Thalassemia is a hereditary hemoglobinopathy and transfusion-dependence is one essential factor in distinguishing the various thalassemia phenotypes and their severity.1 The ”-thalassemias are a group of recessively inherited hemoglobin synthesis disorders resulting from mutations in the ”-globin gene and consequently in defective ”-chain production which lead to an imbalance in ”/”-globin chain synthesis, ineffective erythropoiesis, reduced red blood cell (RBC) survival, and subsequent hemolytic anemia. 1
Non-transfusion-dependent thalassemias (NTDT) is a term used to label patients who are not transfusion dependent for survival, although they may require occasional or even frequent transfusions for defined periods of time depending on the severity of illness in certain clinical settings2. NTDT encompasses three clinically distinct forms: ”-thalassemia intermedia, hemoglobin E/”-thalassemia (mild and moderate forms), and ”-thalassemia intermedia (hemoglobin H disease). 3
A hypercoagulable state has been identified in NTDT patients which can be present since childhood. The hypercoagulable state in patients with NTDT has been primarily attributed to abnormalities in platelets and the pathological red blood cells, although several other factors are believed to be involved leading to clinical thrombosis.
As a result of this hypercoagulable state there is increased occurrence of venous thrombosis4. The prevalence of overt strokes in NTDT patients with a history of thrombosis ranges between 5% and 9% 5. These silent strokes usually present with no obvious clinical features, but they have been proposed to have long term neurocognitive side effects, which can have damaging consequences to cognitive function as overt stroke6.
The occurrence of silent stroke has been studied and the prevalence ranges from 5% – 60%. The exact prevalence of these lesions remains to be extensively studied. There is minimal literature available in India. There is an urgent need to establish the prevalence of this complication in NTDT in this part of the globe, so as to generate a proper management protocol.
REVIEW OF LITERATURE
NONTRANSFUSION DEPENDENT THALASSEMIA (NTDT)
Non transfusion dependent thalassemias present to medical attention in later childhood, sustain more favorable hemoglobin levels 7 to 10 gm/dl, and generally remain transfusion-independent2. They may, however, require transfusion therapy in cases of poor growth, extensive evidence of extrameullary hematopoiesis, acute stressors like intercurrent illness, surgery or pregnancy, later in adulthood when the disease severity progresses, or in certain settings where a benefit of transfusion therapy has been established2. Many patients with NTDT are set on a life of unnecessary regular transfusions, particularly if they present during a period of intercurrent infection requiring a few transfusion7. It is essential to evaluate the patient carefully over the first few months after the genetic diagnosis of ”-thalassemia is established and not to initiate patient
on any treatment modality, especially regular transfusion therapy, too hastily7. The patient’s well-being, particularly with respect to activity, growth, development, and the early appearance of skeletal changes or other morbidities, severity of anemia & severity of organomegaly are the factor to be taken into consideration before the phenotype is clearly established and the treatment modality is selected7. NTDT encompasses three clinically distinct forms: ”-thalassemia intermedia, hemoglobin E/”-thalassemia (mild and moderate forms), and ”-thalassemia intermedia (hemoglobin H disease) 3(11).
The term NTDT lacks specific molecular correlates and the diagnosis remains largely clinical. Most NTDT patients are homozygotes or compound heterozygotes for ”-thalassemia, meaning that both ”-globin loci are affected and the disease is inherited autosomal recessively.8 Less commonly, only a single ”-globin locus is affected, the other being completely normal; in these instances, NTDT is dominantly inherited8. The phenotype of NTDT may also result from the increased production of ”-globin chains by a triplicated or quadruplicated alpha genotype associated with beta-heterozygosity9.
In NTDT, the genetic basis for phenotypic diversity is best explained in terms of primary, secondary, and tertiary genetic modifiers9. The primary modifiers represent the broad diversity of mutations that affect the beta-globin genes, ranging from mild promoter mutations that cause a slight reduction in beta-globin chain production to the many different mutations that result in the ”0-thalassemias, that is, a complete absence of ”-globin product10. The secondary genetic modifiers are those that are involved directly in modifying the degree of globin-chain imbalance in beta thalassemia10. The coinheritance of alpha-thalassemia has this effect, and, because there are numerous different molecular forms of alpha-thalassemia of different severity, this interaction provides further scope for a wide range of different beta thalassemia phenotypes11. Similarly, the degree of globin chain imbalance can be reduced by the more effective synthesis of the gamma chains of fetal hemoglobin after birth. Tertiary modifiers are those that are not related to globin chain production but that may have an important effect on the complications of the disease.
The primary pathological abnormality is imbalance in the production of ”- and ”-globin chains. The effect of this imbalance in globin chain synthesis leads to ineffective erythropoiesis and hemolysis15. As a result of this decreased erythroid red blood cells output resulting in chronic hypoxia there is increased erythroid expansion leading to development of bony and facial changes, organomegaly and extramedullary hematopoietic tumours15. Another cause for clinical symptoms is secondary to iron overloading in these patients. Iron absorption is essentially controlled by hepcidin, a small peptide secreted by the hepatocytes, which blocks iron uptake in the intestine and iron release from the reticuloendothelial system12. Hepcidin expression is enhanced by iron overload and inflammation, whereas it is inhibited by anemia and hypoxia13. In NTDT, ineffective erythropoiesis is the central process that leads to inappropriately low hepcidin levels and increased intestinal iron absorption. Growth differentiation factor 15 inhibits hepcidin expression by opposing the effect of BMP, thereby leading to increased intestinal iron absorption and increased iron release from macrophages16. Consequently, the secretion of ferritin is reduced and its serum level relatively decreased16. Therefore, in thalassemia intermedia, the determination of serum ferritin underestimates the extent of iron accumulation
HYPERCOAGULABILITY IN NTDT
A hypercoagulable state has been identified in NTDT patients, which can be present since childhood15. This hypercoagulable state is attributed to abnormal platelets, activated endothelial cells and pathological red blood cells17. Patients with NTDT have chronically activated platelets and enhanced platelet aggregation, as confirmed by the increased expression of in vivo platelet activation markers CD62P (P-selectin) and CD6315. It has been demonstrated that NTDT patients have 4 to 10 times higher prostacyclin and thromboxane A2 metabolites, both markers of hemostatic activity, as compared to healthy individuals15. Splenectomized NTDT patients also have higher platelet counts but with a shorter life-span due to enhanced consumption15. Increased platelet adhesion is a common finding in splenectomized ”-thalassemia patients and is a strong contributor to occlusive thrombus formation18.
The red cells of patient with thalassemia are abnormal with less deformability increased rigidity and aggregation15. Due to the imbalance in the alpha and non alpha globin units there is Oxidation of globin subunits in thalassemia erythroid cells which leads to the formation of hemichromes15. As a result of this heme disintegration there is release of toxic iron species. The free iron in turn catalyzes the formation of reactive oxygen species, leading to oxidation of membrane proteins and formation of red-cell senescence antigens like phosphatidylserine , which cause the thalassemic red blood cell to become rigid, deformed, and tend to hyperaggregate19. Thalassemic red blood cells with such negatively charged phospholipids increase thrombin generation. Splenectomized patients have a substantially higher number of these negatively charged pathological red blood cells and in turn show higher thrombin generation20. NTDT patients have higher levels of procoagulant microparticles of red blood cell, leukocytic, and endothelial origins compared to controls21.
Elevated white blood cells are known risk factors for thrombosis in sickle cell anemia. White blood cells activated by the toxic oxygen species release proinflammatory cytokines including TNF”, IL1” and IL6. These cytokines activate endothelial cells to express more adhesion molecules like E- selectin and ICAM. These molecules on interaction with ligands on activated white blood cells like integrin and P-selectin increase the anchoring of white blood cells. White blood cells because of their larger size than RBC cause vessel lumen occlusion. They at the same time release cytokines which recruits other activated platelets and RBC. Further worsening of the situation occurs when the enzymes and peptides released by activated white blood cells cause blood vessel damage to cause arteriopathy lesions. This is evident from study by Catherine et al, where the predictors of arteriopathy lesions were early school age, recurrent upper respiratory tract infections and sickle cell disease. In all these 3 situations white blood cells are elevated and are significant risk factors for thrombotic/ vasoocclusive episodes in sickle cell anemia patients.
Activation of endothelial cells occurs by reactive oxygen species & cytokines released by white blood. This leads increased expression of E-selectin, intercellular adhesion molecule-1, von Willebrand factor, and vascular cell adhesion molecule-1 in the endothelial cells as a result of released reactive oxygen species15. Inherited thrombophilia does not have a role in the hypercoagulable state of NTDT but low levels of antithrombin III, proteins C, and protein S have been documented22.
THROMBOSIS IN NTDT
The presence of hypercoagulable state in patients of NTDT predisposes these patients to increased incidence of thrombotic events. Data on the incidence of thrombotic events in NTDT patients is limited15. Nine Italian pediatric thalassemia centers undertook multicentric study to estimate the prevalence of thrombotic events, where 4% of 683 patients with ”-thalassemia major and 9.6% of 52 patients with ”-thalassemia intermedia had experienced a thrombotic events23. In a cohort study involving 83 splenectomized patients with ”-thalassemia intermedia who were followed up for over 10 years, 29% of patients experienced a venous thrombotic event. In a study by Taher et al, they studied 8,860 thalassemia patients in the Mediterranean area and Iran, and observed that thrombotic events, mostly venous, occurred 4.38 times more frequently in NTDT patients than regularly-transfused ”-thalassemia major25. The OPTIMAL CARE (Overview on Practices in Thalassemia Intermedia Management Aiming for Lowering Complication rates Across a Region of Endemicity) study evaluated 584 patients with ”-thalassemia intermedia and established that thrombotic disease, mostly venous, was the 5th most common complication, affecting 14% of the patient population24.
Cappellini et al. followed 83 patients with NTDT over 10 years, 82 of whom were splenectomized, and found that 24 patients (29%) developed either deep vein thrombosis or pulmonary embolism or portal vein thrombosis26. Thromboembolic events in splenectomized NTDT patients is not an acute event and is a result of chronic underlying process. The median time to thrombosis is 8 years, emphasizing the need for a long-term treatment modality for prevention of this problem26.
Logothetis et al. described transient ischemic attacks in about 20% of 138 cases of TM in Greece27. Similar evidence of transient ischemic attacks accompanied by a clinical picture of headache, seizures and hemiparesis was shown in 2.2% of patients with TM by Pignatti et al24. Due to a higher prevalence of conventional stroke-related risk factors overt stroke occurs more frequently in TM than in NTDT, a high prevalence of silent strokes in NTDT patients has been documented in literature. Manfre et al gave the first documentation of silent stroke in NTDT patients in 1972. The study showed a 37.5% rate of ischemic lesions on brain MRI in patients with NTDT who were neurologically intact and had no conventional stroke-related risk factors28. In a cross-sectional brain MRI study by Taher et al conducted in Lebanon in splenectomized adults with NTDT who did not have any significant neurological or cognitive signs or symptoms or any stroke-related risk factors. 60 % patients had one or more ischemic lesions detected in the subcortical white matter on brain MRI29.
