Pulmonary Complications of Hematologic Diseases


Hematologic diseases and their specific therapies can adversely affect several aspects of cardiopulmonary function, by reducing the oxygen-carrying capacity of the blood, by impairing pulmonary vascular function and pulmonary immune defenses, and by direct pulmonary parenchymal damage. This chapter reviews the clinical manifestations, epidemiology, pathophysiology, and treatment of pulmonary complications of disorders of the hematologic system.

Red Blood Cell Disorders


Anemia is defined as a reduction in the number of circulating red blood cells. By decreasing oxygen-carrying capacity of the blood, anemia can impair cardiopulmonary function. However, because multiple compensatory mechanisms exist to adjust to a reduced oxygen-carrying capacity of blood, the signs and symptoms induced by anemia depend on the degree of anemia, the rate at which it evolves, the oxygen demands of the patient, and the presence of chronic cardiopulmonary disease. For instance, in resting adults subjected to acute isovolemic anemia, oxygen delivery can be maintained at hemoglobin concentrations as low as 5 g/dL, a finding also seen in individuals with chronic severe anemia. As such, anemia, even when severe, rarely causes heart failure or pulmonary edema and, when it does, it is likely that heart failure from chronically high cardiac output is superimposed on some other coexisting cardiac abnormality. From the hemodynamic standpoint, as the hemoglobin level decreases (particularly with hemoglobin values < 7 g/dL), cardiac output increases, filling pressures tend to decrease, and systemic and pulmonary vascular resistances decrease. Further, these changes are readily reversible after red blood cell transfusions. Finally, when anemia is severe, in the absence of coexisting cardiopulmonary disease, gas exchange is usually well maintained.

Pulmonary hypertension (PH) is an increasingly recognized complication of chronic hereditary and acquired hemolytic anemias. Remarkably, virtually every cause of hemolytic anemia has been associated with PH ( Table 94-1 ). The PH associated with hemolytic disorders is now considered in the Group 5 classification of pulmonary hypertension, characterized by unclear/multifactorial mechanisms (see Chapter 58 ). In contrast, there have been no reports of PH associated with nonhemolytic anemias such as the anemia of chronic disease or iron deficiency anemia. This suggests that the hemolytic component of anemia is necessary for the development of PH, which is discussed in detail in the section on hemoglobinopathies .

Table 94-1

Hemolytic Disorders Associated with Pulmonary Hypertension

  • Hemoglobinopathies

    • Sickle cell disease

    • Thalassemia intermedia and major

    • Hb Mainz unstable hemoglobin hemolytic anemia

  • Red cell membranopathies

    • Hereditary spherocytosis

    • Hereditary stomatocytosis

    • Paroxysmal nocturnal hemoglobinuria

    • Alloimmune hemolytic anemia

  • Red cell enzymopathies

    • Pyruvate kinase deficiency

    • Glucose-6-phosphate dehydrogenase deficiency

  • Microangiopathic hemolytic anemia

    • Thrombotic thrombocytopenic purpura

    • Hemolytic uremic syndrome

    • Hemolysis from mechanical heart valves

    • Left ventricular assist devices and cardiopulmonary bypass procedures



Polycythemia is characterized by an abnormally high hematocrit (>48% and >52% in women and men, respectively), hemoglobin concentration (>16.5 or >18.5 g/dL in women and men, respectively), or red blood cell count. These measurements are dependent on the plasma volume as well as the red blood cell mass; thus, before classifying a patient as having true polycythemia, one must exclude a decrease in plasma volume as the etiology for these abnormalities. A state of chronically reduced plasma volume with elevated hemoglobin or hematocrit has been called Gaisböck disease, spurious polycythemia, stress erythrocytosis, apparent polycythemia, and pseudopolycythemia and has been associated with use of diuretics, alcohol, obesity, hypertension, and renal disease. On the other hand, patients with true elevations in red blood cell mass have absolute polycythemia and are categorized in primary and secondary forms. Primary polycythemia is caused by an acquired or inherited mutation such as in the JAK2 gene, leading to an abnormality within red blood cell progenitor cells; it includes polycythemia vera and rare familial variants (e.g., activating mutations of the erythropoietin receptor or Chuvash polycythemia). Secondary polycythemia is caused by a circulating factor stimulating erythropoiesis, usually related to the production of erythropoietin as a physiologic response to chronic hypoxia (e.g., chronic hypoxic lung disorders and disorders affecting hemoglobin oxygen affinity), but can also result from erythropoietin-secreting tumors.

Most of the symptoms of polycythemia are related to an increase in blood viscosity, which leads to impairment in systemic and pulmonary blood flow. Thromboembolic events have been described in approximately 30% of patients with polycythemia vera and account for 31% of their deaths. Pulmonary nodular lesions secondary to thrombosis, stasis, or infarction have been described in these patients. Late-stage polycythemia vera, which is characterized by myelofibrosis, may cause pulmonary and pleural masses that are related to extramedullary hematopoiesis.

PH has been reported in patients with polycythemia and other myeloproliferative disorders (recently reviewed by Machado and Farber). In a retrospective review, 26 patients—12 with myeloid metaplasia, 5 with essential thrombocythemia, 6 with polycythemia vera, 2 with myelodysplastic syndrome, and 1 with chronic myeloid leukemia—had echocardiographic or hemodynamic evidence of PH. Interestingly, the median survival after the diagnosis of PH was 18 months and the majority of deaths were related to cardiopulmonary causes, suggesting that the presence of PH was a direct cause of death. In a subsequent retrospective study of 10 patients with PH and myeloproliferative disorders (8 with polycythemia vera and 2 with essential thrombocytosis), 6 patients had chronic, thromboembolic PH and 4 had unexplained PH; these findings were most likely associated with the myeloproliferative disorder because other risk factors for PH could not be identified. In a cohort of 14 patients with Chuvash polycythemia, the prevalence of PH (defined as an estimated pulmonary artery systolic pressure ≥ 35 mm Hg) was 36%. The pathogenesis of PH in these disorders is not well established but can involve chronic thromboembolic disease or hypercoagulability and hyperviscosity leading to in situ throm­bosis, as well as hemoglobin-dependent scavenging of endothelium-derived nitric oxide (NO) and up-regulation of HIF-1α–mediated pathways, such as endothelin-1.

