Cerebrovascular Disease Associated with Sickle Cell Pulmonary Hypertension

Abstract

In patients with sickle cell disease, anemia is a recognized risk factor for stroke, death, and the development of pulmonary hypertension. We have proposed that hemolytic anemia results in endothelial dysfunction and vascular instability and can ultimately lead to a proliferative vasculopathy leading to pulmonary hypertension. Consistent with this mechanism of disease, we now report a case series of six patients with obliterative central nervous system vasculopathy who also have pulmonary hypertension and high hemolytic rate. These patients, identified in the course of a prospective screening study for pulmonary hypertension, presented with neurological symptoms prompting neuroimaging studies. Compared to 164 other patients of similar age in the screened population, those with newly diagnosed or clinically active cerebrovascular disease have significantly lower hemoglobin levels and higher levels of lactate dehydrogenase. A review of the literature suggests that many clinical, epidemiological, and physiological features of the arteriopathy of pulmonary hypertension closely overlap with those of stroke in sickle cell disease, both known to involve proliferative vascular intimal and smooth muscle hypertrophy and thrombosis. These cases suggest that cerebrovascular disease and pulmonary hypertension in sickle cell disease share common mechanisms, in particular, reduced nitric oxide bioactivity associated with particularly high-grade hemolysis. Clinicians should suspect occult cerebrovascular disease in sickle cell patients with pulmonary hypertension.^†

Introduction

Stroke is a common complication of sickle cell disease, with a lifetime risk of 25 to 30% and a peak incidence in childhood at 7 years of age [1,2]. Considerable progress has been made in surveillance and prevention of ischemic stroke in children with sickle cell disease, using prophylactic transfusion therapy in children with high flow velocity transcranial Doppler measurements [3]. The lowest risk of ischemic stroke in sickle cell disease occurs in the third decade of life, a time when the risk of hemorrhagic stroke is the highest. Despite considerable investigation, the mechanism of ischemic stroke in patients with sickle cell disease remains unclear, appearing to involve a combination of large vessel arteriopathy and thrombotic disease, in contrast to the suspected microvascular occlusion in other sickle cell complications [4,5]. The risk of stroke is associated with severity of anemia and relative systolic hypertension [1,2]. Interestingly, these have also emerged as risk factors for pulmonary hypertension. While conducting a screening study for pulmonary hypertension in patients with sickle cell disease, we encountered six patients with symptomatic cerebrovascular disease, all of whom also had pulmonary hypertension. The clinical profiles in these cases are comparable to that found in “hemolysis-associated pulmonary hypertension,” consisting of pulmonary vasculopathy with reduced nitric oxide bioactivity associated with particularly high-grade hemolysis [6]. The apparent overlap in these six patients of cerebrovascular disease with hemolysis-associated pulmonary hypertension suggests that these two syndromes involve similar mechanisms. A review of the literature indicates previously unrecognized similarities in the epidemiology, histopathology, and physiology of these two disorders.

Patients and Methods

Patients

Patient 1 is a 22-year-old African American male with homozygous sickle cell disease and a prior history of stroke, transiently treated with transfusions and then hydroxyurea therapy. On screening, he was found to have pulmonary hypertension and particularly low hemoglobin, with very high reticulocyte count and indirect bilirubin and lactate dehydrogenase (LDH) (Table I). Two weeks after recovering from vaso-occlusive crisis and acute chest syndrome, he presented with acute right-side weakness. Magnetic resonance imaging and angiography (MRI/MRA) showed a new infarct in the left centrum semiovale and segmental occlusions of proximal branches of the left anterior cerebral artery (ACA) and middle cerebral artery (MCA) (Figure 1).

TABLE I. Patient Characteristics.