Evaluation of intracranial blood flow velocity in neurologically asymptomatic NTDT patients using transcranial doppler ultrasonography revealed that mean flow velocities in the intracranial circulation of patients with ”-thalassemia intermedia are higher than healthy controls, but were lower than those associated with ischemic stroke risk in patients with sickle cell disease (>2 m/s) 15.This reiterates the fact that chronic anemia predisposes the patient to develop silent stroke. Evaluation of NTDT patients with PET CT had revealed multiple predominantly left sided lesions of decreased neuronal function.
Risk factors that predisposes to the development of thromboembolic events are multitude, which have been evaluated in detail by number of studies. Patients with NTDT who developed thromboembolic events were mostly splenectomized, nontransfused and had a hemoglobin level below 90 g/l, which justifies a higher thrombotic drive as abnormal RBCs are expected to remain in circulation for longer. In the OPTIMAL CARE study the main independent risk factors for thrombotic events were splenectomy, age >35 year, iron overload (serum ferritin level ‘1000 ng/ml), and a hemoglobin level <9 g/dl24. Splenectomized NTDT patients who experience thrombosis are characterized by high nucleated red blood cell of > 300”103 /mm3 and platelet count > 500”103 /mm3 further confirming the dual role of platelets and red blood cells in this setting15. In a study by Taher et al age above 20 years, splenectomy, personal or family history of thrombotic events were identified as the main risk factors for thrombosis in ”-thalassemia intermedia patients29.
Occurrence of number of complications in NTDT increases with advancing age. With advancing age, there is a statistically significant trend towards a higher rate of leg ulcers, thrombosis, pulmonary hypertension, extramedullary hematopoiesis, hypothyroidism and osteoporosis30.The total number thrombotic events in children less than 10 years was < 3% of the total thrombotic events30. This reiterates the fact that there has to be chronic anemia and hypoxia to develop thromboembolic events. Inherited thrombophilia does not have a role in the hypercoagulable state of NTDT, but low levels of antithrombin III, proteins C, and protein S have been documented. The role of this low anticoagulant protein in predisposing to thromboembolic event is remains to be proven15 .In a study by Taher et al it was observed that patient with factor VII polymorphism of H7 allelle was protective to the occurrence to thromboembolic event39. So there is a possible role of thrombophilia in evelopment of thrombotic events which needs to be further studied. PULMONARY HYPERTENSION IN NTDT Chronic anemia and hypoxia, iron overload, splenectomy, hypercoagulability, and microthrombotic disease of the pulmonary circulation have all been implicated in the pathophysiology of pulmonary hypertension in NTDT. Recently, decreased arginine bioavailability and nitric oxide depletion secondary to hemolysis have also been associated with pulmonary hypertension in patients with thalassemia15. Prevalence rates of pulmonary hypertension ranged between 10% and 78.8% in NTDT patients15.Noninvasive measurement of pulmonary arterial systolic pressure (PASP) is not possible. In the absence of right ventricular outflow tract obstruction, right ventricular systolic pressure (RVSP) is taken as equivalent of pulmonary arterial systolic pressure38. RVSP is estimated from peak tricuspid regurgitant jet velocity (TRJV), using the simplified Bernoulli equation and combining this value with an estimate of the Right arterial (RA) pressure. RVSP = 4'' (TRJV)2 + RA pressure. RA pressure is commonly estimated by IVC diameter and the presence of inspiratory collapse38. The diagnosis was usually established based on a tricuspid-valve regurgitant jet velocity (TRJV) exceeding 2.5 m/s corresponding to a pulmonary arterial systolic pressure exceeding 30-35 mm Hg15. Although pulmonary hypertension is neither associated with myocardial siderosis nor with left ventricular dysfunction in NTDT patients, it is a leading cause of right-sided heart failure. Evidence of pulmonary hypertension and transfusion naivety were the common features associated with higher risk of thromboembolic events in splenectomized patients. Presence of pulmonary hypertension as one of the risk factor has not been evaluated, but is proposed to be one of the risk factors for thrombotic events. IMAGING IN IDENTIFICATION OF SILENT STROKE Thromboembolic events of central nervous system are better evaluated with Magnetic Resonance Imaging. MRI has proven to be highly sensitive in the identification of Silent cerebral infarction [SCI]. Due to its sensitivity to detect increased amounts of water within brain tissue at sites of acute and chronic infarctions, especially with fluid-attenuated inversion recovery (FLAIR) images, MRI has revolutionized the ability to see not only acute symptomatic infarctions, but also SCIs, especially those within the deep white matter29. Silent cerebral infarct is defined as abnormal magnetic resonance imaging of the brain in the setting of a normal neurological examination without a history or physical findings associated with an overt stroke31. Lesion detection is dependent upon the MR technique, the slice thickness and the magnetic field strength. As further advances in imaging are made and medical centers transition from 1.5 Tesla to 3.0 Tesla magnets, more individuals with NTDT will be detected with SCI. In a study by Taher et al 60% patients had evidence of one or more SCIs on brain MRI all involving the subcortical white matter. Most patients had evidence of multiple lesions29. The frontal subcortical white matter was more commonly involved followed by the parietal and occipital subcortical whitematter. Small to medium size lesions of size of 1.5 cm were found in 94% of patients29. The diffuse nature of the lesions and their high prevalence in the frontal and parietal white matter was similar to those found in patients of sickle cell disease. Advanced age and transfusion naivety was independently associated with a higher occurrence and multiplicity of lesions. Evaluation of NTDT patients with magnetic resonance angiography (MRA) revealed that the lesion confined to the location of areas perfused by small penetrating arteries. In study by Taher et al it was observed that 27.6% had evidence of arterial stenosis on MRA29. The majority of lesions had mild narrowing and mostly involved the internal carotid artery. These findings are surely pathological, as the prevalence of mild large-vessel stenosis found incidentally in healthy individuals does not exceed 4%32. This large vessel stenosis failed to explain SCI in a majority of patients. The prevalence and effects of these arterial stenosis needs to be studied THE ROLE OF INTERVENTION Transfusion therapy Regular transfusion therapy maintains stable hemoglobin and reduces the occurrence of chronic hypoxia, thereby suppressing the chronic stimulus to increase the rate of hematopoiesis15. This decreases the rate of ineffective erythropoiesis and the levels of pathological red blood cells with thrombogenic potential. Transfusion therapy may in fact explain the lower rate of thrombotic events in regularly-transfused ''-thalassemia major patients than NTDT. The role of blood transfusion in the primary or secondary prevention of thrombotic events in NTDT patients has not been evaluated in clinical trials but from the observation that it prevents the development of silent cerebral infarct in patients with sickle cell anemia it is believed that it may reduce the occurrence of silent cerebral infarct in patients with NTDT15. With advances in safe and effective iron chelation therapy and improvement in blood screening programme for transfusion transmitted infection it may be a plausible choice in NTDT patients33. Should this be offered to all ? This question needs to be researched. Anti platelet therapy Thrombocytosis following splenectomy and existence of chronically activated platelets in circulation predisposes the patient to develop silent cerebral infarct15. The lower recurrence rate of thrombosis in NTDT patients, who took aspirin after their first event, when compared to those who did not, suggests a prophylactic role for aspirin in TI. The cut off value of platelet count for starting aspirin therapy is debatable and in the recent NTDT guidelines it is recommended to start patient on aspirin beyond a platelet count of 500''1003/mm3 15. Even in patient with normal platelet counts, there could be a role for aspirin therapy considering the observation that it could delay occlusive thrombus formation in carotid arteries of thalassemic mice but this needs extensive studies to be followed as a recommendation in patients with NTDT. Fetal hemoglobin induction Induction of fetal hemoglobin production to improve anemia using hydroxycarbamide has been used in NTDT patients34. Treatment with hydroxycarbamide was also shown to decrease plasma markers of thrombin generation. It reduces the coagulation activation by reducing phospholipid expression on the surface of both RBC and platelets and decreasing RBC adhesion to thrombospondin by decreasing the number of pathological RBC and platelets34. In addition to being a nitric oxide donor, hydroxycarbamide may also decrease hemostatic activation by its effect in decreasing the white blood cell count and particularly monocytes that express tissue factor15. But their routine use needs to be studied extensively. Reconsider splenectomy One of the essential functions of spleen is scavenging. The spleen keeps scavenging procoagulant platelets and red blood cells36. About 80% of pathological red blood cells are removed extravascularly by macrophages present mainly in the spleen. Spleen may be a reservoir of excess iron and may have a possible scavenging effect on iron free species such as non-transferrin bound iron, which may explain the higher serum level of this free iron species in splenectomized NTDT patients 15. By keeping a check of these procoagulant microparticles, toxic iron species, procoagulant platelets and red blood cells, there is a decreased incidence of thrombotic events in non splenectomised patients36. So it is suggested that splenectomy should be reserved for cases of: 1) worsening anemia leading to poor growth and development when transfusion therapy is not possible or iron chelation therapy is unavailable; 2) hypersplensim leading to worsening anemia, leukopenia, or thrombocytopenia and resulting in recurrent bacterial infection or bleeding; and 3) splenomegaly accompanied by symptoms such as left upper quadrant pain or early satiety or massive splenomegaly (largest dimension >20 cm) with concern about possible splenic rupture.36
Iron chelation therapy
An independent association between iron overload and thrombotic disease in patients with NTDT is suggested by observational studies, further studies are needed to confirm such observation35. The role of iron chelation therapy in this setting has not been evaluated. But overall NTDT patients are iron overloaded despite being transfusion independent due to chronic anemia leading to increased iron absorption from gut and increase release of iron from phagocytes due to low levels of serum hepcidin. Iron chelation is needed once serum ferritin is in excess of 800ng/ml as recommended by Thalassemia International Federation guidelines for NTDT15.
Effects of silent cerebral infarction
One important question to ask is whether the observed silent infarcts are truly silent or does it require careful consideration and intervention. In the general population, silent infarcts in the white matter are associated with impaired cognitive skills suggesting they can be nearly as damaging to cognitive function as overt stroke36. Children with silent cerebral infarct have lower cognitive test scores when compared to children with a normal MRI of the brain. Specific areas of deficit have been associated with silent infarct, including executive functions like selective attention, working memory, and processing speed, visual motor speed and coordination, vocabulary, visual memory and abstract reasoning and verbal comprehension36. As a consequence of these specific deficits, academic achievement in mathematics and reading are also affected, with one study reporting that the 35% of children silent cerebral infarct had twice the chance of academic difficulties as those without silent cerebral infarct.
There is no study so far to present the prevalence of silent cerebral infarction from India. This study is intended to estimate the prevalence of this peculiar complication in non-transfusion dependent thalassemia. To evaluate the association of possible risk factors in literature in Indian children and adolescent and find out the possible interventions that can be planned to facilitate management.
To study the frequency of neuroimaging abnormalities in non-transfusion dependent thalassemia children and adolescents >10 years of age.
1. To evaluate the frequency of silent cerebral infarction in non- transfusion dependent thalassemia children utilizing brain MRI.
2. To study the frequency of modifiable and non modifiable risk factors of hypercoagulable state in non- transfusion dependent thalassemia children and evaluate their association with neuroimaging abnormalities .
MATERIALS AND METHODS
Nature of the study : Prospective observational study
Period : JULY 2014 to JUNE 2015
Settings : Pediatric Hematology Oncology unit
Advanced Pediatric centre,
Sample size : Targeted sample size will be 25.