In patients with polycythemia vera, the mainstay of therapy is phlebotomy and myelosuppression with hydroxyurea, to keep the hematocrit below 42% in women and 45% in men, and antithrombotic therapy with aspirin or anagrelide. In case series, treatment of the myeloproliferative disorder has been shown to improve PH. In patients with chronic hypoxic disorders such as chronic obstructive pulmonary disease (COPD) and interstitial lung disease, long-term oxygen therapy should reduce the number of patients who become severely polycythemic. In patients with COPD and severe polycythemia, phlebotomy—to achieve a hematocrit of about 50%—is associated with both a decrease in mean pulmonary artery pressure and in pulmonary vascular resistance, as well as an improvement in exercise performance. However, the use of phlebotomy should be reserved as adjunctive therapy in the management of markedly polycythemic and symptomatic patients who remain significantly polycythemic despite appropriate long-term oxygen therapy.


Sickle Cell Disease

Sickle cell anemia, the most common and most severe form of sickle cell disease, is seen in individuals who are homozygous for a single nucleotide substitution in the β globin gene. This results in the synthesis of hemoglobin S (Hb S), a structural variant that, when deoxygenated, is much less soluble than normal hemoglobin (Hb A). Deoxygenated Hb S polymerizes and aggregates inside sickled erythrocytes as they traverse the microcirculation. Rigid, dense, and sickled cells can become entrapped in the microcirculation, a process that is enhanced by their increased propensity to adhere to endothelium. Mechanistic studies in transgenic mice expressing exclusively human Hb S suggest that microvascular occlusion results in ischemia and reperfusion injury, which promotes inflammatory, thrombotic, and oxidant stress. In patients with sickle cell disease, vaso-occlusion leads to the frequent episodes of bone pain and acute chest syndrome that complicate sickle cell disease. Furthermore, the membrane of erythrocytes containing intracellular Hb S polymer is constantly exposed to mechanical and oxidant injury as the erythrocytes traverse the microcirculation. Ultimately, cumulative membrane damage shortens red cell life span so that sickle cell disease is characterized by a chronic hemolytic anemia. Intravascular hemolysis releases cell-free hemoglobin into the plasma, which scavenges NO and releases red blood cell arginase into the plasma, which catabolizes arginine, the substrate for NO synthesis. Intravascular hemolysis thus produces a state of endothelial dysfunction, vascular proliferation, and pro-oxidant and proinflammatory stress.

It is estimated that around 250,000 children worldwide are born with homozygous sickle cell anemia every year. Approximately 0.15% of African-Americans are homozygous for sickle cell disease, and 8% have the sickle cell trait. Despite significant improvements in the life expectancy of patients with sickle cell disease, estimates of the median age at death range from 48 to 58.5 years for women and 42 to 53 years for men.

Acute and chronic pulmonary complications of sickle cell disease are common but often underappreciated by health care providers. Acute complications include asthma and the acute chest syndrome (ACS), while chronic complications include pulmonary fibrosis, PH, and cor pulmonale. Pulmonary complications account for a large proportion of deaths among adults with sickle cell disease. According to the Cooperative Study of Sickle Cell Disease (CSSCD), a prospective multicenter study of 3764 patients, more than 20% of adults presumably had fatal pulmonary complications of sickle cell disease. Among the 299 patients enrolled in the long-term follow-up study of patients who participated in the Multicenter Study of Hydroxyurea in Sickle Cell Anemia (MSH), pulmonary disease was the most common cause of mortality, accounting for 28% of all deaths.

Acute Chest Syndrome

ACS represents a lung injury syndrome in patients with sickle cell disease that is related to the increased membrane permeability that characterizes the acute respiratory distress syndrome (ARDS). In a patient with sickle cell disease, ACS is clinically defined by the development of a new pulmonary opacity involving at least one complete lung segment, consistent with alveolar consolidation, not atelectasis, and accompanied by chest pain, fever, tachypnea, wheezing, or cough.


ACS is the second most common cause of hospitalization in patients with sickle cell disease and the leading cause of both admission to an intensive care unit and premature death. More recently, however, increased awareness, the chronic use of hydroxyurea, and the early and aggressive use of transfusion therapy appear to have decreased ACS-related morbidity and mortality, as evidenced by a lower rate of ACS in a recently completed multicenter trial of inhaled NO for vaso-occlusive crises.

ACS can develop in any of the sickle hemoglobinopathies but is more common in individuals with homozygous sickle cell disease (Hb SS). In the CSSCD, there was a 29% incidence of the ACS in 3751 subjects over a 2-year period, representing an attack rate of 12.8 episodes per 100 patient-years for Hb SS disease. The incidence was higher in children than in adults (24.5 events vs. 8.8 events per 100 patient-years). Up to half of ACS episodes develop in association with vaso-occlusive pain crisis, and a sizable proportion of patients will have a painful event within 2 weeks of the diagnosis. Because up to 20% of patients admitted with acute vaso-occlusive pain crisis will develop ACS in the first 3 days of hospitalization, physicians must be vigilant for this common and potentially lethal complication.

Steady-state laboratory parameters associated with an increased risk for the development of ACS include an elevated white blood cell count, a higher steady-state hemoglobin level, and a lower steady-state fetal hemoglobin level. In children, a number of studies now suggest that asthma is a risk factor for the development of ACS. Other clinical events that appear to increase the risk of (or are associated with) the development of ACS include major surgical procedures, acute rib infarcts, avascular necrosis of the hips, pregnancy, use of narcotics, acute anemic events, and previous pulmonary events.