	SCD patients with identified pulmonary hypertension and cerebrovascular disease								SCD comparison
Patient number	1	2	3	4	5	6	n	Mean	n	Mean (SD)
Gender	M	F	F	F	M	M	6	0.5 female	164	0.59 female
Age (years)	22	23	34	31	37	47	6	32	164	32 (8)
Systolic blood pressure (mm Hg)	129	99	103	126	128	152	6	123	148	118 (15)
Diastolic blood pressure (mm Hg)	60	58	48	58	51	78	6	59	148	65 (10)
Hemoglobin (g/dL)	7.0	8.1a	8.7	6.9	7.4	8.5	6	7.8	158	9.0 (1.6)**
Reticulocytes (1000/μL)	342	344	370	360	294	190	6	317	150	278 (134)
Indirect bilirubin (mg/dL)	5.8	7.7	4	10	3.2	1.2	6	5.3	155	2.4 (1.5)
Direct bilirubin (mg/dL)	0.2	0.6	0.4	0.4	1	0.5	6	0.5	156	0.5 (0.4)
Lactate dehydrogenase (U/L)	551	1065	486	771	442	619	6	656	139	374 (145)*
Creatinine (mg/dL)	0.5	0.6	0.5	0.5	0.6	0.5	6	0.5	156	0.9 (2.4)**
Fetal hemoglobin (%)	5.1	1.2	9.2	5.2	0.7	6.9	6	4.7	158	8.5 (6.2)*
Arginine (μM)	34	31	31	45	53	40	6	39	142	39 (16)
Ornithine (μM)	74	78	57	41	77	68	6	66	142	63 (23)
Arginine/ornithine ratio	0.46	0.40	0.54	1.10	0.69	0.59	6	0.63	142	0.71 (0.41)
Citrulline (μM)	16	18	23	12	19	3	6	15	142	22 (12)
Proline (μM)	255	322	211	167	262	257	6	246	142	215 (79)
Tricuspid regurgitant jet velocity (m/s)	2.5	3.7	2.5	2.5	2.6	3	6	2.8	164	2.2 (0.6)*
Brain natriuretic peptide (pg/mL)	240	252	177	156	103	349	6	213	81	100 (117)*

Note. Clinical and laboratory characteristics are provided for the six patients with sickle cell disease, pulmonary hypertension, and cerebrovascular disease. Values are from steady state at the time of original screening for pulmonary hypertension and are representative of subsequent serial values from these patients. Mean values are compared among the six patients in this case series with cerebrovascular disease and pulmonary hypertension (patients 1–6) and all screened patients of a similar age range with sickle cell disease, excluding patients with hemoglobin SC disease. Due to extreme skewing of brain natriuretic peptide with renal insufficiency, its comparison group is limited to patients with creatinine levels in the same range as patients 1–6 (i.e., creatinine < 0.7 mg/dL). For the SCD comparison group, mean values are provided, with standard deviations shown in parentheses. No significant differences were seen among the six patients and the comparison in the following variables: alanine aminotransferase, aspartate aminotransferase, erythrocyte sedimentation rate, C-reactive protein, alkaline phosphatase, ferritin, iron, transferrin, or transferrin saturation.

^aOn transfusion therapy; n.d., not done.

^*P < 0.05, two-tailed unpaired Students t test;

^**P < 0.01.

Fig. 1.

Watershed infarcts due to cerebrovascular disease in the circle of Willis. Chronic infarcts in the left frontal and left parietal watershed regions are identified in patient 1 on axial FLAIR MRI (arrowheads, a). High-grade focal stenosis of the M1 segment of the left MCA is present on MRA (long arrow – b). A moya–moya type pattern is developing in the territory of the left PCA (short arrow, b).

Patient 2 is a 23-year-old Puerto Rican woman with homozygous sickle cell disease who presented with recurrent transient ischemic attacks (TIA), strokes, and dyspnea on exertion. She was found to have severe pulmonary hypertension by echocardiography, confirmed by right heart catheterization. MRI/MRA demonstrated numerous old bilateral, watershed cerebral infarcts and stenosis of the left ACA and MCA (Figure 2). Despite chronic transfusion therapy, she maintained a low hemoglobin, very high reticulocyte count, and indirect bilirubin and LDH levels (Table I).