Both the following criteria should be satisfied for inclusion in the study.
1. Diagnosed patient of non-transfusion dependent thalassemia.
2. Age >10 years
1. Neurological abnormality detetcted on physical examination.
2. Cerebral vascular malformation detected in brain MRI
3. Patients with diabetes mellitus
4. Patients with hypertension.
5. Contraindications to undergo MRI examination.
6. Patients on treatment with anticoagulant therapy
7. Patients with known cardiac diseases
Patients of Non-Transfusion dependent thalassemia, satisfying the study criteria and visiting the pediatric hematology clinic of the Advanced Pediatric Centre, PGIMER were enrolled.
NTDT is diagnosed when patients fulfilled at least two of the following criteria37:
(i) Age at initial presentation ‘ 3 years.
(ii) Administration of first red cell transfusion after the age of 3 years, before/after the diagnosis of thalassemia was established.
(iii) Pretransfusion hemoglobin (Hb) level of ‘7.0 g/dl at the time of first transfusion
(iv) Need for red cell transfusion < 5 times per year. Patients with Non transfusion dependent thalassemia with age above 10 years were enrolled after obtaining detailed informed consent. The demographic details were collected. Details of evaluation like mean hemoglobin and (if regularly transfused) pretransfusion hemoglobin, peak serum ferritin and treatment details like blood transfusion frequency, splenectomy, hemoglobin F induction therapy, anti platelet therapy and the duration of these therapies will be noted[Annexure I]. Clinical examination of these children was performed to rule out overt CNS findings [Annexure I]. At enrollment 6 ml blood sample was collected by venipuncture for complete blood count, serum ferritin and plasma protein C, protein S and antithrombin III evaluation. Plasma protein C and free protein S was evaluated using clot based assay and antithrombin III was evaluated using chromogenic assay. Patients were evaluated by brain MRI using the sequences of T1, T2, FLAIR, DWI and MR Angiography for presence of silent cerebral infarction. Infarction was defined as areas of abnormal increased signal intensity on T2 and FLAIR weighed sequences. MRI brain with silent cerebral infarction was evaluated for number, size & location of lesions and presence of brain atrophy. When infarction was not associate with any neurological symptom or sign it was defined as silent cerebral infarction. Infarction associated with neurological symptom or sign was defined as overt cerebral infarction. Cerebral arteriopathy was defined as intracranial artery narrowing on magnetic resonance angiography. The severity, site and number of arterial narrowing lesions were recorded. The concordancy of site of arteriopathy lesion and infarction was evaluated. Echocardiographic evaluation for pulmonary arterial hypertension would be done using tricuspid regurgitant jet velocity and right atrial pressure. Diagnosis of pulmonary hypertension was made if the tricuspid regurgitant velocity is greater than 2.5 m/s. Pulmonary hypertension was classified as possible pulmonary hypertension if TRJV >2.5 m/s in an asymptomatic individual. Likely pulmonary hypertension was defined as TRJV> 3.2 m/s in an asymptomatic individual or TRJV > 2.4 m/s in a symptomatic individual. The above details were entered in the proforma given in Annexure I. Patients with positive investigational findings were managed as per guidelines for NTDT patients.
Modifiable risk factors include blood transfusion therapy, splenectomy procedure, mean hemoglobin, peak serum ferritin, platelet count, total nucleated RBC count, therapy with hydroxyurea, aspirin and iron chelator and pulmonary hypertension. Non modifiable risk factors include age, family and/or personal history of thrombosis, plasma protein C, protein S and antithrombin III levels.
The data was entered into SPSSv22 software. Baseline variables were analyzed by descriptive statistics. The data on risk factors was compared in patients with neuroimaging abnormalities using chi square test for categorical variable and mann whitney test for quantitative variables. The level of significance was set at < 0.2. Odds ratio was estimated for all the risk factors. Factors with significant odds evaluation were studied to estimate the attributable fraction and population attributable fraction. The significant variables were evaluated in multivariate binomial logistic regression analysis for statistical significance and a cut off P value < 0.5 was taken as significant. Results A total of 35 patients were enrolled during the study period. There were a total of 29 males and 6 females. The mean age at diagnosis was 6.9 '' 2.9 years and median age at enrolment was 14 (IQR12-15) years. Twenty six, 7 and 2 patients were diagnosed to be suffering from '' thalassemia intermedia, E-'' thalassemia and hemoglobin H disease type of NTDT respectively. The state of domicile of patients is given in table I. The mean Hb F at diagnosis was 61''29%. Table I ' Baseline characteristics of patients studied S.no. Baseline characteristics No. of patients (%) (n=35) 1. Sex ' Males Females 29 (82.8) 6 (17.2%) 2. Age at diagnosis 6.9 '' 2.9 years 3. Age at enrolment 14 (IQR12-15) years 4. Type of NTDT - '' thalassemia intermedia E-'' thalassemia Hemoglobin H disease 26 (74.2%) 7 (20%) 2 (5.8%) 5. Mean Hb F at diagnosis 61''29%. 6. State of domicile- Bihar Chandigarh Haryana Himachal Pradesh Jammu & Kashmir Punjab Uttar pradesh 1 (2.9%) 3 (8.6%) 8 (22.9%) 5 (14.3%) 1 (2.9%) 14 (40%) 3 (8.6%) Treatment details: The study cohort included 24 (68.5%) who were transfused at least once in lifetime till enrolment and 11 (31.5%) who were transfusion na''ve. The median age at first transfusion was 6 (IQR 4-8) years. Out of the 24 children transfused, 21 (87.5%) were minimally transfused ('2 units/ year) and 3 (12.5%) were on regular transfusion (> 10 units transfused / year) for a short period of time.
The median number of units transfused in the study cohort was 7 (IQR 2-16 units). Seventeen (48.4%) patients were splenectomized at a median age 8 (IQR 7.5-10.5) years. All 35 patients were on daily hydroxyurea ingestion at a mean dose of 14.7” 2.9 mg/kg/day for mean duration 7.9 ” 3.9 years. None of the patients were on aspirin ingestion at enrolment. Oral deferasirox and deferiprone were consumed by 3 (8.57%) patients each for a mean duration of 4 ” 1.67 years. Table II gives the treatment characteristics of patients studied.
Table II -Treatment characteristics of patients studied (n=35)
S.no. Treatment characteristics No. of patients (%) (n=35)
1. Blood transfusion- transfusion na”ve
Minimally transfused 11 (31.5%)
2. Median age at first transfusion 6 (IQR4-8)
3. Median no. of units transfused 7 (IQR 2-16)
4. Splenectomized ‘ Yes
No 17 (48.4%)
5. Median age at splenectomy 8 (IQR 7.5-10.5 years)
6. Hydroxyurea administration All patients
7. Mean dose of hydroxyurea 14.7” 2.9 mg/kg/day
8. Mean duration of hydroxyurea administration 7.9 ” 3.9 years
9. No. of patients on iron chelator- Deferiprone
Deferasirox 3 (8.5%)
Anthropometry and pubertal status:
Normal stature and short stature was present in 37.1% and 62.9% patients respectively. Sixty five percent of the study cohort was underweight. Delayed puberty was present in 3(8.5%) patients. Hemolytic facies was present in 28 (80%) of the cohort.
Hematological parameters of study cohort:
Hematological parameters of the cohort recorded since diagnosis is as follows: mean hemoglobin since diagnosis was 7.5” 1.41 gm/dl, median hemoglobin for the year before enrolment was 8.4(IQR 7.7-9.3)gm/dl, peak platelet count since diagnosis was 602 ” 322 ”103/mm3, mean platelet count since diagnosis was 367” 170 ”103/mm3, median peak serum ferritin since diagnosis was 425 (IQR283-1120) ng/ml, median peak serum ferritin for 1 year prior to enrolment was 326 (IQR 246-840) ng/ml & the median nucleated RBC count was 2.5 (IQR 0.37-37.8) ”103/mm3
Table III- Hematological parameters of study cohort (n=35)
S.no. Hematological parameter Value
1. Mean hemoglobin since diagnosis 7.5” 1.41 gm/dl
2. Median hemoglobin for the year before enrolment 8.4(IQR 7.7-9.3)gm/dl
3. Peak platelet count since diagnosis 602 ” 322 ”103/mm3
4. Mean platelet count since diagnosis 367” 170 ”103/mm3
5. Median serum ferritin since diagnosis 425 (IQR283-1120) ng/ml
6. Median serum ferritin for 1 year prior to enrolment 326 (IQR 246-840) ng/ml
7. Median nucleated RBC count 2.5 (IQR 0.37-37.8) ”103/mm3
Neuroimaging abnormalities (NIA):
Magnetic resonance imaging performed in 35 patients enrolled showed neuroimaging abnormalities in 15 (42.9%). Out of the 15 patients, 12 (80%) had detectable infarction in MRI imaging, 1 (6.6%) had pituitary hypoplasia and 8 (53.3%) patients had cerebral arteriopathy. Of these 12 patients with infarction, 9 (75%) had no symptoms (silent cerebral infarction) and 3 (25%) had overt cerebral infarction with transient ischemic attack in one and hemiplegia in other 2 patients. Table IV gives the details of neuroimaging abnormality in the study cohort. Table V gives the clinical and laboratory parameters of patients with & without neuroimaging abnormalities and their correlation.