During acute hospitalization for vaso-occlusive crisis, the development of ACS is often preceded by an abrupt drop in hemoglobin levels (mean decrease of 0.78 g/dL from steady-state values) and increases in markers of hemolysis, such as lactate dehydrogenase (LDH). ACS may also be preceded by a drop in the platelet count, and levels less than 200,000 per µL constitute an independent risk factor for ACS severity, associated with increased risk of neurologic complications and mechanical ventilation. In our units, we monitor trends in hemoglobin and platelet counts and pay careful attention to patients with drops in steady-state values.


Three major mechanisms appear to be involved in the pathogenesis of ACS: infection, bone marrow fat embolization, and direct red cell intravascular sequestration causing lung injury and infarction ( Table 94-2 , Fig. 94-1 ).

Table 94-2

Causes of 670 Episodes of Acute Chest Syndrome *

0–9 yr
( n = 329)
10–19 yr
( n = 188)
≥20 yr
( n = 153)
Fat embolism, with or without infection 59 (8.8) 24 16 19
Chlamydia 48 (7.2) 19 15 14
Mycoplasma § 44 (6.6) 29 7 8
Virus 43 (6.4) 36 5 2
Bacteria 30 (4.5) 13 15 12
Mixed infections 25 (3.7) 16 6 3
Legionella 4 (0.6) 3 0 1
Miscellaneous infections 3 (0.4) 0 3 0
Infarction 108 (16.1) 50 43 15
Unknown # 306 (45.7) 139 88 79

Reproduced with permission from Vichinsky EP, Neumayr LD, Earles AN, et al: Causes and outcomes of the acute chest syndrome in sickle cell disease. National Acute Chest Syndrome Study Group. N Engl J Med 342:1855–1865, 2000, Table 4.

* Data on one episode were excluded because the patient’s birth date was not known.

Nineteen of the episodes of pulmonary fat embolism were associated with infectious pathogens.

This category included episodes in which Chlamydia alone was identified but not episodes involving mixed infections or pulmonary fat embolism.

§ This category included only episodes in which Mycoplasma pneumoniae or Mycoplasma hominis was identified, but not episodes involving mixed infections, Mycobacterium tuberculosis, or pulmonary fat embolism.

This category included two cases of tuberculosis and one case of Mycobacterium avium complex infection.

A pulmonary infarction was presumed when the results of the analysis for pulmonary fat embolism, bacterial studies, viral isolation studies, and serologic tests were complete and were all negative.

# The cause of episodes for which some or all of the diagnostic data were incomplete and no etiologic agent was identified was considered to be unknown.

Figure 94-1

Pathogenesis of the acute chest syndrome.

Three major triggers are associated with the development of acute chest syndrome (ACS): infection, bone marrow fat embolization, and direct red cell intravascular sequestration causing lung injury and infarction. Lung injury results in ventilation-perfusion mismatch/shunt and hypoxemia, which leads to increased hemoglobin S polymerization, and erythrocyte vaso-occlusion. This worsens bone marrow infarction and pulmonary vaso-occlusion to promote a vicious cycle. Fat embolization can be diagnosed by Oil Red O staining of pulmonary alveolar macrophages, revealing the characteristic red lipid inclusions, as shown in the panel. Common infectious organisms and other causes of ACS are listed in Table 94-2 . NO, nitric oxide.

The most common etiology of ACS in both children and adults is infection by a community-acquired pathogen ( eFig. 94-1 ), followed by an excessive inflammatory lung injury response. More than 80% of adults with sickle cell disease report a history of having been admitted to the hospital for “pneumonia.” The National Acute Chest Syndrome Study Group analyzed 670 episodes of ACS in 538 patients with sickle cell disease to determine the cause, outcome, and response to therapy, and respiratory samples obtained from sputum and bronchoalveolar lavage were analyzed for viral and bacterial infections. Among the infectious agents identified most frequently were atypical bacteria and viruses, including Chlamydia pneumoniae (29%), Mycoplasma pneumoniae (20%), Legionella pneumophila (2%), respiratory syncytial virus (10%), parvovirus (4%), rhinovirus (3%), parainfluenza virus (2%), influenza A virus (2%), cytomegalovirus (2%), Epstein-Barr virus (1%), and herpes simplex virus (1%). Community-acquired encapsulated bacteria were rarely isolated, despite the fact that patients with Hb SS disease rarely have normal splenic function. Staphylococcus aureus was isolated in 5% of cases and Streptococcus pneumoniae in only 4% of cases. Cases of severe ACS related to outbreaks of seasonal influenza have also been described.

Fat embolization syndrome is the second most common cause of the ACS. It arises as a complication of vaso-occlusive pain crisis involving multiple bones, which results in bone marrow edema, infarction, and necrosis. As a consequence, bone marrow contents are released into the systemic circulation and trapped in the pulmonary circulation, producing acute PH, severe lung inflammation, and hypoxemia ( Fig. 94-2 ) ( eFig. 94-2 ). Bone marrow fat released into the bloodstream is also converted by secretory phospholipase A 2 to free fatty acids, which can produce direct inflammatory lung injury.

Figure 94-2

Fat embolization in acute chest syndrome.

A, Chest CT scan of a patient with acute chest syndrome and fat embolization syndrome. B, Postmortem examination specimen of a patient who died suddenly during an episode of vaso-occlusive crisis and acute chest syndrome demonstrating bone marrow elements ( arrows ) lodged in the small pulmonary artery.

Finally, in approximately 20% of patients, direct lung infarction or vaso-occlusion is associated with the development of ACS, with a small percentage of patients actually developing triangular-shaped pulmonary infarction ( eFig. 94-3 ), sometimes followed by central cavitation. Direct pulmonary arterial in situ thrombosis is also seen in patients with ACS. A French study evaluated the presence of pulmonary artery thrombosis by computed tomography (CT)–pulmonary angiography in 125 consecutive patients with 144 episodes of ACS and noted a 17% prevalence of subsegmental thromboembolism without evidence of peripheral venous thrombosis in any of their cases.