Fig. 2.

Early subacute infarct at the level of the centrum semiovale on axial MRI images performed 16 h after ictus in patient 2. Hyperintense lesion on DWI (arrow, a) suggests acute to subacute infarct, which is confirmed by the restricted diffusion (hypointensity) on the computed ADC map (arrow, b). Hyperintensity on FLAIR indicates subacute to chronic infarct (arrow, c). Other foci of hyperintensity on FLAIR suggest chronic infarcts in the watershed areas of the centrum semiovale (arrowheads). MRA demonstrates high grade stenosis at the left MCA trifurcation (thin white arrow, d) with attenuation of distal MCA branches. Prominence of the distal left PCA branches (black arrows, d) indicates the development of collateral circulation.

Patient 3 is a 34-year-old African American woman on hydroxyurea therapy for homozygous sickle cell disease, with a history of chronic leg ulceration. On screening study, she was found to have lower than average hemoglobin, severe reticulocytosis, and high LDH (Table I). She had pulmonary hypertension, for which she entered an IRB-approved trial of sildenafil therapy. She suffered an episode of vaso-occlusive crisis complicated by the fat embolization syndrome, including encephalopathy, renal insufficiency, pulmonary infiltrate, and lipid-laden macrophages identified from induced sputum by oil red O stain. Two weeks later, she developed acute right hemiparesis and dysarthria. MRI/MRA confirmed new cerebral infarct, with evidence of unsuspected prior infarcts and chronic occlusion of both internal carotid arteries (Figure 3).

Fig. 3.

Chronic watershed infarcts resulting from nearly complete occlusion of the cervical internal carotid arteries at the carotid bifurcation. Axial FLAIR MRI from patient 3 demonstrates (a) right frontal cortical infarct and bilateral, left greater than right, chronic infarcts in the white matter of the centrum semiovale at the boundary of the anterior and middle cerebral artery territories. Volume rendered 3D TOF MRA of the intracranial vasculature (b, inferior projection; c, left lateral projection) shows complete absence of flow-related enhancement in the intracranial carotid arteries (asterisk). Note the large vertebrobasilar system and, in particular, the marked enlargement of the posterior communicating arteries that supply collateral flow to the anterior circulation (white arrows). An angiographic “string sign” is seen in the cervical ICA just above the carotid bifurcation (d, left anterior oblique projection of 2D time of flight MRA).

Patient 4 is a 31-year-old Jamaican woman with homozygous sickle cell disease and a history of prior stroke at 17 years of age. On screening, she was found to have pulmonary hypertension and TIA episodes have continued to recur. She chronically has severe anemia, reticulocytosis, and prominent elevation of serum LDH and indirect bilirubin (Table I). MRI/MRA showed very severe right-side encephalomalacia with partial obliteration of the ACA, complete obliteration of the MCA and PCA, and some irregularity in the left MCA and PCA.

Patient 5 is a 37-year-old African American man with homozygous sickle cell disease, found on screening to have mild pulmonary hypertension, more severe than average anemia, reticulocytosis, and elevation of LDH (Table I). In evaluation of a carotid bruit, MRI indicated multiple focal old infarctions of the left frontal and parietal lobes and 80% narrowing of the left internal carotid artery. Six months later, he developed acute right-side weakness and dysarthric speech associated with new ischemic infarction in the watershed zone of the left cerebral hemisphere, with narrowing and irregularity of both internal carotid arteries.