Table IV- Neuroimaging abnormality in the study cohort (n=35)
S.no. Type of neuroimaging abnormality No. of patients (%)
1. Cerebral infarction
Silent cerebral infarction
Overt cerebral infarction 12(34.28%)
2. Pituitary hypoplasia
3. Cerebral arteriopathy 8 (22.85%)
Table V- Correlation of clinical and laboratory parameters with neuroimaging abnormality
S. no Variables With neuroimaging abnormality (n=15) Without neuroimaging abnormality (n=20) P value
1. Age at enrolment (yrs) 14(12-15) 14(12.2-15.7) 0.667
2. Age at diagnosis (yrs) 6.93” 3.41 6.7 ”2.6 0.883
3. Sex ‘ M:F 4:1 17:3 0.708
4. Diagnosis- Hb H disease
” thalasemia intermedia
E-” thalssemia intermedia –
5(34%) 2 (10%)
5. Hb F at diagnosis (%) 56.3” 36.1 64.8”27.4 0.376
6. Transfused ‘ Yes
No 12 (80%)
3 (20%) 13 (65%)
7 (35%) 0.346
7. Freq. of transfusion
12 (60%) 0.557
8. Age at first transfusion (yrs) 5.5(4-8) 7(4-8.5) 0.77
9. Splenectomy 7 (46.67%) 10(50%) 0.81
10. Age at splenctomy(yrs) 8(7-10) 9.5(7.7-11) 0.363
11. Mean hemoglobin (gm %) 7.5” 1.1 8.1”1.13 0.285
12. Mean hemoglobin last year (gm %) 8.3 (7.6-9.1) 8.6(7.8-9.5) 0.422
13. Peak platelet count (” 10 3 /mm3) 616” 352 592”307 0.441
14. Mean platelet count (” 10 3 /mm3) 345” 167 384”175 0.974
15. Peak serum ferritin (ng/ml) 343(293-980) 580(260-1435) 0.902
16. Peak serum ferritin last year (ng/ml) 326(293-902) 402(235-904) 0.922
17. Peak nRBC count (” 10 3 /mm3) 2(0.73-37.8) 3.3(0.24-83.3) 0.845
18. Mean nRBC count (” 10 3 /mm3) 1.3(0.47-16.1) 2.4(0.24-40) 0.87
19. Pulmonary hypertension 3 (20%) 5 (25) 0.737
20. Tricuspid regurgitant jet velocity (m/s) 1.7(1.7-2.4) 1.7(1.5-2.8) 0.901
21. Protein c levels (%) 56(47-61) 49(42-57) 0.209
22. Protein s levels (%) 63.8”16.1 73.3”25.54 0.256
23. Antithrombin III levels (%) 95(87-104) 93(87-102) 0.781
24. Mean TLC (” 10 3 /mm3 ) 8.51”1.64 7.05”1.34 0.007
25. Mean ANC (” 10 3 /mm3) 3.55”0.84 3.45”1.16 0.601
26. Hydroxyurea dose (mg/kg/day) 14.8”2.6 14.5”3.22 0.974
27. Hydroxyurea duration (yrs) 6.08”3.85 8.3”3.9 0.116
28. Low protein C levels 6(40%) 13(65%) 0.15
29. Low antithrombin III levels 2(13.3%) 1 (5%) 0.398
30. Low protein S levels – – –
Silent cerebral infarction in our cohort:
Patients with silent cerebral infarction had a median age of 14 (IQR 12-15 years). Seven (77.8%) patients had a diagnosis of ” thalassemia and 2 (22.3%) patients were diagnosed E-” thalassemia. In patients with silent cerebral infarction 1 was put on regular transfusion regimen at diagnosis of thalassemia syndrome in an outside hospital for a mistaken diagnosis of thalassemia major. Five (55.5%) patients were minimally transfused and 3 (33.3%) were transfusion na”ve. All patients were on regular hydroxyurea therapy for hemoglobin F induction since diagnosis for mean duration of 6.8”2.34 years at a mean dose 14.4”3.2 mg/kg/day. Splenectomy was performed in 5 (55.5%) and not in 4 (44.5%) patients. Eight (88.8%) patients had single lesion and 1 (11.2%) had 2 lesions. The distribution of the infarction was: 5(55.5%) in frontoparietal, 2 (22.2%) parietal, 1 each (11.1%) in the frontal and periventricular region. The lesion was small in size (< 1.5 cm) in 8 (88.8%) patients and large (> 1.5 Cm) in 1 (11.2%) patient.
Table V ‘ Demographic and treatment details of patients with silent cerebral infarction (n=9)
S.no. Demographic and treatment characteristics Value
1. Median age 14(IQR 12-15 years)
2. Diagnosis- ” thalassemia
E-” thalassemia 7 (77.8%)
3. Transfusion therapy-
Regular transfusion regimen
4. Mean dose of hydroxyurea 14.4”3.2 mg/kg/day
5. Mean duration of hydroxyurea 6.8”2.34 years
6. Splenectomized 5 (55.5%)
Overt cerebral infarction:
The mean age of patients with overt cerebral infarction was 14”3 years. The mean age at diagnosis of this group of patients was 4.3” 2 years. Two children were suffering from E-” thalassemia and 1 was diagnosed with ” thalassemia. All of them had received transfusion once in their lifetime but none were on regular transfusion. Splenectomy was performed in one patient. All of them were on regular oral hydroxyurea therapy for a median duration of 3 (IQR 0.2-13) years at mean dose 14 ”1 mg/kg/day. The clinical symptom in 2 (66.6%) patients was hemiplegia and 1 (33.3%) patient with transient ischemic attack. All 3 had only single lesion in imaging. The lesion was small in size in 2 (66.6%) and 1 (33.3%) had large lesion with subarachnoid hemorrhage in frontal region. The localization of infarction was in frontoparietal in all 3. All 3 patients had cerebral arteriopathy. Of them 1 had moyamoya disease in MRI, 1 had segmental attenuation of all 3 cerebral arteries and 1 had focal cerebral arteriopathy of anterior cerebral artery. The site of the infarction was concordant to the site of arterial lesion in 2 (66.6%) patient and in 1 (33.3%) discordant. The characteristics of patients with overt cerebral infarction are given in table VII.
Table VII- Treatment details, clinical characteristics, neuroimaging findings of patients with overt cerebral infarction
S.no Characteristics Patient I Patient II Patient III
1. Age at diagnosis (yrs) 6 5 2
2. Age at enrolment (yrs) 19 11 11
3. Diagnosis of patient Beta E- beta E-beta
4. Transfusion history Minimally transfused Minimally transfused Minimally transfused
5. Splenectomised Yes No No
6. Hydroxyurea therapy
7. CNS symptom Transient ischemic attack Hemiplegia Hemiplegia
8. Neuroimaging findings
Number of lesions
Number of vessel
Anterior cerebral artery
All 3 cerebral artery
Internal carotid artery
Cerebral arteriopathy was present in 8 (22.8%) patients contributing to 53.3% of the total neuroimaging abnormality in study cohort. The median age of this subgroup of patients were 15 (IQR13-19 years). Six patients were diagnosed with ” thalassemia and 2 patients with E- ” thalssemia. One (12.5%) patient was transfusion na”ve, 6 (75%) were minimally transfused and 1 (12.5%) was on regular transfusion for few years for a mistaken diagnosis of thalasemia major in an outside hospital. Splenectomy was performed in 50% of this subgroup cohort. All the patients were on oral hydroxyurea therapy for a median period of 6 (IQR 3-8 years) and dose 16 (IQR15-17) mg/kg/day. The location of the lesions is as follows: 3 (37.5%) each in middle cerebral artery and anterior cerebral artery, 1 in distal part of internal carotid artery presenting as moyamoya disease and 1 had segmental attenuation of all 3 cerebral arteries. The severity of arteriopathy lesion is as follows: 4 (50%) had mild (< 50%) vessel narrowing, 2(25%) had moderate narrowing (50-75%), 1each (12.5%) had severe (> 75%) and total narrowing of arterial lumen. Six patients (75%) with cerebral arteriopathy had cerebral infarction also. Of these 6 patients, 4(66%) had location of infarction corresponding to location of arteriopathy. Of these 6 patients, 2 (33%) were symptomatic infarction and rest 4 were silent infarction.
Table VIII ‘ Demographic & treatment details of patients with cerebral arteriopathy (n=8)
S.no. Demographic and treatment characteristics Value
1. Median age 15 (IQR13-19 years)
2. Diagnosis – ” thalassemia
E- ” thalssemia 6(75%)
3. Transfusion details- Transfusion na”ve
Regular transfusion 1 (12.5%)
4. Splenectomized 4 (50%)
5. Hydroxyurea therapy ‘ Median dose
Median duration 16 (IQR15-17) mg/kg/day
6 (IQR 3-8 years)
Figure II- Cerebral arteriopathy- Site, severity and number of lesion
Figure III- Presence, concordancy and symptomaticity of infarction with cerebral arteriopathy
Pulmonary hypertension in study cohort:
Pulmonary hypertension was present in 8 (22.9%) patients. The median age of patients with pulmonary hypertension was 14.5 (IQR 12-18) years. Five patients (62.5%) were diagnosed with ” thalassemia, 2 (25%) with E- ” thalassemia and 1 (12.5%) with Hb H disease. One patient was transfusion na”ve and the rest were minimally transfused. Fifty percent patients with pulmonary hypertension were splenctomised. All the patients with pulmonary hypertension were treated with hydroxyurea with a median dose of 15(IQR 13-17)mg/kg/day for a median period of 7.5 (IQR 5-10) years. Of these 8 patients, 3 (37.5%) had possible pulmonary hypertension and 5 (62.5%) had likely pulmonary hypertension. None of the patients were symptomatic due to pulmonary hypertension. The median TRJV was 1.7 (IQR 1.45-2.5) m/sec. Figure IV gives number of patients with possible and likely pulmonary hypertension.
Table IX ‘ Demographic and treatment details of patients with pulmonary hypertension
S.no. Demographic and treatment characteristics Value
1. Median age 14.5 (IQR 12-18) years
2. Diagnosis- ” thalassemia
E- ” thalassemia
Hb H disease 5 (62.5%)
3. Splenectomized 4 (50%)
4. Transfusion na”ve
Minimally transfused 1 (12.5%)
5. Hydroxyurea therapy- median dose
Median duration 15(IQR 13-17)mg/kg/day
7.5 (IQR 5-10) years
Figure IV- Type of pulmonary hypertension in study cohort
Deficiency of anticoagulant proteins:
Figure V- No. of patients with normal and low levels of anticoagulant proteins
The median level of protein C was 54 (IQR 47-60) IU/ml, antithrombin III 94 (IQR 87-103) IU/ml and mean free protein S levels were 69.22 ” 22.22 IU/ml. Compared to age appropriate cut off, protein C was low in 54.2%, anti-thrombin III was low in 8.57% and free protein S was normal in all patients. Figure V gives the distribution of normal and low values of anticoagulant proteins in study cohort. All the patients had normal liver function test. None of the patients were on aspirin at enrolment. Iron chelators were taken for a median period of 4.5 (IQR 2-5) years in 6 patients (17.1%). Of these 6 patients, 3 each were on oral deferiprone and deferasirox.
Neuroimaging abnormalities (NIA) and their association with risk factors:
Various possible risk factors were evaluated for the occurrence for NIA. Table X gives the results of the association of the possible risk factors with NIA.