A potential role for hemolysis-derived plasma-free hemoglobin and free heme as novel mechanisms in ACS is emerging. Ghosh and colleagues administered lysed red blood cells into the circulation of sickle cell mice, which resulted in increased pulmonary vascular permeability without affecting permeability in other organs; moreover, preliminary studies from the same group have demonstrated that intravenous administration of heme to sickle cell mice induces severe and lethal acute lung injury. These data suggest that plasma-free hemoglobin and/or heme specifically may contribute directly to lung injury. The relationship between increased intravascular hemolysis and thrombocytopenia indicates that a possible thrombotic thrombocytopenic purpura (TTP)–like mechanism may exist in a subset of patients with ACS as suggested by studies demonstrating that free hemoglobin may inhibit activity of ADAMTS13, a protease whose deficiency underlies most cases of TTP.

Clinical Features and Evaluation.

The clinical features at presentation are age dependent, which likely reflects the different etiologies of ACS in different age groups, with children having a higher proportion of infectious etiologies in comparison with adults, who tend to have fat embolization as a major etiology. Overall, 80% of patients present with fever, 62% with cough, and approximately 40% have chest pain, tachypnea, dyspnea, and abdominal, arm, leg, rib, or sternal pain. Most adult patients present with severe extremity or chest pain and 24 to 72 hours later develop ACS. Reactive airway disease is observed in 13% of cases of ACS and is much more common in children.

ACS is associated with signs of systemic inflammation, with mean peak temperatures of 38.9° C and mean white blood cell counts of 23,000 cells/µL 3 . As mentioned earlier, although a high steady-state hemoglobin level is a major risk factor for developing ACS, the acute presentation is often associated with a drop in hemoglobin levels (mean decrease of 0.78 g/dL from steady-state levels) and increases in markers of hemolysis. A platelet count less than 200,000 per µL appears to be a marker of ACS severity, associated with increased risk of neurologic complications and mechanical ventilation. Secretory phospholipase A 2 levels are elevated early in the course of ACS, even before the development of radiographic changes, and have been used to predict the onset of the syndrome.

Because the clinical manifestations can be indistinguishable from the other causes of ACS, the diagnosis of pulmonary fat embolization syndrome relies on the identification of Oil Red O-positive lipid accumulations within alveolar macrophages (see Fig. 94-1 ). In a 30-center clinical trial, the National Acute Chest Syndrome Study Group identified fat embolization syndrome in 16% of ACS cases in adults and children on the basis of positive lipid accumulations in alveolar macrophages. Traditionally bronchoscopy has been used as the diagnostic modality of choice for the diagnosis of pulmonary fat embolization syndrome. However, induced sputum may be a noninvasive alternative; in one study, induced sputum sampling of alveolar macrophages was found to have a similar yield as samples obtained from bronchoalveolar lavage, with a modest but significant correlation between the two measurements (r = 0.65).

Some patients with ACS manifest evidence of systemic fat embolization, also called the acute multiorgan failure syndrome. This syndrome should be suspected in patients presenting with acute multiorgan failure characterized by the development of acute hypoxic respiratory failure, right heart failure, renal and hepatic dysfunction, alterations in mental status, seizures, thrombocytopenia, and coagulopathy. In patients with ACS, those with lipid-laden macrophages in induced sputum have significantly more extrathoracic pain, evidence of fat emboli, more neurologic symptoms, a lower platelet count, and higher transaminase levels than do those without lipid-laden macrophages. This suggests that the fat embolization syndrome is both a major cause of ACS and has a more severe course with systemic complications.

The mean length of hospitalization for ACS is 10.5 days, compared with 3 to 4 days for uncomplicated vaso-occlusive pain crisis. Thirteen percent of all patients with ACS require mechanical ventilation, and the overall mortality is 3% for all patients and 9% for adults. Risk factors for mechanical ventilation and poor outcome include a platelet count less than 200,000 per µL (likely indicative of the fat embolization syndrome), a larger number of lobes involved on chest radiograph, and a self-reported or medical record history of cardiac disease. The latter complication is now thought to represent occult PH and cor pulmonale. In fact, in a study of 84 consecutive hospitalized patients with ACS, 13% of patients manifested right heart failure, a subgroup that had the highest risk for mechanical ventilation and death. Thus, PH and right heart dysfunction likely represent a major comorbidity during ACS, and right heart failure should be considered in patients presenting with shock or severe hypoxemia. Interestingly, despite these issues, the outcome of patients with severe ACS on mechanical ventilation is better than that of patients with ARDS, with a mortality rate of 19% in contrast to the approximate 30% mortality rate reported in current ARDS studies.


In the outpatient setting, patients with a history of ACS should be treated with hydroxyurea because hydroxyurea has been shown to reduce the risk of developing ACS by approximately 50%. A chronic transfusion regimen is also effective in reducing the incidence of ACS. Because the triggers and risk factors for ACS are well known, clinical surveillance plus aggressive and early therapy of this patient population are likely to improve prognosis. This could be accomplished by close monitoring of patients during a vaso-occlusive crisis, during a febrile illness, and in the postoperative state.

If ACS develops, a range of treatment strategies is recommended, as outlined in Table 94-3 . Oxygen therapy should be given routinely to maintain oxygen saturation above 92%. Aggressive pain management and incentive spirometry can minimize chest wall splinting, with consequent relief of atelectasis and alveolar hypoxia. In fact, the use of incentive spirometry has been shown to decrease the incidence of new pulmonary opacities in patients admitted with vaso-occlusive pain affecting the chest wall.

Table 94-3

Treatment of the Acute Chest Syndrome (ACS)

Oxygen therapy to maintain arterial hemoglobin oxygen saturation above 92%
Pain control and incentive spirometry to reduce chest wall splinting and pulmonary atelectasis
Close clinical observation
Monitor P o 2 /F io 2 ratio:
Particular attention to diagnosis of worsening respiratory function
Asthma therapy if indicated
Empirical antibiotics
Cover typical and atypical respiratory pathogens.
Consider the regional and seasonal risk of methicillin-resistant Staphylococcus aureus.
Anticipate influenza A or B infections and treat/prevent accordingly.
Transfusion therapy
Main indication for transfusion therapy in ACS is worsening respiratory function.
Simple transfusion is as effective as erythrocytapheresis in the usual patient.
Patients with high initial hemoglobin concentrations (≥9 g/dL) or patients with more severe disease should receive erythrocytapheresis.
Transfused blood should be matched to Rh, C, E, and Kell antigens, and transfusion records documenting history of prior alloantibodies should be obtained.