Patient 6 is a 47-year-old African American man with homozygous sickle cell disease, found on screening to have systolic hypertension, more severe than average anemia, and markedly elevated LDH, with moderate pulmonary hypertension (Table I). He presented with paresthesias of the hand, and MRI/MRA showed occlusion of the right ACA, with multiple small areas of hyperintense signal in periventricular and subcortical white matter, consistent with small-vessel ischemic injury. This serendipitous finding did not closely correlate neuroanatomically with the affected hand, and nerve conduction velocity and electromyography studies did not disclose an etiology for the symptom.

Methods

The patients in this case series were identified during follow-up on a prospective screening protocol for pulmonary hypertension in patients with sickle cell disease. The methods for prospective screening for pulmonary hypertension and the data on the first 200 patients screened have been published, with an additional 40 patients screened subsequently [7]. Pulmonary hypertension was defined as a tricuspid regurgitant jet velocity ≥ 2.5 m/s on Doppler echocardiography. All patients signed informed consent for screening and laboratory research studies on an NIH IRB-approved protocol. All laboratory values were obtained on blood samples procured during a baseline outpatient clinic visit while in steady state and not during acute vaso-occlusive events. The values are representative of serial steady-state laboratory values obtained over many months (data not shown). Although prospective neuroimaging was not part of the study, during the course of follow-up on this protocol, six patients with homozygous sickle cell disease were identified with ongoing or new neurological symptoms prompting magnetic resonance imaging of the brain. In order to develop a suitable comparison group for the six homozygous sickle cell patients in this case series, from our screening group of 240 patients with sickle cell disease we excluded those with HbSC and included only those patients of the same age range, providing a comparison group of 164 patients. The older patients were excluded in the comparison group due to the markedly different incidence of pulmonary hypertension [7]. The patients with hemoglobin SC disease were excluded because both stroke and pulmonary hypertension are far less frequent in this subgroup, and none of our six patients in this case series has hemoglobin SC disease [7,8]. Due to the well-known skewing of brain natriuretic peptide with renal insufficiency, its comparison group is further limited to patients with creatinine levels in the same range as patients 1–6 (i.e., creatinine < 0.7 mg/dL). Amino acid levels were measured by the Mayo Clinic clinical laboratory. Comparisons were made using the two-tailed, unpaired t test. Results were considered significant if P < 0.05.

Results

Six adults with homozygous sickle cell disease presented in our screening program with ischemic stroke or other neurological symptoms that led to the new diagnosis of cerebrovascular disease. In our entire current cohort of 240 patients with sickle cell disease screened for pulmonary hypertension (101 of these with 2-year follow-up), 80 have pulmonary hypertension. All 6 patients with new cerebrovascular disease have concurrent pulmonary hypertension (Table I).

All six patients with pulmonary hypertension and cerebrovascular disease had evidence of hemolytic anemia more severe than the average patient in our screening study (Table I). This was indicated by a significantly lower hemoglobin level (P = 0.008) and higher serum LDH level (P = 0.03) than a comparison group consisting of patients with sickle cell disease of the same age range, excluding those with hemoglobin SC disease. In addition, the fetal hemoglobin level was significantly lower in the six patients compared to the comparison group (P = 0.03). Other markers of hemolytic severity also showed trends in these six patients: elevated indirect bilirubin (P = 0.07), higher reticulocyte count (P = 0.2), and elevated aspartate aminotransferase (AST) level (P = 0.08) despite normal alanine aminotransferase level, a pattern suggestive of release of erythrocytic AST due to intravascular hemolysis. Remarkably, the hemolytic pattern of reticulocytosis, high LDH, and indirect bilirubin level was not ablated in the patient on chronic transfusion, although her transfusion program was not intensive. The six patients also had lower serum creatinine levels than the comparison group (P = 0.006), although the clinical significance of this is unclear. Significant differences were not seen in several markers of liver disease, iron burden, or inflammation. Mean platelet and leukocyte counts did not differ significantly between the six patients and the control group. Supporting the diagnosis of pulmonary hypertension, the mean tricuspid regurgitant jet velocity, an echocardiographic indicator of pulmonary artery systolic pressure, was significantly higher in the six patients than in the comparison group (P = 0.03) (Table I). This is likewise true for brain natriuretic peptide, which our group has found to be a relatively sensitive and specific indicator of pulmonary hypertension in patients with sickle cell disease (R.F.M. et al., manuscript in preparation).