Table X ‘ Results of the various risk factors evaluated for thromboembolic events
Risk factor Mean/ median value in NTDT cohort Mean/median value in NTDT patients with neuroimaging abnormality (n=15) Mean/median value in NTDT patients without imaging abnormality (n=20) P value Odds ratio (95%CI)
Median age (yrs) 14 (12-15) 14(12-15) 14 (12.25- 15.75) 0.681 0 (-1.0, 2.0 )
Mean age at diagnosis (yrs) 6.9 ” 2.9 6.91 ” 3.41 6.77” 2.68 0.882 0(-2.0, 2.0)
Females (%) 29 (82.8)
6 (17.2) 12 (80)
3 (15) 0.698 0.95(0.14, 6.11)
Mean Hb F (%) 61” 29 56.33” 31.31 64.3 ”27.98 0.382 6(-13.0, 31)
Splenctomised (%) 17 (48.4) 7 (46.67) 10 (50) 0.845 0.917(0.227,3.704)
Age at splenectomy (yrs) 8 ( 7.5-10.5) 8 (7-10) 9.5 (7.7-11) 0.364 1(-1.0,3.0 )
Median age at first transfusion (yrs) 6 (4-8) 5.5 (4-8) 7 (4-8.5 ) 0.769 0(-2.0, 3.0)
Median no. of units transfused (Units) 7 (2-16) 7.5 (2.25-24.5) 3 (2-10) 0.376 -4(-15.0,1.0)
Median weight (kg) 29.2 (26.2-37) 31 (26-37) 29 (27-38) 0.681 0.9(-5.0,6.0)
Mean height (cm) 143” 12 134” 13.76 143 ”11.3 0.908 0.35(-9.0, 9.5)
Median Hb last year (gm %) 8.4(7.7-9.3) 8.3 (7.6-9.1) 8.54”1.08 0.419 0.3(-0.5,1.2)
Mean Hb (gm %) 7.5” 1.41 7.5 ”1.1 7.85 (7.35-9.1) 0.283 0.4(-0.4,1.1)
Peak platelet count
(” 10 3 /mm3) 610 ” 320 610” 350 590” 307 0.987 3(-2.5, 16.1)
Mean platelet count
(” 10 3 /mm3) 360” 170 340 ”160 380” 175 0.438 4.2(-10,13.4 )
Mean TLC”103/mm3 7.68”1.62 8.51”1.64 7.05”1.34 0.008 -1500(-2600,-200)
Mean ANC”103/mm3 3.50”1.02 3.55”0.84 3.45”1.16 0.610 -200(-700,500)
Peak ferritin (ng/ml) 425(283-1120) 343(293-980) 580 (260-1435) 0.934 17.5(-281,542)
Peak ferritin last year (ng/ml) 326 (246-840) 326(293-902) 402(230-924) 0.934 9(-163, 398)
Peak n RBC count/mm3 2500 (370-37800) 2000(730-37000) 3300 (234-81000) 0.856 97(-10625,10550)
Mean n RBC count/mm3 2130 (275-16100) 1300(470-16100) 2300 (242-40000) 0.882 97(-2602,4063)
Median protein C (%) 54(47-60) 56(47-71) 49 (42.7-57) 0.214 -6(-16, 3.0)
Mean protein S (%) 69.2 ”22.2 63 ”16.1 73” 25.4 0.254 8.9(-8,24.7)
Mean antithrombin III 94.1 ” 10.86% 93.8 ”13.1% 94” 9.15% 0.780 -1(-8.0,7.0)
Mean hydroxyurea duration (yrs) 7.9 ”3.9 6.08” 3.05 8.3” 3.9 0.122 2(0, 5.0)
Mean hydroxyurea dose (mg /kg/day) 14.7” 2.9 14.8” 2.61 14.5”3.2 0.987 0(-3.0,2.0)
The TIF guidelines had described a set of risk factors for thrombosis in NTDT patients. The association between risk factors for thrombosis and neuroimaging abnormalities was evaluated. The adult age (> 18 years), personal or family history thrombosis, history of splenectomy, anemia (Hb< 9 gm %), thrombocytosis (platelet count> 5,00,000), peak serum ferritin > 800ng/ml, mean nucleated RBC count> 300/mm3, evidence of pulmonary hypertension, deficiency of anticoagulant proteins and diagnosis of beta and E- beta thalassemia was not associated significantly with neuroimaging abnormality. There was significant association between neuroimaging abnormality and mean TLC (P value- 0.008).There was a trend towards significant association with presence of 5 or more risk factors (0.069) and 7 or more risk factors (P value-0.086). The odds of development of neuroimaging abnormality was 6 , 1.39 , 1.33 , 2.9 , 5.4 and 3.4 times with mean Hb < 9 gm%, peak platelet count> 500”103/mm3,nRBC count > 300/mm3, deficiency of anticoagulant proteins, presence of 5 or more risk fact ors and presence of 7 or more risk factors respectively ( table XI)
Table XI- Association of risk factors (as per TIF) for thrombosis and neuroimaging abnormality
Risk factors for thrombosis Abnormal neuroimaging N=15 (%) Normal neuroimaging N=20 (%) P value Odds ratio (95% CI)
Age > 18 years 2 (13.3) 8 (40) 0.088 0.87(0.12-6.003)
Personal or family history of thrombosis Nil Nil –
Splenctomy 8 (53.3) 9 (45) 0.845 0.87(0.225-3.34)
Hemoglobin < 9 gm% 12 (80) 17 (85) 0.08 6.0(0.67-56.52) Thrombocytosis 8 (53.3) 9 (45) 0.625 1.397(0.36-5.35) Ferritin > 800 ng/ml 6 (40) 7 (35) 0.686 0.75(0.18-3.034)
Nucleated RBC count > 300/mm3 10 (66.6) 15 (75) 0.727 1.33(0.264-6.73)
Evidence of pulmonary hypertension 3 (20) 5 (25) 0.727 0.75(0.14-3.791)
Diagnosis of ” & E-” thalassemia 13 (86.67) 20 (100) 0.207 1.83(1.34-2.503)
Deficiency of anticoagulant proteins 6 (40) 14 (70) 0.383 2.92(0.23-35.68)
Transfusion na”ve or minimally transfused 12 (80) 20 (100) 0.382 0.34(0.02-4.17)
5 or more risk factors 13(86.67) 11(40.7) 0.069 5.43(0.94-29.93)
7 or more risk factors 8(53.3) 5(25) 0.086 3.42(0.81-14.36)
Table XII gives the attributable fraction and population attributable fraction of the above risk factors.
Table XII- Attributable and population attributable fraction for neuroimaging abnormality
S. no. Risk factor Odds ratio (95% CI) Relative risk Attributable risk fraction Population attributable risk fraction
1. Hb< 9 gm% 6.0 (0.67-56.52) 1.714 41.6% 36.35% 2. Platelet count > 500”103/mm3 1.397 (0.36-5.35) 1.154 13.3% 6.8%
3. nRBC count > 300/mm3 1.33 (0.26-6.73) 1.125 11.11% 8.7%
4. Deficiency of anticoagulant proteins 2.92 (0.23-35.68) 1.781 43.8% 6.3%
5. 5 or more risk fact ors 5.43 (0.94-29.93) 1.78 43.82% 34.8%
6. 7 or more risk factors 3.42 (0.81-14.36) 1.773 43.59% 22.3%
Cerebral infarction (silent and overt) and their association with risk factors:
The association between cerebral infarction (overt and silent) and risk factors for thrombosis was evaluated. There was no significant association between cerebral infarction and risk factors of adult age (> 18 years), personal or family history thrombosis, history of splenectomy, anemia (Hb< 9 gm %), thrombocytosis(platelet count> 5,00,000), peak serum ferritin > 800ng/ml, mean nucleated RBC count> 300/mm3, evidence of pulmonary hypertension and deficiency of anticoagulant proteins. There was significant association between infarction in imaging and mean TLC (P value- 0.041) and there was a trend towards significant association with presence of 5 or more risk factors (0.055). The odds of development of cerebral infarction were 1.33, 3.88, 1.82, 1.76, 4.4, 1.2, 8.4 & 3.96 times for risk factors of age > 18 yrs, mean Hb < 9 gm%, peak platelet count > 500”103/mm3, nRBC count > 300/mm3, deficiency of anticoagulant proteins, evidence of pulmonary hypertension presence of 5 or more risk factors and presence of 7 or more risk factors and respectively ( Table XII).
Table XIII- Association of risk factors for thrombosis and cerebral infarction
Risk factors for thrombosis Cerebral infarction present N=12 (%) No cerebral infarction N=23 (%) P value Odds ratio(95% CI)
Age > 18 years 2 (16.7) 8 (34.7) 0.771 1.33(0.19-9.31)
Personal or family history of thrombosis Nil Nil –
Splenctomy 7 (58.3) 10 (43.3) 0.904 1.09(0.27-4.408)
Hemoglobin < 9 gm% 10 (84.6) 19 (82.5) 0.213 3.88(0.41-36.78) Thrombocytosis 7 (58.3) 10 (43.3) 0.404 1.82(0.443-7.47) Ferritin > 800 ng/ml 4 (33.3) 9 (39.2) 0.070 0.218(0.039-1.22)
Nucleated RBC count > 300/mm3 9 (75) 16 (69.5) 0.529 1.765(0.297-10.47)
Evidence of pulmonary hypertension 3 (25) 5 (21.7) 0.827 1.2(0.23-6.18)
Diagnosis of ” & E-” thalassemia 10 (84.67) 23 (100) 0.293 1.57(1.21-2.034)
Deficiency of anticoagulant proteins 4 (33.3) 16 (69.56) 0.217 4.4(0.356-54.36)
Transfusion na”ve or minimally transfused 10 (84) 22 (95.6) 0.971 1.04(0.08-12.87)
5 or more risk factors 11(91.66) 13(48.14) 0.055 8.46(1.19-76.89)
7 or more risk factors 7(58.33) 17(62.96) 0.079 3.967(0.905-17.38)
Table XIV gives the attributable fraction and population attributable fraction for cerebral infarction
Table XIV – Attributable and population attributable fraction for cerebral infarction
s. no. Risk factor for cerebral infarction Relative risk Odds ratio (95% CI) Attributable risk fraction Population attributable risk fraction
1. Hb < 9 gm% 1.411 3.882 (0.41-36.78) 29.7% 24.7% 2. Peak platelet count > 500”103/mm3 1.228 1.82(0.443-7.447) 18.5% 9.9%
3. nRBC count > 300/mm3 1.191 1.765(0.297-10.47) 6.27% 1.5%
4. Deficiency of anticoagulant proteins 2.063 4.4(0.356-54.367) 51.52% 8.35%
5. Age 1.111 1.33(0.191-9.311) 9.9% 1.4%
6. Splenectomized 1.030 1.091(0.27-4.408) 2.9% 1.4%
7. Evidence of pulmonary hypertension 1.067 1.2(0.233-6.185) 6.27% 1.5%
8. 5 or more risk factors 1.678 8.462 (1.19-76.89) 40.4% 31.7%
9. 7 or more risk factors 1.673 3.967 (0.90-17.38) 40.26% 20.02%
Silent cerebral infarction and their association with risk factors:
The association between silent cerebral infarction and risk factors for thrombosis was evaluated. It was found not to have significant association with adult age (> 18 years), transfusion regimen, personal or family history thrombosis, history of splenectomy, anemia (Hb< 9 gm %), thrombocytosis(platelet count> 5,00,000), peak serum ferritin > 800ng/ml, mean nucleated RBC count> 300/mm3, evidence of pulmonary hypertension & deficiency of anticoagulant proteins. There was significant association between silent cerebral infarction and presence of 7 or more risk factors (0.05). The odds of development of silent cerebral infarction with the following risk factor is 1.45, 2.4, 2.7, 2.9, 1.5, 5 and 5.4 times respectively with splenectomy, mean Hb, 9 gm%, peak platelet count > > 500”103/mm3, mean nRBC count > 300/mm3, deficiency of anticoagulant proteins, presence of 5 or more risk factors and presence of 7 or more risk factors. Table XVI gives the details of risk factors studied for association with silent cerebral infarction and the odds ratio for the same.