Given the high prevalence of infectious etiologies for ACS, we recommend the use of empirical antimicrobial therapy in all patients. Considering the high prevalence of atypical bacteria and encapsulated organisms, empirical coverage should include agents, such as the macrolides or fluoroquinolones, effective against these organisms. It is also important to consider alternative organisms such as methicillin-resistant S. aureus or influenza viruses, especially in patients not responding to therapy or during the annual influenza season.

Blood transfusion remains the mainstay of ACS therapy, although its value has not yet been demonstrated in randomized trials. Acute red cell transfusion increases arterial P o 2 and hemoglobin oxygen saturation and may rapidly resolve the pulmonary event. Transfusion can also increase blood viscosity and the consequent risk of vaso-occlusion, so it is recommended that the hemoglobin level not rise above 11 g/dL. The National Acute Chest Syndrome Study Group found no significant differences in outcomes between patients treated with simple transfusion or with erythrocytapheresis, a red cell exchange, suggesting that simple transfusion is preferred as initial therapy. For patients with high initial hemoglobin concentrations (>9 g/dL) or with more severe disease, however, erythrocytapheresis is recommended. However, because most patients have a significant decrease in hemoglobin at presentation, the transfusion of 2 to 4 units of packed red cells over 24 to 48 hours can usually be performed without complication. In order to decrease the risk of delayed hemolytic transfusion reactions related to alloimmunization against minor red blood cell antigens, all transfused blood should be matched to Rh, C, E, and Kell antigens.

Some potentially promising treatments have not proven to have a role in ACS. Although treatment with corticosteroids has been shown to reduce the severity of pain and length of hospitalization, this therapy is complicated by a high rate of rebound pain and hospitalization. A study evaluating a slow tapering protocol to maintain the beneficial effects of corticosteroids while limiting rebound pain and re-admission was terminated early because of slow accrual. The placebo-controlled study by Gladwin and coworkers evaluated the role of inhaled NO therapy for patients presenting in vaso-occlusive crisis and did not demonstrate an effect of inhaled NO compared with placebo on the duration of pain crisis, narcotic use, pain scores, or the development of ACS. In a prospective, randomized, open, single-center study of 67 adult patients with ACS, noninvasive mechanical ventilation (NIV) use improved respiratory rate and gas exchange but did not reduce the number of patients remaining hypoxemic at day 3 and was associated with greater patient discomfort. Additionally, NIV did not change transfusion rates, pain scores, narcotic dose, or hospital length of stay but instead prolonged length of stay in the step-down unit.

Pulmonary Hypertension

PH has emerged as a major threat to the well-being and longevity of patients with sickle cell disease. PH is defined hemodynamically by a mean pulmonary artery pressure ( ) greater than or equal to 25 mm Hg. Sickle cell disease could represent one of the most common causes of PH. Because there are 30 million individuals worldwide with sickle cell disease, of whom 10% to 30% may have PH, there may be as many as 3 to 9 million patients with the complication.


A variety of studies, both retrospective and prospective, have identified a high prevalence of PH in patients with sickle cell disease. Retrospective studies have reported that 20% to 30% of patients with sickle cell disease have an elevated pulmonary artery systolic pressure estimated noninvasively by Doppler echocardiography by a tricuspid regurgitant jet velocity (TRV) that is 2 standard deviations above the normal mean value (≥2.5 m/sec). Autopsy studies suggest that up to 75% of sickle cell patients have histologic evidence of PH at the time of death ( Fig. 94-3 ).

Figure 94-3

Pulmonary arteriopathy in sickle cell–related pulmonary hypertension.

A, Low-power photomicrograph demonstrating pulmonary arterial smooth muscle hypertrophy (hematoxylin and eosin stain). B, Higher-power photomicrograph of one of the arteries showing a plexogenic lesion, with medial thickening, laminar intimal hyperplasia, recanalization, and fibrosis (hematoxylin and eosin stain).

These data are now corroborated by three prospective studies. In the National Institutes of Health (NIH) PH echocardiographic screening study, 23% of patients with sickle cell disease had borderline to mild elevations in pulmonary artery systolic pressures (defined by a TRV of > 2.5 to 2.9 m/sec, which corresponds to a pulmonary artery systolic pressure of 30 to 39 mm Hg) and 9% had moderately to severely elevated pressures (defined by a TRV > 3.0 m/sec, which corresponds to a pulmonary artery systolic pressure of approximately 40 to 45 mm Hg). Similar rates were found in echocardiographic screening studies performed at two other centers.

Measurement of N-terminal pro-brain natriuretic peptide (NT-proBNP), a prohormone released by the right and left ventricular myocardium under pressure stress, in stored plasma samples has shown a high prevalence of abnormal values that may correlate with PH. Of those enrolled in the MSH in 1996, 30% of individuals had elevated levels suggesting the possible presence of PH. Similarly, among patients enrolled in the CSSCD from 1978 to 1988, 27.6% of adults had elevated levels of NT-proBNP. In both studies, elevated NT-proBNP was independently associated with a higher risk of death.

PH has been shown to increase over time and with increasing age. Two centers independently reported the follow-up of adult sickle patients who on initial echocardiographic screening had normal TRVs. After 2 to 3 years of follow-up, 13% to 15% of these patients developed high TRVs, suggesting an increasing incidence of PH of about 4% to 7% per year. Increasing age is associated with an increased risk of elevated TRV. For example, in the NIH PH echocardiographic screening study, patients with an elevated TRV were significantly older than patients without (38 ± 19 years for patients with TRV > 3.0 m/sec and 39 ± 12 years for patients with TRV 2.5 to 2.9 m/sec, compared with 34 ± 10 years for patients with TRV < 2.5 m/sec; P = 0.02). An increasing number of studies suggests that PH is developing in children with sickle cell disease; however, few children have TRV values greater than 3.0 m/sec, and the implications in terms of functional capacity and associated mortality have not been determined.

