Introduction

Thrombosis is a well-recognized complication of nephrotic syndrome (NS). The incidence of thromboembolic complications (TECs) in NS in children is reported to be approximately 3 % [1]. This percentage, however, may be an underestimate of the true incidence as unless suspected and investigated, many such events apparently go unrecognized. No prospective longitudinal studies to detect clinical or subclinical thromboembolic events are available in the pediatric age group. We describe here the clinical profile and outcome of 35 thrombotic events seen in 34 children with NS over a period of 7 years.

Patients and methods

We reviewed the case records of all children diagnosed as having NS with TECs who were admitted to the Advanced Pediatrics Centre, Post Graduate Institute of Medical Education and Research, Chandigarh, India from January 2004 to December 2010. The center serves as a tertiary care referral center for North West India and registers approximately 250–300 new children with NS every year. The management protocol is based on the Consensus Statement on Management of Nephrotic Syndrome by the Indian Pediatric Nephrology Group [2–4] and all the patients analyzed in this report were hospitalized and investigated accordingly.

Cerebral venous thrombosis (CVT) was diagnosed by contrast-enhanced computed tomography (CT) of the brain followed by CT venography or magnetic resonance venography studies. Pulmonary thromboembolism (PTE) was clinically suspected if children developed unexplained respiratory distress, cough, wheeze or hypoxemia. The initial investigation in such cases was a lung perfusion scan after an intravenous injection of 74 MBq of 99mTc-MAA. The results were interpreted as high, moderate or low probability of PTE. High probability was defined as ≥2 perfusion defects in the presence of normal ventilation. CT angiography of the chest was performed in children with moderate probability on ventilation perfusion scans. Deep venous thrombosis (DVT) was detected by ultrasound Doppler of the limb. Anticoagulation therapy was initiated as soon as TECs were identified. In children with intracranial arterial involvement, acetyl salicylic acid was used. Children received unfractionated heparin (70–100 U/kg) for the first 4–5 days and later overlapped with warfarin for 48–72 h. Oral warfarin was maintained as a single daily dose for a period of 6–12 months. No child received fibrinolytic therapy. Tests to identify procoagulant state were carried out during follow-up after stopping anticoagulant therapy at the Coagulation Laboratory of the Department of Hematology, PGIMER, Chandigarh. Thrombophilic work-up was not performed at admission. The tests included functional activity of protein C and protein S, antithrombin III functional activity, lupus anticoagulant, anticardiolipin immunoglobulin (Ig) G and IgM, and DNA testing for Factor V Leiden. Tests for anti-β2-glycoprotein were not available during this period. Our laboratory is registered in the World Health Organization’s United Kingdom National External Quality Assurance Scheme for Coagulation and Immunochemistry for quality control.

Results

During the study period a total 34 children (22 boys and 12 girls) had 35 thrombotic events. The baseline characteristics of these children are summarized in Table 1. CVT was the commonest complication seen in 11 (31.4 %) children followed by PTE in 9 (25.7 %) and DVT involving the peripheral limb arteries in 5 (14.2 %) children. Superior vena caval thrombosis was seen in one child. Arterial thrombosis resulting in central nervous system infarcts was observed in 7 (20 %) children and 2 children had thrombosis of the peripheral arteries (Table 2).

Table 1 Baseline characteristics of the children at the time of presentation (n = 35)
Full size table
Table 2 Frequency of various thromboembolic complications, clinical profile and outcome
Full size table

Patient characteristics

Age

The mean age of the children at the time of admission for the event was 7.7 ± 2.7 years (range 2.5–12 years) while the mean age at the onset of NS was 5.9 ± 4.2 years (Table 1).

Type of nephrotic syndrome

Of the 35 events, 11 episodes occurred in children with steroid-resistant NS (SRNS) and the same number of events were recorded in children with steroid-dependent/frequently relapsing NS (SDNS/FRNS) (Table 1). 7 episodes were seen in children with infrequently relapsing NS but during a relapse of disease. Notably, we recorded TECs during the first episode of NS in 5 children and it was the predominant presenting complaint in 3 of them.