Table XVI- Association of risk factors for thrombosis and silent cerebral infarction
Risk factors for thrombosis Silent cerebral infarction present N=9 (%) No silent cerebral infarction N=26 (%) P value Odds ratio (95% CI)
Age > 18 years 1 (11.1) 9 (34.7) 0.752 0.68(0.06-7.108)
Personal or family history of thrombosis Nil Nil – –
Splenctomy 4 (44.4) 13 (50) 0.627 1.45(0.31-6.69)
Hemoglobin < 9 gm% 8 (88.9) 21 (80.7) 0.439 2.4(0.248-23.26) Thrombocytosis 4 (44.4) 13 (50) 0.208 2.72(0.557-13.65) Ferritin > 800 ng/ml 3 (33.3) 10 (38.4) 0.109 0.146(0.016-1.33)
Nucleated RBC count >300/mm3 6 (66.6) 19 (73.07) 0.33 2.94(0.31-28.027)
Evidence of pulmonary hypertension 2 (22.2) 6 (23.07) 0.958 0.952(0.15-5.561)
Diagnosis of ” & E-” thalassemia 7(77.77) 26 (100) 0.392 1.37(1.16-1.69)
Deficiency of anticoagulant proteins 3 (33.3) 17 (65.38) 0.752 1.5(0.119-18.83)
Transfusion na”ve or minimally transfused 8 (88.9) 24 (92.3) 0.752 0.66(0.05-8.37)
5 Or more risk factors 8(88.88) 16(59.25) 0.128 5.0(0.54-46.28)
7 or more risk factors 6(66.6) 7(26.92) 0.05 5.42(1.05-27.83)
Table XV gives the attributable fraction and population attributable fraction for silent cerebral infarction.
Table XV- Attributable fraction and population attributable fraction for silent cerebral infarction
S. no. Risk factors for silent cerebral infarction Relative risk Odds ratio (95%CI) Attributable risk fraction Population attributable risk fraction
1. Splenectomy 1.102 1.45(0.31-6.69) 9.2% 4.7%
2. Hb< 9 gm% 1.2 2.4(0.248-23.26) 16.67% 13.79% 3. Peak platelet count >500”103/mm3 1.288 2.72(0.557-13.65) 22.36% 12.3%
4. nRBC count >300/mm3 1.243 2.94(0.31-28.02) 19.54% 15.78%
5. Deficiency of anticoagulant proteins 1.125 1.5(0.119-18.83) 11.11% 1.06%
6. 5 or more risk factors 1.364 5.0(0.54-46.28) 26.68% 19.97%
7. 7 or more risk factors 1.604 5.42(1.05-27.83) 37.65% 18.32%
Cerebral arteriopathy and their association with risk factors
On analysis of association between cerebral arteriopathy and risk factors for thrombosis, there was no statistically significant association between cerebral arteriopathy and attributable risk factors : adult age (> 18 years), transfusion regimen, personal or family history thrombosis, history of splenectomy, anemia (Hb< 9 gm %), thrombocytosis(platelet count> 5,00,000), peak serum ferritin > 800ng/ml, mean nucleated RBC count> 300/mm3, evidence of pulmonary hypertension, diagnosis of beta and E- beta thalassemia & deficiency of anticoagulant proteins. There was significant association between presence of arteriopathy and presence of thrombosis (cerebral and deep venous thrombosis) (p value -0.003). When 5 or more of the above risk factors were present there was a significant occurrence of focal cerebral arteriopathy (p value- 0.029). Table XVII gives the association between cerebral arteriopathy and risk factors for thrombosis. The odds of development of cerebral arteriopathy were 2.6, 2, 1.16, 1.78, 20 and 1.5 times respectively with the risk factors of age > 18 years, Hb <9 gm%, evidence of pulmonary hypertension, efficiency of anticoagulant proteins, evidence of thrombosis and presence of or more risk factors (Table XVII). Table XVII- Association of risk factors for thrombosis and cerebral arteriopathy Risk factors for thrombosis Cerebral arteriopathy present N=8 (%) No cerebral arteriopathy N=27 (%) P value Odds ratio (95% CI) Age > 18 years 2 (25) 8 (29.7) 0.324 2.66(0.36-19.03)
Personal or family history of thrombosis Nil Nil – –
Splenctomy 6 (75) 11 (40.7) 0.927 1.07(0.22-5.21)
Hemoglobin < 9 gm% 7 (88.9) 22 (80.7) 0.546 2.0(0.204-19.61) Thrombocytosis 6 (75) 11 (40.7) 0.927 1.077(0.22-5.21) Ferritin > 800 ng/ml 4 (50) 9 (33.3) 0.981 1.02(0.20-5.209)
Nucleated RBC count > 300/mm3 6 (75) 19 (70.3) 0.869 0.85(0.136-5.39)
Evidence of pulmonary hypertension 2 (25) 6 (22.2) 0.869 1.16(0.186-7.34)
Diagnosis of ” & E-” thalassemia 7(87.5) 26 (96.2) 0.428 1.32(1.08-1.60)
Deficiency of anticoagulant proteins 3 (37.5) 17 (62.9) 0.651 1.78(0.14-22.7)
Transfusion na”ve or minimally transfused 6 (75) 26 (96.2) 0.658 0.56(0.04-7.11)
Evidence for thrombosis 7 (87.5) 7 (25.92) 0.003 20(2.07-192.53)
5 or more of the above risk factors 8 (100) 17 (62.9) 0.037 1.5(1.13-1.99)
Table XVIII gives the attributable fraction and population attributable fraction for cerebral arteriopathy.
Table XVIII -Attributable fraction and population attributable fraction for cerebral arteriopathy
S. no. Risk factors for cerebral arteriopathy Relative risk Odds ratio (95% CI) Attributable risk fraction Population attributable risk fraction
1. Age > 18 years 1.333 2.66(0.36-19.03) 24.98% 4.6%
2. Hb< 9 gm% 1.143 2.0(0.204-19.61) 12.52% 10.21% 3. Evidence of pulmonary hypertension 1.037 1.16(0.186-7.34) 3.56% 0.8% 4. Deficiency of anticoagulant proteins 1.172 1.78(0.14-22.7) 14.67% 1.4% Pulmonary hypertension and their association with risk factors The association of pulmonary hypertension with its proposed risk factors was evaluated and it was found that risk factors of adult age (> 18 years), transfusion regimen, personal or family history of thrombosis, history of splenectomy, anemia (Hb< 9 gm %), thrombocytosis (platelet count> 5,00,000), serum ferritin > 800ng/ml, nucleated RBC count> 300/mm3, diagnosis of beta and E- beta thalassemia, presence of thrombosis & deficiency of anticoagulant proteins were not significantly associated with pulmonary hypertension. A mean hemoglobin of < 7 gm% (P value-0.067) over the past 1 year and nucleated RBC count > 300/mm3 was showing a trend towards significant association with pulmonary hypertension (0.06).The odds of development of pulmonary hypertension is 2, 2, 1.78, 1.33, 1.78 and 1.5 times respectively with the following risk factors : Hb <9 gm%, peak serum ferritin > 800 ng/ml, deficiency of anticoagulant proteins, clinical evidence of thrombosis and presence of 5 or more risk factors (Table XIX). Table XIX gives the association between risk factors and pulmonary hypertension.
Table XIX- Association of risk factors and pulmonary hypertension
Risk factors for thrombosis Pulmonary hypertension present N=8 (%) No pulmonary hypertension N=27 (%) P value Odds ratio (95% CI)
Age > 18 years 3 (37.5) 7 (25.9) 0.869 0.82(0.078-8.60)
Personal or family history of thrombosis Nil Nil –
Splenectomy 2 (25) 15 (55.7) 0.927 1.07(0.22-5.21)
Hemoglobin < 9 gm% 6 (75) 23 (85.1) 0.542 2.0(0.204-19.68) Last year mean hemoglobin < 7 gm% 3 (37.5) 2 (7.41) 0.067 0.13(0.018-1.01) Thrombocytosis 3 (37.5) 14 (51.8) 0.927 1.07(0.22-5.21) Ferritin > 800 ng/ml 2 (25) 11 (40.7) 0.392 2.0(0.404-9.909)
Nucleated RBC count > 300/mm3 5 (62.5) 20 (74.07) 0.06 0.174(0.30-0.99)
Diagnosis of ” & E-” thalassemia 7(87.5) 26 (96.2) 0.346 0.26(0.015-4.86)
Deficiency of anticoagulant proteins 5 (37.5) 15 (55.5) 0.651 1.78(0.14-22.76)
Transfusion na”ve or minimally transfused 7 (87.5) 25 (92.5) 0.324 1.33(1.09-1.62)
Evidence for thrombosis 4 (50) 10 (37.02) 0.511 1.7(0.346-8.34)
5or more risk factors 6(75) 18(66.6) 0.656 1.5(0.251-8.97)
Table XIX gives the attributable fraction and population attributable fraction for pulmonary hypertension.
Table XIX-Attributable fraction and population attributable fraction for pulmonary hypertension
S. no. Risk factors for pulmonary hypertension Relative
risk Odds ratio (95% CI) Attributable risk fraction Population attributable risk fraction
1. Hb<9 gm% 1.143 2(0.20-19.68) 12.51% 10.26% 2. Peak ferritin > 800 ng/ml 1.182 2(0.404-9.909) 15.39% 6.32%
3. Deficiency of anticoagulant proteins 1.172 1.78 (0.14-22.76) 14.67% 1.4%
4. Evidence of thrombosis 1.133 1.7(0.346-8.34) 11.73% 5.05%
5. 5 or more risk factors 1.091 1.5(0.251-8.97) 8.34% 5.8%
Association of white blood cell count and absolute neutrophil count with NIA and pulmonary hypertension:
Mean total leucocyte count (TLC) and absolute neutrophil count (ANC) was evaluated for its association with neuroimaging abnormalities including infarction (silent and symptomatic) & cerebral arteriopathy. The mean TLC & ANC in the study cohort was 7680”1628 and 3501”1028 respectively. Mean TLC was significantly associated with the occurrence of neuroimaging abnormalities, cerebral infarction (including silent and asymptomatic) and arteriopathy in MRI imaging. Mean absolute neutrophil count was not significantly associated with neuroimaging abnormality including infarction and arteriopathy. Mean TLC and ANC was not significantly associated with pulmonary hypertension occurrence. Table XXI gives the association between TLC and ANC with neuroimaging abnormality and pulmonary hypertension.
Table XXI-Association of mean TLC & ANC with neuroimaging abnormality and pulmonary hypertension
Clinical parameters P value for mean TLC P value for mean ANC
Neuroimaging abnormality 0.08 0.61
Infarction in MRI 0.041 0.898
Silent infarction 0.323 0.897
Arteriopathy 0.040 0.954
Pulmonary hypertension 0.363 0.428
Multivariate binomial regression analysis of the identified risk factors:
Multivariate binomial regression analysis was performed based on significant factors derived from univariate logistic regression analysis. The output from the multivariate analysis showed significant association between median TLC and neuroimaging abnormality (Pvalue-0.05), single serum ferritin values >800ng/ml and cerebral infarction (both silent and symptomatic) (p value- 0.046). Seven or more risk factors described for thrombosis by thalassemia international federation guidelines (0.028) and serum ferritin value > 800ng/ml (0.034) was significantly associated with silent cerebral infarction. Arteriopathy and pulmonary hypertension had no significant factors derived by multivariate regression analysis. However there was a trend towards significance between documented thrombosis (symptomatic or silent) and focal cerebral arteriopathy (P value-0.098). Similarly there was a trend towards significant association between median TLC and cerebral infarction (P value-0.067).Tables XXII- XX gives the output of multivariate binomial logisitic regression analysis for neuroimaging abnormalities, cerebral infarction in imaging, silent cerebral infarction, focal cerebral arteriopathy and pulmonary hypertension respectively.