Epidemiologic risk factors associated with PH include a history of renal or cardiovascular complications, increased systemic systolic blood pressure, elevations in the marker of hemolysis (LDH), elevated alkaline phosphatase, and low transferrin levels. In men, a history of priapism was also an independent factor associated with PH. These risk factors have also been observed in recently published studies using right heart catheterization to define PH. Interestingly, the development of PH was not associated with the number of vaso-occlusive episodes, markers of inflammation, fetal hemoglobin levels, or platelet counts. Although a high pulmonary artery pressure could also result from the high cardiac output state associated with chronic anemia, this hyperdynamic state does not seem to be a major contributor to significant elevations in pulmonary artery pressures because there has been no report of PH associated with nonhemolytic anemia. In sum, PH represents a component of the systemic vasculopathy of sickle cell disease (characterized by systemic hypertension, renal failure, and priapism).

Finally, three studies have provided new insights into the prevalence of PH in sickle cell disease using the gold standard diagnostic test for the disease, right heart catheterization. In the NIH screening study (2001–2010, median follow-up of 4.4 years), 86 of the 533 subjects underwent right heart catheterization and, of these, 56 (10.5%) were diagnosed with PH. Similarly, in a screening study of 80 patients from Brazil, 32 (40%) had an elevated TRV and, of those who underwent right heart catheterization, 8 (10% of the total population) had PH. In a third large screening study of 398 patients with sickle cell disease in France using right heart catheterization, 6% were shown to have PH. Of note, the French study excluded approximately 10% of patients, those with “severe” renal, liver, or lung disease, with severity defined as a creatinine clearance of less than 30 mL/min, an abnormal prothrombin time (international normalized ratio > 1.7), and chronic restrictive lung disease defined by a TLC of less than 70% predicted. It is not clear why these patients would be excluded from a prevalence study of sickle cell disease–related PH, especially considering the fact that all of these complications develop as a direct consequence of sickle cell disease and all three represent significant published risk factors for developing PH in sickle cell disease.


An elevated pulmonary artery pressure increases the risk of death for patients with sickle cell disease. In the NIH study, compared with patients with TRV less than 2.5 m/sec, the rate ratios for death for TRVs of 2.5 to 2.9 m/sec and greater than 3.0 m/sec were 4.4 and 10.6, respectively. In support of these findings, De Castro and colleagues found that 6 of 42 patients (14%) with PH and only 2 of 83 patients (2%) without this finding died during a 2-year follow-up period. Similarly, in the study by Ataga and colleagues, 9 of 36 patients with PH and only 1 of 57 patients without PH died during the 2.5-year follow-up period (relative risk, 9.3). Consistent with these data, in a cohort of 632 patients with SCD from the United States and England, 11.2% had TRV ≥ 3.0 m/sec and 24.1% had NT-proBNP level ≥ 160 pg/mL. Of 22 deaths during follow-up, 50% had a TRV ≥ 3.0 m/sec. At 24 months, the cumulative survival was 83% with TRV ≥ 3.0 m/sec and 98% with TRV < 3.0 m/sec. The hrs for death were 11.1 (95% CI 4.1-30.1; p < 0.0001) for TRV ≥ 3.0 m/sec, 4.6 (1.8-11.3; p = 0.001) for NT-proBNP ≥ 160 pg/mL, and 14.9 (5.5-39.9; p < 0.0001) for both TRV ≥ 3.0 m/sec and NT-proBNP ≥ 160 pg/mL.

When documented by right heart catheterization, the presence of PH is a major risk factor for death in patients with sickle cell disease. Castro and colleagues reported a 50% two-year mortality rate in patients with PH; with each increase of 10 mm Hg in , there was a 1.7-fold increase in the rate of death. In the NIH study, the mortality rate was significantly higher in the PH group (20 deaths, 36%) than in either the group without PH by right heart catheterization (3 deaths, 10%,) or the general sickle cell group with normal Doppler echocardiographic estimates of pulmonary artery systolic pressure (50 deaths, 13%). Similarly, in both the Brazilian and French cohorts, the mortality rate was significantly higher in the PH group (38% and 23%, respectively) than in the other patients. In the NIH study, specific hemodynamic variables were independently and significantly related to mortality, including , diastolic P ap , systolic P ap –pulmonary capillary wedge pressure, transpulmonary gradient, and pulmonary vascular resistance. These data suggest that mortality in adults with sickle cell disease and PH is proportional to the physiologic severity of pulmonary vascular disease.


Epidemiologic studies suggest that the central risk factor for the development of PH in patients with sickle cell disease is the severity of hemolytic anemia ( Fig. 94-4 ). The association between the development of PH and the intensity of hemolytic anemia has been observed in three prospective screening studies of adult patients with sickle cell disease, in an expanding number of pediatric studies, and in studies using right heart catheterization to define PH. This suggests that hemolysis is related mechanistically to PH. That relationship is plausible and biologically significant because free hemoglobin inactivates the intrinsic vasodilator NO. Hemolysis also releases arginase, which depletes L-arginine, the substrate for NO synthesis. These combined mechanisms result in a state of decreased NO bioavailability and “resistance” to NO-dependent vasodilation.

Figure 94-4

Multifactorial pathogenesis of pulmonary hypertension in sickle cell disease, focusing on changes at a small pulmonary artery.

PH is thought to arise from multifactorial effects of hemolysis, as well as from the anemia leading to increased cardiac output and multiorgan damage from iron overload, renal failure, and asplenia. EPO, erythropoietin; ET-1, endothelin-1; HIF, hypoxia inducible factor; NO, nitric oxide; PS, phosphatidylserine; , cardiac output; TF, tissue factor; VEGF, vascular endothelial growth factor.