Discussion

In conducting our prospective screening study of adults with sickle cell disease for pulmonary hypertension, 6 patients presented with symptomatic or newly diagnosed cerebrovascular disease during the screening period. Despite the fact that only one third of the 240 patients we have screened have pulmonary hypertension, remarkably, all 6 patients with newly diagnosed or clinically active cerebrovascular disease fall in this group. Conversely, we have not encountered new neurological disease in the 160 patients without pulmonary hypertension during this screening study. Because neuroimaging was not performed prospectively in this screening study for pulmonary hypertension, its true prevalence is likely to be even higher than that represented by this case series. Since our follow-up has been more intensive in patients with pulmonary hypertension, it is possible that this may have contributed to increased sensitivity of detection for neurological symptoms in this population. However, the concurrence of pulmonary hypertension and cerebrovascular disease in this case series is quite striking and has not been previously reported.

Pulmonary hypertension is a complication of sickle cell disease receiving increasing recognition [9]. Defined by tricuspid regurgitant jet velocity ≥ 2.5 m/s, pulmonary hypertension has been found in 32% of adults with sickle cell disease [7]. In our prospective screening study of pulmonary hypertension, it is the principal clinical factor associated with early mortality. Risk factors for pulmonary hypertension have included a history of priapism, renal and liver disease, and markers of accelerated hemolysis during steady state compared to other patients with sickle cell disease, including low hemoglobin and high LDH levels.

All six patients with concurrent cerebrovascular disease and pulmonary hypertension had more severe than average hemolysis. This is indicated by significantly higher than average LDH levels at steady state, falling in the top quartile of the comparison group. Serum LDH in patients with sickle cell disease is directly related to severity of intravascular hemolysis [10,11]. These six patients exemplify a new mechanism of vascular disease associated with pulmonary hypertension in patients with sickle cell disease [12]. In this model, serum LDH is a surrogate marker for the concurrent release of hemoglobin into blood plasma, where it inactivates nitric oxide, resulting in systemic and pulmonary vasoconstriction, accompanied by endothelial activation and dysfunction [13]. A growing body of epidemiological, animal, and laboratory data support this model [14–20]. A pattern of vasculopathy featuring pulmonary hypertension, lower extremity cutaneous ulceration, and priapism is seen in sickle cell disease, particularly in those patients with the most severe hemolysis [7,10,19]. However, data have not yet been presented to associate this mechanism with cerebrovascular disease in sickle cell disease. Based on the recent literature cited above, these six patients with particularly severe intravascular hemolysis would be predicted to have low nitric oxide bioavailability due to consumption by cell-free plasma hemoglobin.

These patients also likely have diminished production of nitric oxide. LDH is also a marker for the release of red cell arginase, which depletes the plasma of L-arginine, the substrate required for nitric oxide production [21,22]. In agreement with this, the patients in this series have relatively low plasma levels of arginine. Five of the six patients have a lower than average arginine to ornithine ratio, which has been associated with increased plasma arginase activity, pulmonary hypertension, and early mortality in sickle cell disease [7,23,24]. The six patients also have relatively high proline levels, a finding that also has been associated with high arginase activity [25]. Thus, LDH may be a marker of at least two biologically linked processes that cause impaired nitric oxide bioavailability: accelerated destruction of nitric oxide by cell-free plasma hemoglobin and diminished production of nitric oxide due to arginine depletion by plasma arginase [10].

