VI enrolled subjects, performed laboratory experiments and collected the data. Mariana Hildesheim analyzed the data and prepared the tables. Patricia O'Neal enrolled subjects. Jonathan Wilson analyzed data and prepared figures. Oswaldo Castro enrolled subjects and contributed to the design and interpretation of the data. Victor R. Gordeuk enrolled subjects, contributed to the interpretation of the data and to the writing of the manuscript. Gregory J. Kato designed the research, contributed to the analysis of the data and to the writing of the manuscript. The authors are grateful to Mary K. Hall, CIP for expert protocol management. We thank Darlene Allen for data collection and Britny Hall for manuscript preparation. We thank all the patients who participated in this study. Caterina P. Minniti*, James G. Taylor VI*, Mariana Hildesheim*, Patricia O'Neal , Jonathan Wilson*, Oswaldo Castro , Victor Gordeuk , Gregory J. Kato*, * Cardiovascular and Pulmonary Branch, Sickle Cell Vascular Disease Section, NHLBI, National Institutes of Health, Bethesda, Maryland;, Center for Sickle Cell Disease, Department of Medicine, Howard University, Washington, District of Columbia.