Hemolysis and decreased NO bioavailability also induce platelet activation, thrombin generation, and tissue factor activation. There is also a correlation between the rate of hemolysis and the levels of procoagulant factors in blood in patients with sickle cell disease. Hemolysis is also associated with the formation of red blood cell microvesicles expressing phosphatidylserine, which activates tissue factor. These factors all contribute to an increased risk of thrombosis. In addition, sickle cell patients with functional asplenia and thalassemia patients with surgical splenectomy have increased levels of cell-free hemoglobin and red cell microvesicles, providing a potential mechanism for the hypercoagulability associated with both diseases, with possible exacerbation by asplenia. Finally, the accumulation of redox-active heme and iron from lysed red blood cells further contributes to the generation of reactive oxygen species that can exacerbate ischemia-reperfusion injury, thrombosis, and vascular proliferative responses. Another downstream effect of hemolytic anemia includes increased endothelin-1-mediated vasoconstrictive and proliferative responses. In patients with sickle cell disease, both at steady state and during vaso-occlusive pain crises, plasma endothelin-1 levels are increased. In vitro, sickled erythrocytes increase endothelin-1 production by cultured human endothelial cells. All told, hemolysis acts via multiple mechanisms to block NO-induced vasodilation, activate procoagulant activity, injure endothelial cells, and exacerbate vasoconstriction and proliferation.

Functional or surgical asplenia could also contribute to the development of PH in patients with sickle cell disease. Splenectomy has been reported to be a risk factor for the development of PH, particularly in patients with hemolytic disorders. Loss of splenic function may trigger platelet activation, promoting pulmonary microthrombosis and red cell adhesion to the endothelium. The spleen also plays a critical function in the removal of senescent and damaged erythrocytes. Following splenectomy, moreover, the rate of intravascular hemolysis increases.

Historically, PH in patients with sickle cell disease was thought to derive from the repetitive episodes of pulmonary vaso-occlusive crisis and ACS leading to pulmonary fibrosis, vascular obstruction, and chronic hypoxemia. However, this association has not been supported by epidemiologic studies in which the number of episodes of vaso-occlusive crisis and ACS have not been shown to be associated with PH.

Clinical Features and Evaluation.

The diagnostic evaluation of patients with sickle cell disease suspected of having PH should follow the same guidelines established for other causes of PH. Given the high prevalence of PH in this population and the associated high mortality, we recommend that all adults with sickle cell disease undergo universal noninvasive screening with Doppler echocardiography or with assessment of steady-state plasma NT-proBNP levels. In patients with sickle cell disease, as described earlier, the echocardiographic estimation of pulmonary artery systolic pressures (where pulmonary artery systolic pressure = 4*TRV 2 + estimate of the right atrial pressure) correlates reasonably well with measured pulmonary systolic pressures by right heart catheterization.

Although no prospective data on prevalence and risk of PH are available for children, we currently recommend that children with hypoxemia, high hemolytic rates (hemoglobin values < 7 g/dL with high LDH values), and/or recurrent ACS be screened. It is important that such screening be performed in the steady state because pulmonary pressures rise during vaso-occlusive painful crisis.

The diagnosis of PH in patients with sickle cell disease can be challenging ( Fig. 94-5 ; and ). Exertional dyspnea, the most typical presentation of PH, is also a cardinal symptom of chronic anemia, and therefore a high index of suspicion for PH is necessary. Other conditions that commonly present in patients with sickle cell disease, such as left ventricular dysfunction, pulmonary fibrosis, and liver cirrhosis, could also present in a similar fashion and result in PH. Patients with PH tend to be older and have higher systolic arterial blood pressure, lower hemoglobin levels, higher indices of hemolysis (such as high bilirubin or LDH values), lower hemoglobin oxygen saturation, greater degree of renal and liver dysfunction, and a higher number of lifetime red blood cell transfusions. As such, the diagnostic evaluation of these patients should include an aggressive search for other conditions that might contribute to PH, such as iron overload, chronic liver disease, human immunodeficiency virus (HIV), nocturnal hypoxemia, and thromboembolism.

Figure 94-5

Imaging features of sickle cell disease and pulmonary hypertension.

A, Echocardiographic four-chamber view of the heart, illustrating severe right ventricular (RV) and right atrial (RA) dilation and moderate tricuspid regurgitation (blue color Doppler). Below, Doppler tracing from a patient with severe pulmonary hypertension reveals a jet velocity of more than 4 m/sec. B, Axial chest CT scan showing enlargement of pulmonary arteries ( arrows ) in severe pulmonary hypertension. C, Chest CT showing mild pulmonary fibrosis typical of patients with sickle cell disease and pulmonary hypertension. D, Perfusion scan demonstrating patchy areas of abnormal perfusion. Ventilation scans were normal (not shown).

A diagnostic right heart catheterization is essential to confirm the diagnosis and exclude diastolic dysfunction (see later). The 6-minute walk test is a useful surrogate for functional capacity in this patient population even considering the high prevalence of confounding factors such as the presence of avascular necrosis of the hip. Specifically, the 6-minute walk test inversely correlates with the severity of PH, and PH-specific therapy improves walk distance.

The Doppler echocardiogram is an important tool for population screening and estimation of pulmonary artery pressure, but it cannot be used to diagnose or define PH in an individual patient. For a given patient, a diagnosis of PH is based on the measurement of the mean pulmonary artery pressure via a right heart catheterization. In a single individual, the use of a TRV ≥ 2.5 m/sec has low specificity for a diagnosis of PH, with only 25% of patients with a TRV ≥ 2.5 m/sec having the diagnosis of PH. A TRV ≥ 2.9 m/sec has a higher positive predictive value of 64%, but high false-negative rate of 42%. Combining a TRV ≥ 2.5 m/sec, a high NT-proBNP level (>164.5 pg/mL), and a low 6-minute walk distance of less than 333 m, the positive predictive value for PH is 62%, with a false-negative rate of 7%.

Abnormal pulmonary function can be found in most patients with sickle cell disease. The pulmonary function abnormalities are characterized by mild restrictive lung disease (mean total lung capacity values of about 79% of predicted), abnormal diffusing capacity, and radiographic signs of mild pulmonary fibrosis, and the severity of these defects seems to be slightly greater in those patients with PH. However, in these patients the degree of pulmonary function abnormalities is rarely severe enough to be a major contributor to the etiology of PH.