Risk factors

The following factors were associated with risk of thrombotic event:

  1. (i)

    Degree of proteinuria

  2. (ii)

    Serum albumin levels

  3. (iii)

    Associated infection

  4. (iv)

    Thrombocytosis

  5. (v)

    Anemia

  6. (vi)

    Hemoconcentration

  1. (i)

    Degree of proteinuria: Proteinuria (≥40 mg/m2/h) at the time of thrombotic phenomena was seen in the majority (82.8 %) of children; however, 2 children (18.1 %) had sub-nephrotic range proteinuria, while 4 children (8.5 %) developed these complications while in remission. Of these 4 children, 3 (27.2 %) developed CVT while 1 child was diagnosed with intracranial arterial thrombosis (Table 2).

  2. (ii)

    Serum albumin levels: The mean serum albumin level in the cohort was found to be 1.7 ± 1.2 g/l and hypoalbuminemia was seen in 82.8 % of children (Table 1).

  3. (iii)

    Infection: Concurrent focus of infection, an important predisposing factor for TECs was found in 31.4 % of children. The site of infection identified was bacterial peritonitis in 17.1 % of children, cellulitis in 5.7 %, bacteremia in 5.7 % and meningitis in 2.8 %. Streptococcus pneumoniae as well as Staphylococcus aureus were isolated on blood culture in each of 2 children. Co-infection was identified as a risk factor in 83.3 % of children with DVT, and 55.5 % with PTE, but was identified in only 1 child (9 %) with both CVT and intracranial arterial thrombosis (Table 2).

  4. (iv)

    Other factors predisposing to thrombosis like thrombocytosis and anemia were seen in 28.5 and 14.2 % of children, respectively. Hemoconcentration and azotemia were seen in 22.8 and 28.5 % of children, respectively (Table 1).

Tests for inherited thrombophilic conditions and antiphospholipid antibodies (APLAs)

These tests were performed 6 weeks after stopping anticoagulation in 18 children. 2 children with CVT were identified as having a functional protein S deficiency along with NS. The mother of one of the children had a history of DVT during pregnancy and on further evaluation was found to be protein S deficient.One child with PTE had low antithrombin functional activity. APLAs were found in 4 children who had intracranial arterial thrombosis—lupus anticoagulant was positive in 1 child while IgM anticardiolipin antibodies were positive in the other 3 children. On repeat testing, these were found to be negative in 2 children after 1 year of follow-up.

Clinical profile

The clinical presentations included:

  1. (i)

    Cerebral venous thrombosis

  2. (ii)

    Intracranial arterial thrombosis

  3. (iii)

    Pulmonary thromboembolism

  4. (iv)

    Deep venous thrombosis of lower limbs

  5. (v)

    Thrombosis of deep vessels of neck

  6. (vi)

    Thrombosis of peripheral arteries

  1. (i)

    CVT: Headache was the most important presenting symptom in all the children with CVT, followed by vomiting in 81.1 % of children (Table 3). Focal seizures (18 %) and alteration in sensorium (18 %) occurred less commonly. A fundal change in the form of papilledema was found in 7 children (63.3 %) and was an important clue to diagnosis. Lateral rectus gaze palsy was seen in 4 (36.3 %) children. CT was the first line of investigation performed in all children. Sinus thrombosis in the form of empty delta sign was evident on contrast-enhanced CT (CCT) in all children (Fig. 1). Venography studies either by magnetic resonance imaging (MRI) or CT were performed for better delineation of venous sinuses in 54.5 % of children. Superior sagittal sinus thrombosis was the commonest sinus involved in 72.7 % of children followed by transverse, sigmoid and straight sinus. Parenchymal lesions in the form of venous infarct or bleeding were found in 4 (36.3 %) children.