Table XXII ‘ Output of multivariate logistic regression analysis for neuroimaging abnormalities
Parameters Nagelkerke R2 Hosmer & Lemeshow test PPV & NPV Odds ratio (95%CI) P value
Adult age 0.393 0.691 60%
85% 0.807(0.043-15.15) 0.886
Mean hemoglobin < 9 gm% 0.589(0.027-12.93) 0.737 5 or more risk factors +ve 0.229(0.026-2.024) 0.185 7 or more risk factors +ve 0.569(0.102-3.177) 0.521 Mean TLC 1.001(1.0-1.001) 0.049 PPV- positive predictive value, NPV- negative predictive value Table XXIII- Output of multivariate logistic regression analysis for cerebral infarction Parameters Nagelkerke R2 Hosmer & Lemeshow test PPV & NPV (%) Odds ratio (95%CI) P value Serum ferritin > 800ng/ml 0.508 0.547 58.3
87 0.072(0.00-0.959) 0.046
5 or more risk factors +ve 0.095(0.006-1.589) 0.102
7 or more risk factors +ve 0.26(0.032-1.897) 0.178
Mean TLC 1.001(1.0-1.001) 0.067
PPV- positive predictive value, NPV- negative predictive value
Table XXIV- Output of multivariate logistic regression analysis for silent cerebral infarction
Parameters Nagelkerke R2 Hosmer & Lemeshow test PPV & NPV (%) Odds ratio (95%CI) P value
Serum ferritin > 800ng/ml 0.450 0.978 55.6
92.3 0.048(0.003-0.723) 0.034
5 or more risk factors +ve 0.128(0.005-3.271) 0.211
7 or more risk factors +ve 16.592(1.235-222.891) 0.028
Platelet counts> 5 lakh/mm3 4.449(0.230-86.04) 0.323
PPV- positive predictive value, NPV- negative predictive value
Table XXV- Output of multivariate logistic regression analysis for cerebral arteriopathy
Parameters Nagelkerke R2 Hosmer & Lemeshow test PPV & NPV (%) Odds ratio (95%CI) P value
Documented thrombosis 0.528 0.882 62.5
92.6 0.130(0.012-1.459) 0.098
5 or more risk factors +ve 0.0(0.00) 0.999
Mean TLC 1.001(1.0-1.001) 0.180
PPV- positive predictive value, NPV- negative predictive value
Table XXVI- Output of multivariate logistic regression analysis for pulmonary hypertension
Parameters Nagelkerke R2 Hosmer & Lemeshow test PPV & NPV (%) Odds ratio (95%CI) P value
Nucleated RBC count > 300/mm3 0.202 0.656 37.3
96.3 3.222(0.401-25.87) 0.271
Last year mean Hb < 7 gm% 3.664(0.333-40.28) 0.288 Discussion A total of 35 patients with NTDT were enrolled and evaluated for neuroimaging abnormalities in our study. The median age of children in our study was 14 (IQR 12-15) years. The mean age of patients included in a study on asymptomatic neuroimaging abnormalities (NIA) is varied being 32.1'' 11 years in an analysis by Taher et al was (1), 29 '' 7.1 years by Manfre et al (2). Karimi et al in 2 consecutive studies in 2010 & 2012 have evaluated patients with a mean age of 24.3 '' 4 years & 23.1''8.2 years (3, 4). Teli et al have assessed younger patients with a mean age of 12 '' 3 years (5). Our centre is a pediatric centre, and we predominantly manage children and early adolescents. This would explain the lower age of patients evaluated. Studies have reported a prevalence of neuroimaging abnormalities to range between 15-60% in literature (Table 1) we got 43% neuroimaging abnormalities in our cohort. One of the reasons for a varied prevalence would be age. Most studies have been done in an older age group. Teli et al, the only study where a younger age group was studied, (mean age of 12 '' 3 years) as we have done got no patient with an abnormality. Age > 35 years is a risk factor for thromboembolic manifestation (6). It can be proposed that thromboembolic manifestations increase with advancing age, though they may begin to occur in late childhood and early adolescent age group as seen in our study. This opens a new debate whether these thromboembolic manifestations are a mere manifestation of adult age alone or they have origin in early age group & the early incidence is dependent on the number of risk factors present. Table I gives the baseline characters and risk factors for neuroimaging abnormalities across different studies.
Table I- Baseline characters and risk factors for MRI abnormalities across different studies (1-5, 7)
Study cohort No. of patients Mean or median age (yrs) Splenectomised
(%) No. of patients with MRI abnormality
(%) Transfusion history Time since splenectomy
Taher et al 30 32.1” 11 100 60 40% occasional* 17” 9.9 13:17
Manfre et al 16 29 NA 37.5 NA NA 11:5
Karimi et al (2010) 30 24.3” 4.3 100 26.7 16% occasional* NA 11:19
Teli et al 24 12”4.6 25 0 16% occasional*
Karimi et al (2012) 95 23.1”8.2 62.5 15.8 48% regular^ NA 32:63
Our study 35 14(12-15) 48 43 64% occasional*
8% regular^ 6 (4.5-6) 29:6
*- transfusion 1-2 times /year, ^- transfusion every 1-3 months for a pretransfusion Hb ‘ 9 gm% (6).
Our study cohort included both splenctomised and nonsplenectomised patients. The median age at splenectomy in our cohort was 8 (IQR 7.5-10.5) years. The average age of splenectomy was lower in our cohort compared to the cohort in the study by Taher et al (1) where the age was around 15 years. The median age for occurrence of thromboembolic events post splenectomy has been observed to be 8 years (8). In our cohort the median time post splenectomy was 6 years compared to 17” 9.9 years in the cohort of Taher et al. Inclusion of non-splenectomised patients, earlier splenectomy and shorter time post splenectomy in our cohort could be a reason for lower incidence of neuroimaging abnormalities compared to the study by Taher et al (1). It can be inferred that splenectomy is a significant risk factor for MRI abnormality with an increasing cumulative incidence with increasing time, more so > 8 years post splenectomy (8). Probably, abnormalities would occur earlier than 8 years whenever there are other risk factors.
Twenty eight percent patients in our cohort were transfusion na”ve, 8% had been on regular transfusion for the first 2 years from diagnosis, in outside hospitals for a diagnosis of thalassemia major. Occasional transfusion was received by 64% of patients. Blood transfusional policy for NTDT has been different across different studies. The cohort of our study and in the study by Taher et al and Karimi (2010) et al did not have any patients on regular transfusion. All these 3 studies had more number of neuroimaging abnormalities than the other reported studies, though this conclusion is limited by the number of studies & lack of data on transfusional policy in the other studies. Transfusion naivety was quoted as a significant risk factor over occasional transfusion/regular transfusion for MRI abnormality by Taher et al (1). In our study it was found that occasional transfusion was not protective against MRI abnormalities, as 50% of occasionally transfused patients had MRI abnormalities. Regular transfusion was thought to be protective to development of neuroimaging abnormalities as observed by Karimi et al (3) as their prevalence of abnormalities was low. However, there is no clear evidence as to the protective effect of regular/occasional transfusion in these patients. The best transfusion policy to protect against neuroimaging abnormalities needs further research.
Hemoglobin >9 gm% was found to be protective against thromboembolic events (6). In our study mean hemoglobin value > 9 gm was present in only 1 patient in the group with neuroimaging abnormalities. 30% patients without neuroimaging abnormalities had hemoglobin > 9 gm%. Though the mean hemoglobin was not significantly associated with MRI abnormalities, the mean hemoglobin was lower in patients with MRI abnormalities. The peak platelet count was significantly elevated as expected in patients who had undergone splenectomy. The peak platelet count was > 5 lakh/mm3 in 8/15 & 9/20 patients with and without MRI abnormality respectively. The peak platelet count was lower than that observed in study by Taher et al (1) & Karimi (2010) et al (4). Since both the studies included patients post splenectomy they had a higher mean platelet count. Peak serum ferritin > 800 ng/ml was present in 13 patients including 5 (38%) patients with MRI abnormality in our study. The median serum ferritin was lower than that reported by Taher et al (1), Karimi (2010) et al (3) and karimi (2012) et al (3). The probable reason for this observation could be due to less number of children who were on regular blood transfusion and lower median age of the study cohort. There were 78% patients with elevated nucleated RBC in their peripheral blood. Of them 12(44.4%) had MRI abnormalities. The mean nucleated RBC count was lower in our cohort compared to the cohort studied by Taher et al (1). This could be conceivably due to younger age of our cohort, all patients being on hydroxyurea therapy and lesser number of patients post splenectomy. Table II gives the comparison of laboratory parameters across various studies. It can be inferred from table II, that the number of pathological RBC, platelet and toxic oxygen species derived elevated free iron are the major determinants of thromboembolic events. Splenectomy leads to elevation of all 3 above parameters thereby leading to higher thromboembolic events. However, this has not translated into a higher incidence of MRI abnormality in studies by Karimi et al (3, 5) despite having higher values for the above parameters. There must be some other unknown players, in addition to the factors mentioned above, for the pathogenesis of these thromboembolic events.
Table II- Comparison of laboratory parameters across various different studies (1-5, 7)
Study cohort Mean Hb (gm %) Mean peak platelet count (” 10 3 /mm3) Mean nucleated RBC count (” 10 3 /mm3) Mean peak serum ferritin (ng/ml) Prevalence of neuroimaging abnormality (%)
Taher et al 8.6”2.1 791.2”355.3 367.4”319.8 1176”641.9 60
Karimi et al (2010) 8.4 879 NA 519 26.7
Karimi et al (2012) 8.8”1 567”312 NA 1404”1499 15.8
Teli et al 9.3”1.3 388(251-1221) NA 66(20-927) 0
Our study 7.5” 1.41 610” 320 2.5 (0.37-37.800) 425(283-1120) 43
Another interesting feature was total leucocyte count which was significantly higher in children with neuroimaging abnormality (P value-0.007). White blood cells have been contemplated as a risk factor for silent cerebral infarction in patients with sickle cell anemia. Activated white blood cells are known to express more adhesion molecules, produce more procoagulant particles and proinflammatory cytokines. This sets off an inflammatory state activating a multitude of cells including endothelial cells, platelets and RBC. This has been seen to contribute to a hypercoagulable state in sickle cell anemia (9). A study by Kinney et al found total leucocyte count > 11.8”103/mm3 to be a risk factor for silent cerebral infarction in sickle cell anemia (10). The mean total leucocyte count for patients with neuroimaging abnormality was 8.51”1.61”103/mm3 vs. 7.05”1.34”103/mm3 in patients without neuroimaging abnormality. Hydroxyurea therapy lowers the total leucocyte count and decreases the tissue factor expression by monocytes. This might explain the lower total leucocyte count in the study, where the entire cohort was on hydroxurea therapy for 7.9” 3.9 years. Further studies are required to evaluate the association between WBC and hypercoagulability in NTDT.