A ventilation-perfusion scan is an indispensable component of the evaluation because chronic thromboembolic PH, if amenable to pulmonary endarterectomy, is a potential curable cause of PH reported in patients with chronic hemolytic disorders. In the majority of cases, scintigraphic evidence of thromboembolic disease is uncommon and the most commonly seen abnormality is of patchy areas of abnormal perfusion, similar to findings described in other forms of PH. Chronic thromboembolic PH is present in approximately 5% of sickle cell patients with severe PH, and the disease has been successfully treated surgically in two patients with sickle cell disease ( Fig. 94-6 ). As such, patients should undergo imaging studies and, if suggestive of chronic thromboembolic PH, should undergo more invasive studies (i.e., angiography) to exclude this potentially surgically treatable condition.

Figure 94-6

Chronic thromboembolic pulmonary hypertension in a patient with sickle cell disease.

A, High-resolution CT scan demonstrating mosaic pattern of attenuation. Multiple hyperlucent regions represent areas with decreased perfusion ( arrows ). B, Ventilation-perfusion scan demonstrating multiple unmatched perfusion defects ( arrows ). Ventilation scans, top 7 images; perfusion scans, bottom 8 images. C, Digital subtraction pulmonary angiogram that shows diffuse peripheral areas of hypoperfusion, which represent multiple peripheral small filling defects ( arrows ).

Measurement of NT-proBNP levels can be used as a PH biomarker for diagnosis and risk stratification in patients with sickle cell disease. In a contemporaneous cohort, NT-proBNP levels were higher in patients with sickle cell disease–associated PH and correlated directly with the severity of PH and the degree of functional impairment. An NT-proBNP level of 160 pg/mL or greater had a 78% positive predictive value for finding an elevated TRV and was an independent predictor of mortality. In a subset of individuals participating in the MSH follow-up study in 1996, 30% of patients had an NT-proBNP level of 160 pg/mL or greater. An NT-proBNP level of 160 pg/mL or greater in the MSH cohort was independently associated with mortality. Similarly, in patients enrolled in the CSSCD from 1978 and 1988, elevated levels of NT-proBNP were independently associated with a higher risk of death.

In contrast to patients with traditional forms of PH (e.g., idiopathic pulmonary arterial hypertension, scleroderma associated PAH) (see Chapters 58 and 59 ), who are symptomatic with s in the range of 50 to 60 mm Hg, in patients with sickle cell disease the degree of elevation in is mild to moderate, in the range of 30 to 40 mm Hg, with mild elevations in pulmonary vascular resistance. These patients also have coexistent mild elevation in pulmonary capillary wedge pressure, suggesting left heart failure ( Table 94-4 ). Right heart catheterization data from these multiple studies show that the hemodynamic etiology of the PH in patients with sickle cell disease is multifactorial: precapillary PH (defined by an ≥ 25 mm Hg and a wedge pressure ≤ 15 mm Hg) is present in 50% of catheterized patients, whereas pulmonary venous hypertension secondary to left ventricular diastolic dysfunction disease (defined by an ≥ 25 mm Hg and a wedge pressure > 15 mm Hg) is present in 50%.

Further, using echocardiography in a cohort of 141 patients with sickle cell disease, Sachdev and associates found that 47% of patients had PH, diastolic dysfunction, or both (29% had PH alone, 11% had diastolic dysfunction and PH, and 7% had diastolic dysfunction alone). PH and diastolic dysfunction were associated with a relative risk of death of 5.1 and 4.8, respectively, while the relative risk of death when both were present was 12.0. These data suggest that both PH and diastolic dysfunction independently carry mortality risk. In addition, in a series of 483 patients with homozygous sickle cell disease, markers of diastolic dysfunction were independently associated with a low 6-minute walk distance.

Regardless of the hemodynamic etiology, sickle cell disease patients with chronic anemia, who maintain a high compensatory resting cardiac output in order to ensure adequate oxygen delivery, normally have a reduced pulmonary vascular resistance and appear to be poorly tolerant of even small increases in pulmonary vascular resistance. When compared with age-, gender-, and hemoglobin-matched patients with sickle cell disease without PH, individuals with PH exhibited a lower 6-minute walk distance (435 ± 31 vs. 320 ± 20 meters; P = 0.002), lower peak oxygen consumption on cardiopulmonary exercise testing (50 ± 3% vs. 41 ± 2% of predicted; P = 0.02), and higher ventilatory equivalent for CO 2 at anaerobic threshold (31.6 ± 1.5% versus 39.2 ± 1.6: P = 0.035) on cardiopulmonary exercise testing. Patients with sickle cell disease and PH appear to have worse function than is reported for patients with PH from other etiologies. In addition, in patients with sickle cell disease, pulmonary artery vascular resistance sharply rises with exercise, suggesting that pulmonary vascular disease contributes to functional limitation in these patients. Taken together, these data suggest that, in sickle cell disease patients with chronic anemia, mild to moderate PH has a severe adverse impact on functional and aerobic exercise capacity.


Data on the specific management of patients with sickle cell disease and PH are limited. Most of the recommendations are based on expert opinion or extrapolated from data derived from other forms of PH. The general approach should include maximization of sickle cell disease–specific therapy (i.e., treatment of primary hemoglobinopathy), treatment of hypoxia with chronic oxygen therapy, treatment of associated conditions (such as iron overload, chronic liver disease, HIV infection, nocturnal hypoxemia, and thromboembolic disorders). For patients with sickle cell disease with PH ( ≥ 25 mm Hg and wedge pressures < 15 mm Hg with a relatively high pulmonary vascular resistance, > 160 dyn•sec −1 •cm −5 ), targeted therapy with pulmonary vasodilator/antiremodeling agents can be considered ( Fig. 94-7 ).

Jul 21, 2019 | Posted by in CARDIOLOGY | Comments Off on Pulmonary Complications of Hematologic Diseases

Full access? Get Clinical Tree

Get Clinical Tree app for offline access