    Table 3 Clinical profile of children with cortical venous thrombosis (n = 11)
    Full size table
    Fig. 1
    figure 1

    Axial contrast-enhanced cranial CT shows a filling defect in the region of the venous confluence (empty delta sign)

    Full size image
  2. (ii)

    Intracranial arterial thrombosis: This was less commonly seen compared to venous thrombosis. 2 children with arterial thromboembolic phenomenon presented with an acute stroke/hemiparesis-like event (Table 4). The majority of the other children (6) had seizures at the time of presentation; CT of the brain revealed parenchymal infarcts in all these children. Diffusion-weighted MRI in 3 children followed by MR angiography in 2 children revealed involvement of the basilar, vertebral and posterior cerebral artery in one child and involvement of the left middle cerebral artery in the other child. Anticoagulation therapy was initiated in 4 of these 7 children and the remaining 3 children were given acetyl salicylic acid (3–5 mg/kg/day) only.

    Table 4 Clinical profile of children with intracranial arterial thrombosis (n = 7)
    Full size table
  3. (iii)

    PTE: It was demonstrable in 9 children, all boys (Table 5) with a mean age of 7.8 years, the youngest being 3.5 years. PTE was generally seen 1–9 days after hospitalization in children who had presented with relapse of NS, or with edema and heavy proteinuria. Breathing difficulty was seen in 77 % of children followed by cough in 33 %. Chest pain was observed in only 11 % of children. Tachypnoea and hypoxia (pulse oximetry ≤92 %) were the most consistent signs seen in 66 % of children and wheeze was identified in 33 % of children. Some children (55.5 %) were admitted for treatment of infection. Bacterial peritonitis was identified in three children and pneumococcal meningitis in one child. Of the 9 children, echocardiography was performed in 5 children and was found to be normal. D-dimer, a surrogate marker for thrombosis, was performed in only 5 children, of whom 4 were positive. Doppler ultrasound of the lower limbs was performed in all children but was found to be normal. Ventilation perfusion scans revealed high probability in 5 and intermediate probability in 2 children. The left lung was involved more frequently than the right lung. CT angiography performed in 2 children showed involvement of the left major pulmonary artery in both children.

    Table 5 Clinical profile of children with pulmonary thromboembolism (n = 9)
    Full size table
  4. (iv)

    DVT: This was seen in 5 children, of whom 4 had SRNS (Table 6). All children had presented in relapse of NS. Associated infection in the form of cellulitis of the lower limbs was identified in 2 children, spontaneous bacterial peritonitis in 1 child and septicemia in 1 child. Blood culture was positive in 3 children. Femoral veins were most commonly involved with unilateral and bilateral involvement in each of 2 children. Thrombus extended to involve the inferior vena cava, and common and external iliac vein in 1 child. Thrombosis of the upper limb vessels involving the right brachial and axillary vein occurred in 1 child on day 2 of hospitalization while admitted for the treatment of pneumococcal sepsis; a history of peripheral venous cannulation in the same limb was obtained. All the children recovered following anticoagulation and treatment for infection. Ventilation perfusion scans did not reveal any PTEs.

    Table 6 Clinical features of children with deep vein thrombosis lower limb (n = 5) and superior vena cava (n = 1)
    Full size table
  5. (v)

    Thrombosis of deep veins of neck: Thrombosis of the superior vena cava and left internal jugular vein was seen in 1 child who presented with chylopericardial tamponade. This child has been reported previously and she had improved following surgical drainage, nutritional rehabilitation and conservative management with conventional heparin [5].

  6. (vi)

    Thrombosis of peripheral arteries: Involvement of the peripheral limb arteries (posterior tibial artery and the brachial artery) was seen in only 2 children; both children developed progressive gangrene of the limb and a history of arterial puncture prior to the development of gangrene was elicited in both children. Thrombectomy was performed in one child; however, the gangrene progressed and the child succumbed to septicemia and progressive gangrene. The other child recovered completely after initial anticoagulation with conventional heparin and later acetyl salicylic acid was given for a total of 3 months.

Outcome

Final outcome was good with complete recovery seen in 32 (91.4 %) out of 35 events. All children with central nervous system thrombosis as well as with DVT improved. Moreover, children in whom a diagnosis of PTE could be established antemortem also recovered. However, 3 children died in our cohort, of whom 2 had sudden death. Autopsy was performed in both children with sudden death which revealed left main pulmonary artery thrombosis in one child and bilateral pulmonary arterial thrombosis in the other child.