In our cohort, the anticoagulant proteins, protein C and antithrombin III, were lesser than the age appropriate cut off in 54.2% and 8.57% patients respectively. No patient had low protein S. There is no age specific cut off for these anticoagulant proteins for Indian children and the reference value was utilized from a study by Appel et al and Andrew et al(11, 12). Lack of age specific cut off for Indian children is a limitation to this data. Anticoagulant proteins were lower in other studies in thalassemic patients. In study by Teli et al in thalassemia intermedia patients they observed 54.6%, 45.8% and 4.1% patient with lower protein C, protein S and antithrombin III levels. Their prevalence of low anticoagulant protein levels is higher than that seen in our study, mainly with reference to protein S. Without the presence of normative data it is difficult to compare results of protein level between different geographic areas. Neither of the studies could demonstrate a significant association between anticoagulant protein levels and silent cerebral infarction (5). Table III gives the prevalence of low anticoagulant protein levels across different studies. The role of anticoagulant proteins deficiency in hypercoagulability still needs to be evaluated after generating age appropriate reference values for a given geographic location.
Table III- Prevalence of low anticoagulant levels across different studies (5, 13)
S.No Study cohort Study population Mean/Median protein C levels % Mean/Median protein S levels % Mean/Median antithrombin III levels %
1 Our study NTDT 54 (47-60) 69.3”22.2 94.1”10.86
2. Teli et al NTDT 70”15 66.5”14.4 92.4”1.4
3. Shirahata et al Thalassemia major and NTDT 50.4”17.2 58.8”25.5 78.8”12.1
Total prevalence of silent cerebral infarction, symptomatic cerebral infarction and cerebral arteriopathy was 25.7%, 8.5% & 22.8% respectively. Overt cerebral infarction has been described predominantly as isolated case report in literature (14-16). Prevalence of 8.5% is an alarm signaling the seriousness of the concern of hypercoagulability in NTDT patients.Development of an action plan for management of patients with silent cerebral infarction is required.
The prevalence of silent cerebral infarction in the general population of pediatric and adolescent age group is not known. The prevalence of silent cerebral infarction has been 0 to 11% across various studies in adult population (1, 17, 18). It is mentioned that the incidence tends to escalate as age advances (17, 18). As a corollary, the prevalence of NIA in children and young adolescents should be more than the expected in the general population in the same age group. The risk factors for hypercoagulability state associated with thalassemia & thalassemia per se are in all probability responsible for the increased prevalence of NIA. Table IV gives the various studies on the prevalence of cerebral infarction in the general population
Table IV- Prevalence of silent cerebral infarction in general population and NTDT patients (1, 17-19)
S.No. Study cohort No. of patients Mean/Median/age range of cohort No. of patients (%) Type of population
1. Our study 35 14 (12-15) 9 (25.7) NTDT
2. Taher et al 30 31.9”11 18 (60) NTDT
3. Karimi et al (2010) 30 24.3”3.49 8 (26.7) NTDT
4. Hopkins et al 243 16-65 (5.3) General population
5. Vernooij et al 2000 63.3(45.7-96.7) 145 (7.2) General population
6. Katzman et al 1000 3-83 (0.8) General population
Table V- Characteristics of silent cerebral infarction lesions across different studies (1, 3, 4)
S.no. Lesion characteristics Our study (n=35) Taher et al
(n=60) Karimi (2010) et al
(n=30) Karimi (2012) et al
1. No. of patient with silent cerebral infarction (%) 9 (25.7) 18 (60) 8 (26.7) 15 (15.8)
2. Size of lesions- Small
3. Number of lesions- Single
4. Location of lesions-
NA- not available
Cerebral arteriopathy was seen in 22.8% of our cohort. In a study by Taher et al the prevalence was 27.6% (20). In our study, interestingly, 75% patients with arteriopathy had coexistent cerebral infarction. Twenty five percent of patients with arteriopathy had symptomatic infarction and there was concordancy rate between site of infarction and site of arteriopathy in 50%. This is higher than the rate (5.5%) observed by Taher et al (20). Of the concordant lesions 50% were symptomatic, including a patient with moyamoya disease and a patient with 3 vessel disease.
Table VI- Comparison of the cerebral arteriopathy in our study with cohort of Taher et al (20)
S.No. Characteristic finding Taher et al cohort (n=29) Our cohort (n=35)
1. Prevalence of arteriopathy (%) 8 (27.6) 8 (22.8)
2. No. of lesions 12 10
3. Severity of lesions ‘ Mild
Total occlusion 9
4. Co-existent arteriopathy and infarction 3 patients 6 patients
5. Concordancy rate Nil 4 patients
6. Symptomatic infarction Nil 2 patients
7. Vessel affected- Internal carotid artery
Middle cerebral artery
Anterior cerebral artery
Posterior cerebral artery
2 lesions 1 lesion
We could infer that NTDT patients are also at risk of symptomatic infarction secondary to cerebral arteriopathy. It is difficult to explain the reason for the rest 50% concordant lesions which were silent cerebral infarction, because it is believed that the cerebral infarction are silent as they are sequlae of involvement of smaller arteries resulting in infarction of smaller brain tissue. But involvement of larger arteries leading to smaller infarction & the effects of cerebral arteriopathy and their role in silent cerebral infarction needs evaluation . Table IV gives the details of comparison of cerebral arteriopathy by Taher et al and our study.
Pulmonary hypertension was seen in 22.8% of our cohort. Aessoposs et al had reported a prevalence of 60.9% & 23% in their cohort published in 2001 & 2005 respectively (21, 22). Table VII gives the baseline characteristics and treatment history of children in different cohorts studied for pulmonary hypertension. The prevalence of pulmonary hypertension varies between 10 and 78% (23). Lower age of our cohort and lesser splenectomized patients in our cohort could be the reason for lower prevalence of pulmonary hypertension. There is a need for serial follow up of our cohort to see development of pulmonary hypertension.
Table VII- Baseline characteristics and treatment history of children studied for pulmonary hypertension across different studies (6, 21, 22)
S.No. Baseline characteristics and treatment details Our cohort
(n=35) Aessopos et al 2001 (n=123) Aessopos et al 2005 (n=141) Optimal care study (n=584)
1. Mean/median age of cohort (years) 14(12-15) 25.4”13.86
2. Splenectomy rate % 48 55.5 42 55.7
3. Transfusion history-
4. Prevalence of pulmonary hypertension % 22.8 60.9 23 11
5. Symptomaticity rate % Nil 5.4 NA NA
Mean total leucocyte count was significantly associated with NIA and showed a trend towards association with cerebral infarction in multivariate regression analysis. White blood cells produce significant inflammatory cytokines which activates platelets, endothelial cells and increases reactive oxygen species production in patient with sickle cell anemia (9). The same mechanisms may be operating in NTDT. The impact of white blood cells in hypercoagulable state in NTDT and cerebral infarction needs to be researched. Serum ferritin > 800ng/ml and presence of 7 or more risk factors as defined by Thalassemia international federation (TIF) guidelines for NTDT were significantly associated with silent cerebral infarction. This suggests that as the number of risk factors proposed for hypercoagulability increases the risk for silent cerebral infarction escalates.. It can be concluded that the 2 most important determinants of hypercoagulability and silent cerebral infarction are: number of risk factors operating and time for which their operating in an individual. Increase in patient age & duration of exposure to risk factors will increase the incidence of neuroimaging abnormality and pulmonary hypertension.
1. Total of 35 NTDT children with median age of 14 years were studied.
2. Splenectomy was performed in 48% of study cohort.
3. Thirty two percent patients were transfusion na”ve and 68% were transfusion non-na”ve.
4. Eighty two percent of cohort had mean hemoglobin < 9 gm%. 5. Thrombocytosis was present in 48% of cohort including 23%. 6. Mean serum ferritin was > 800ng/ml in 37% of cohort. Seventeen percent of cohort had NIA and serum ferritin above the defined cut off value.
7. Nucleated RBC count was > 0.3”103/mm3 in 71% of the cohort with 28% of the cohort had both NIA and nRBC count above the defined cut off value.
8. Fifty four percent of the cohort had serum protein C & 8.4% of the cohort had low serum antithrombin III below the age appropriate cut off. None of the patients had low serum protein S than the age appropriate cut off.
9. Neuroimaging abnormalities were prevalent in 45% of the study cohort.
10. The prevalence of silent cerebral infarction in the study cohort was 25.7%
11. Cerebral arteriopathy and pulmonary hypertension was observed in 22.8% of the study cohort.
12. Mean white cell count was significantly associated with NIA (P value-0.049).
13. Serum ferritin value > 800ng/ml was significantly associate with cerebral infarction (P value-0.046)
14. Serum ferritin > 800ng/ml and presence of 7 or more risk factors as defined by Thalassemia international federation guidelines for NTDT was significantly associate with occurrence of silent cerebral inafrction.
15. There was no significant association between the defined risk factors and presence of focal cerebral arteriopathy and pulmonary hypertension.
16. Our cohort was younger than the other cohorts described in literature for evaluation of NIA in NTDT patients.
17. The occurrence of NIA has its onset in childhood and early adolescence. Screening for the same should be initiated since early adolescent age group.
18. The odds of development of neuroimaging abnormality was 6, 1.39, 1.33 & 2.9 times with a mean Hb of < 9 gm%, peak platelet count> 500”103/mm3, nRBC count > 0.3”103/mm3 & deficiency of anticoagulant proteins.
19. Presence of 5 or more risk factors had an odd’s ratio of 5.43 and presence of 7 or more risk factors had an OR of 3.42 for the development of NIA.
20. Overt cerebral infarction was present in 8% of the study cohort.
Hypercoagulability is operable in patients of NTDT since childhood and occurrence of its effect depends on the number of factors predisposing to thrombosis existing simultaneously and the duration for which these factors operate simultaneously. In our analysis, the odds of development of neuroimaging abnormality was 6 , 1.39 , 1.33 , 2.9 , 5.4 and 3.4 times with a mean Hb of < 9 gm%, peak platelet count> 500”103/mm3,nRBC count > 300/mm3, deficiency of anticoagulant proteins, presence of 5 or more risk factors had an odd’s ratio of 5.43 and presence of 7 or more risk factors had an OR of 3.42 for the development of NIA. Hypercoagulability in NTDT patients is not just contributed by platelets, RBC, microparticles and nucleated RBC count, but also contributed by white blood cells (significant in our study ( p<0.008)), which is presently not included as a risk factor for hypercoagulability.The contributing mechanism of white blood cells in pathogenesis of hypercoagulability in NTDT patients, possible role towards the development of cerebral arteriopathy & cerebral infarction needs further systematic investigation. Further follow up of this cohort for further development and progression of thrombotic lesions needs to be performed. Our cohort teaches us that the complications described in NTDT patients tend to have their onset in late childhood and early adolescence. Patients > 10 years of age are at risk of these complications and there is a need to develop consensus statement to screen from a younger age to identify these complications.
Strengths and limitations
1. Our study was prospective observational study.
2. All the risk factors quoted in literature to be associated with hypercoagulability and neuroimaging abnormalities were investigated in our cohort of NTDT patients.
3. Predominant population in our study was children and early adolescents.
4. Patients were enrolled randomly, thereby eliminating selection bias.
5. Identified the importance of earlier age of screening for these complications in NTDT patients.
6. Study population comprised of children native of majority of north Indian states.
1. Lesser number of study populations were adults
2. No data on normative values for anticoagulant proteins in our Indian children and adolescents
3. Lack of data on neuroimaging abnormalities in children and early adolescents in general population.
4. T2* MRI was not performed for evaluation of iron overload status in NTDT children.
5. The number of patients were studied were limited.
...(download the rest of the essay above)