Pulmonary hypertension and stroke have many clinical features in common among patients with sickle cell disease. The Cooperative Study of Sickle Cell Disease identified five independent risk factors for stroke in patients with sickle cell disease: prior TIA, acute chest syndrome within 2 weeks prior to the stroke, higher annual rates of the acute chest syndrome, low steady-state hemoglobin, and higher systolic blood pressure [2]. It is intriguing that the latter two are also risk factors for the development of pulmonary hypertension. There are a number of other epidemiological observations in common between ischemic stroke and pulmonary hypertension (Table II).

TABLE II. Overlap of Risk Factors, Pathology and Possible Pathophysiological Mechanisms for Ischemic Stroke and Pulmonary Hypertension in Patients with Sickle Cell Disease.

Factors	Ischemic stroke	Pulmonary hypertension
Medical history
Prior stroke	Increased risk [1]	Increased risk [7]
TIA	Increased risk [2]	Not determined
Recent or recurrent ACS	Increased risk [2]	No association [7]
Dactylitis	Increased risk [33]	Not determined
Clinical measurements
Systemic systolic HTN	Increased risk [2]	Increased risk [7]
Low oxygen saturation	Increased risk [34]	Increased risk [7]
Ultrasound blood flow assessment	Increased risk with high TCD velocities [35]	Increased risk with high TR velocities [7]
Laboratory markers
Anemia	Increased risk [2,33,36,37]	Increased risk [7]
High LDH, Reticulocyte count	Seen in this case report	Increased risk [7]
Leukocytosis	Increased risk [2,33]	No association [7]
Fetal hemoglobin	Not protective [2,38,39]	Not protective [7]
Low Arginine/Ornithine ratio	Seen in this case report	Increased risk [7] [22]
Pathology
Smooth muscle hyperplasia with overlying endothelial damage and fibrosis in large arterial vessels	Cerebral arteries [26–29]	Pulmonary arteries [30,31]
Thrombosis in situ	Cerebral arteries [29,40]	Pulmonary arteries [30,31]
Pathological pathways
Cell adhesion	Risk associated with VCAM-1 allelic variants [41]	Increased risk with elevation of plasma soluble VCAM-1, ICAM-1, E-selectin [42]
Impaired nitric oxide signaling	Indirect evidence in animal model of SCD cerebral blood flow [43]	Biochemical and physiological evidence [6,44]

There also is considerable overlap in the histopathology of pulmonary hypertension and ischemic stroke in patients with sickle cell disease. In autopsy studies on patients with sickle cell disease, large vessel intimal hyperplasia, distorted endothelium, and in situ thrombosis have been seen in arteriopathy associated both with stroke [26–29] and pulmonary hypertension [30,31]. Although sickle cell disease has generally been considered a microvascular disease, it is clear that pulmonary hypertension and cerebrovascular disease in patients with sickle cell disease share a proclivity for large-vessel arteriopathy.

This case series suggests a clinical association between cerebrovascular disease and pulmonary hypertension in patients with sickle cell disease. We propose that this link might be best investigated by a prospective case control trial, comparing the prevalence of cerebrovascular disease determined by magnetic resonance imaging and angiography in sickle cell patients with or without pulmonary hypertension, matched for age, gender, and phenotype.

The similarities in epidemiology, physiology, and histopathology between the cerebral vasculopathy and pulmonary vasculopathy are remarkable: both are characterized by intimal and smooth muscle proliferative vasculopathy in a large vessel distribution, both by vasoconstriction, anemia, lack of association with vaso-occlusive crisis frequency, and lack of protection afforded by fetal hemoglobin. These factors are summarized in Table II. We suggest that severe hemolysis in patients with sickle cell disease, marked by low hemoglobin levels, high reticulocyte counts, and high LDH levels, may predispose to ischemic stroke and pulmonary hypertension though impairment of nitric-oxide-mediated blood flow. Since nitric oxide has antiproliferative properties in vascular cells, its deficiency is associated with the progressive development of a proliferative, obliterative vasculopathy [32]. This is a model that merits further clinical investigation.