Discussion

Thromboembolism is a well-known but rare complication of pediatric NS. The true frequency of TECs in NS is difficult to determine as many events are subclinical/asymptomatic and go undiagnosed [6]. The reported incidence of TECs in adults with NS ranges from 9−70 % [7] while the frequency of clinically evident TECs in children, as reported in various studies is only between 1.8 and 4.4 % [8–11]. The incidence of subclinical pulmonary embolism by scintigraphic pulmonary ventilation and perfusion studies was found to be 28 % in children with NS [6]. The coagulation disturbances in children are as severe as in adults with NS [6, 12]. We recognized 35 events over a period of 7 years. To the best of our knowledge, this is the largest single center experience of TECs in children with NS.

The hypercoagulable state in NS is multifactorial, attributed predominantly to urinary loss of anticoagulants, increased procoagulatory activity, altered fibrinolytic system, thrombocytosis, and enhanced platelet activation and aggregation [6, 12, 13]. Other thrombophilic factors which also contribute are low albumin, increased cholesterol, associated infections and iatrogenic volume depletion due to inappropriate and overuse of diuretics, venepuncture and immobilization. There are few studies on the contribution of genetic prothrombotic defects in addition to acquired risk factors. Fabri et al. [14] evaluated 53 children with NS for prevalence of the Factor V mutation Arg506–>Gln (Factor V Leiden), the prothrombin variant (20210G–>A), and homozygosity for Ala677–>Val in the methylenetetrahydrofolate reductase gene and concluded that inherited thrombophilia is not a strong risk factor for the development of nonrecurrent thrombosis in children with NS [14]. Martinez and co-workers [15] also investigated the presence of genetic prothrombotic factors in patients with glomerulonephritis with or without a history of TECs and/or NS. They found an increased prevalence of heterozygous Factor V Leiden in patients with a history of thrombotic events [15].

Identification of these conditions is of clinical significance as the management is different especially with regards to the duration of anticoagulation. Among 18 children who were evaluated, we identified two children with associated inherited protein S deficiency, and one child with antithrombin III deficiency. The role of APLAs, especially with congenital NS (CNS) thrombosis, has been recognized previously [16] but its association with childhood NS has not been well recognized. We found 4 children with APLAs with intracranial arterial lesions. However, with this limited information based on single APLA test, it would be imprudent to suggest a cause-and-effect relationship between intracranial artery thrombosis and the presence of APLAs. We screened for antinuclear antibodies in all children and therefore lupus (the most important secondary cause for APLAs) can be ruled out. Therefore, our results suggest that there could be some correlation between APLAs and intracranial arterial thrombosis; however, this needs further studies in a larger group of patients.

Few centers have reported cases of CVT occurring in patients with NS. In a pooled literature analysis [17], only 21 pediatric cases were found. We treated 11 cases with a good outcome; however, it was the commonest complication encountered in our cohort. CVT was seen during the first episode as the presenting manifestation and/or was found to occur during relapse and in remission. Unlike other types of TECs, concurrent infection was not found to be a necessary trigger for CNS thrombosis. The single most sensitive symptom of CVT in our study was persistent headache. We suggest that unexplained headache in children with NS should arouse clinical suspicion of CVT warranting prompt investigations. Superior sagittal sinus was the commonest sinus involved in our children, similar to Fluss et al. [17] who also reported its involvement in all the patients in their study. Even with parenchymal lesions seen in four children the outcome was excellent.

Posterior reversible encephalopathy syndrome (PRES) is an important differential diagnosis in children with kidney disease presenting with seizures, headache and altered sensorium. Diffusion-weighted cranial MRI provides accurate diagnosis [18, 19]. Of 7 children with intracranial arterial lesions, diffusion-weighted MRI was performed in 3 children. PRES could be a possibility in other children in whom detailed imaging studies were not performed.

PTE proved to be the most important life-threatening complication and was identified as the cause of death on autopsy in two of our children. It is to be noted that signs and symptoms of this complication are subtle and require prompt recognition to prevent fatal outcomes. Unexplained tachypnoea, hypoxemia or abnormal respiratory findings should alert the physician towards this possibility [11, 20, 21]. According to the recent Prospective Investigation of Pulmonary Embolism Diagnosis III criteria, ventilation perfusion scan is still the investigation of choice to confirm a diagnosis of PTE [22]. Hence, it is a preferred modality over the more invasive CT angiography. Early ventilation scans help in the diagnosis and improving the outcome [23].

TECs are generally venous, whereas arterial thrombosis occurs less frequently. Literature on arterial thrombosis in NS is scarce. Acute subclavian and brachial artery thrombosis has been described by Witz et al. [24] while Siddiqi et al. [25] reported sustained acute thrombosis of the arterial bypass grafts in two patients with NS. Arterial punctures and associated infection predispose to this complication and were identified as additional risk factors in both cases.

Superior vena cava thrombosis has been rarely described and can result in chylopericardium and chylothorax [5, 26]. Renal vein thrombosis, also a well-known complication, is seen more frequently in adults and in patients with membranous nephropathy [27–30] while case series in children do not report it to be a frequent occurrence [9, 12]. It has been conspicuous by its absence in our series. This is one of the major differences between our study and those reported from developed countries. We do not have an adequate explanation to explain this difference but it could perhaps be a reflection of differences in genetic background between our patients and those reported from the Western world. DVT was easily detected by available methods but further tests to detect renal vein thrombosis in all the children with DVT were not performed.

The therapeutic approach to thrombosis in children is with anticoagulants, (conventional heparin infusion/low-molecular-weight heparin) with/or without fibrinolytic agents (streptokinase, urokinase, tissue plasminogen activator) [31]. Most of our children received conventional heparin with a good outcome. Fresh frozen plasma infusion may be required to correct the antithrombin III levels. Fibrinolytic therapy was not given to any of our children. Tissue plasminogen activator has been efficacious in pediatric patients; however, its risk−benefit ratio in pediatric patients remains unclear. An attempt at surgical removal of peripheral arterial thrombosis in one child was not successful. There is no consensus on the duration of anticoagulation in children [31]. All our children received oral anticoagulation for at least 6 months. Children with associated protein S deficiency may require lifelong anticoagulation.

Our literature search showed only one report concerning prophylactic anticoagulation in children with NS [9]. Prophylactic anticoagulation involves a risk of bleeding and is therefore not routinely recommended. Moreover, it was found that anticoagulation therapy failed to prevent further recurrences of renal vein thrombosis in a short series of patients [8]. One child had recurrence of TECs in our cohort and was not receiving any anticoagulation therapy at that time. The use of prophylactic anticoagulation may be considered during relapse and infections with severe hypoalbuminemia in order to prevent thrombotic complications; however, further studies are required.

Limitations

A major limitation of our study is that the exact incidence of thrombosis in our cohort of children is unclear. It is extremely difficult to classify our pediatric population into SRNS/SDNS categories as the overwhelming majority of our children have been receiving treatment for several months–years from other centers. As ours is a tertiary care center, by the time the patients reach our institute they have received several courses of different immunosuppressive therapies. Strict categorization of these patients into FRNS/SRNS groups is rather inaccurate.

Another limitation of our study is that a complete coagulation profile like D-dimer, fibrinogen assay and tests for procoagulant states could not be performed in all the children due to technical and financial constraints.

We observed that some children developed thrombotic events even when the disease was in clinical remission. We do not have an adequate explanation on the pathogenesis of this event but it is an important clinical observation from our study.

Conclusions

Although rare, TECs are life-threatening complications in children with NS. Thrombotic complications are predominantly venous, but arterial thrombosis can also occur. They are more common in SRNS or SDNS but they can also be encountered during the first episode or a relapse in infrequently relapsing NS. Thrombotic risk factors, such as severe hypoalbuminemia, infections, arterial or venous punctures and volume compromise should be identified and promptly treated. Coexistence of genetic prothrombotic condition can occur, and merits evaluation. A high index of suspicion for thrombotic complications in a child with NS is required as the clinical features may be subtle. Neuroimaging and angiographic techniques help in confirming of diagnosis. Our experience shows that early aggressive heparin therapy followed by oral anticoagulants is required for a favorable outcome.