Prevalence of post-thrombotic syndrome following asymptomatic thrombosis in survivors of acute lymphoblastic leukemia


Lesley Mitchell, Stollery Children’s Hospital, Department of Pediatrics, Dentistry Pharmacy Centre, Room 1130, 11304-89 Avenue, Edmonton, AB T6G 2C7, Canada.
Tel.: +1 780 492 3137, fax: +1 780 492 3350.


Summary. Background: Deep vein thrombosis (DVT) is a complication of treatment of acute lymphoblastic leukemia (ALL) in children but little is known about the long-term outcomes of these DVT. Objective: To determine the incidence of post-thrombotic syndrome (PTS) in (i) children with ALL diagnosed with asymptomatic DVT using radiographic testing and (ii) an unselected group of ALL survivors. Methods: Cross-sectional study in two populations. Group I comprised children in the Prophylactic Antithrombin Replacement in Kids with ALL treated with L-Asparaginase (PARKAA) study diagnosed with DVT by radiographic tests. Group II consisted of non-selected childhood ALL survivors <21 years. PTS was assessed using a standardized scoring sheet. Results: Group I: 13 PARKAA patients (median age 12 years) were assessed, and 7 had PTS (54%; 95% CI, 25–81). All patients had collaterals, three also had increased arm circumference. Group II: 41 patients (median age 13 years) with a history of ALL were enrolled, and 10 had PTS (24%; 95% CI, 11–38). All patients had collaterals; five also had increased arm circumference. Conclusion: There is a high incidence of PTS in survivors of childhood ALL with radiographically diagnosed asymptomatic DVT. A significant proportion of ALL survivors develop PTS, indicating previously undiagnosed DVT.


Deep vein thrombosis (DVT) is a well-recognized secondary complication of treatment of acute lymphoblastic leukemia (ALL) in children [1–4]. The development of DVT in children with ALL is the result of both disease and treatment-related factors. One of the most important DVT risk factors specific for ALL is therapy with L-asparaginase, which causes an antithrombin deficiency as well as disturbances of almost all other coagulation proteins, resulting in a complex procoagulant state [3]. In the general pediatric population, indwelling central venous lines (CVL) are the most common factor associated with DVT in children [5]. Therefore, placement of CVL to facilitate clinical management further adds to the thrombosis risk in the ALL population, which is highlighted by the fact that a large number of DVTs occur in the vicinity of the CVL [3,4,6].

The reported incidence of DVT in children with ALL ranges between 1% and 73% [4,6–13] . The variation in the reported incidence can be explained by differences in diagnostic methods, study design, chemotherapy treatment protocols and definition of the DVT. Prospective studies including only clinically symptomatic DVT report incidences of between 3% and 14% [8,9,11,13]. Studies that used radiographic tests to screen for asymptomatic DVT report an incidence of between 35% and 73% [4,7,14,15] . However, the clinical significance of radiographically detected, asymptomatic DVT is unclear and controversial, as there are no data on the long-term outcome of asymptomatic DVT in children available. Even with this substantial incidence of asymptomatic DVT, the current standard of practice in this patient population is still to not use anticoagulant prophylaxis or screen for DVT [16].

Data on the outcome of symptomatic DVT in children show that about 60% go on to develop post-thrombotic syndrome (PTS) [17–19]. No study so far has determined the incidence of PTS following asymptomatic DVT in children, probably for two reasons. First, there is a lack of defined cohorts of pediatric patients screened for DVT, and secondly, there is a great deal of difficulty in following patients over many years.

The study ‘Prophylaxis with Antithrombin Replacement in Kids with ALL treated with L-Asparaginase’ (PARKAA) was a multicentre randomized controlled trial in which children with ALL were screened for thrombosis after treatment with L-asparaginase [4,20]. As survivors of childhood cancer, the PARKAA cohort continues to be followed in their respective centers. Therefore, the previous establishment of the PARKAA cohort (1997–1999) and the ability to locate these patients provided a unique opportunity to study the long-term outcome of asymptomatic DVT.


The overall objective of the current paper was to assess the incidence of PTS in children with ALL who previously had an asymptomatic DVT. There were two approaches to assess the question: firstly, to assess the outcome of asymptomatic DVT by determining the incidence of PTS in children with a history of ALL with radiographically diagnosed DVT; secondly, to corroborate the findings by determining the incidence of PTS in an unselected group of survivors of childhood ALL.

Patients and methods

Study design and population

The study design was a cross-sectional study performed separately in two populations. The first group (PARKAA group) comprised children who had been diagnosed with DVT in the upper venous system as part of their participation in the PARKAA multicentre study [4,20]. Diagnosis of DVT was made by screening with a combination of venography and ultrasound in the upper venous system and magnetic resonance imaging of the brain. PARKAA was an open, randomized controlled, extended phase II trial in pediatric patients with newly diagnosed ALL undergoing induction therapy. The trial was performed at eight pediatric tertiary centers in Canada (Ottawa, Edmonton, London, Calgary, Vancouver) and the US (Syracuse, Atlanta, Houston). Patients were randomized in a 1:2 ratio to either AT replacement therapy or no AT replacement. The primary and secondary objectives of the study were to determine the incidence of thrombosis in the non-treatment arm and to obtain preliminary data on safety and efficacy of AT prophylaxis in children with ALL, respectively. During the study period from July 1997 to May 1999, a total of 109 patients were enrolled (37 in the treatment arm, 72 in the non-treatment arm) [4,20]. A total of 30 children were diagnosed with a thrombotic event.

The second group consisted of non-selected children and young adults with a history of ALL who were followed in the Childhood Cancer Survivor (CCS) Program at Stollery Children’s Hospital, Edmonton, Canada (CCS group). The sample population was selected for two reasons, the first being that it is the hospital where the senior authors are located, therefore ensuring consistency in PTS assessment. Secondly, patients could be approached for consent for ultrasounds. Stollery Children’s Hospital is a large quaternary care facility, so we believe that the study population is representative of the general ALL population. To be eligible, subjects had to be less than 21 years of age and have had a central venous line in place during their treatment for ALL. The study was approved by the participating centers’ ethics review boards. Informed consent and assent (where applicable) was obtained from parents/guardians and/or patients.

Study methods

Patients were contacted and invited for a follow-up of their DVT at their respective treatment centers (PARKAA group) at five North American centers: Stollery Children’s Hospital, Edmonton, Canada; Children’s Hospital of Eastern Ontario, Ottawa, Canada; Upstate Medical University, Syracuse, USA; BC Children’s Hospital, Vancouver, Canada; and Texas Children’s Hospital, Houston, USA. For the second group (CCS group), the patients were assessed for PTS during their follow-up visit at the Childhood Cancer Survivor outpatient clinic at Stollery Children’s Hospital, Edmonton, Canada. Assessment for presence of PTS in the upper system was performed using a standardized scoring sheet. There has not been any formal assessment of the inter-assessor correlation of the PTS score. The investigators tried to minimize the potential for interpretation bias in two ways. For the PARKAA cohort only three investigators screened for PTS. All three are experienced and have worked together clinically. Two investigators (PM and LM) flew to the other centers in Houston, Syracuse and Vancouver to screen the patients. The third investigator (JH) screened her patients in Ottawa. For the Childhood Cancer Survivor (CCS), the children were screened by one of four attending oncologists and one of the investigators (SK) was always present to assess the patients as well and both would come to a consensus on the diagnosis. Demographic data were recorded for each patient, including age, gender, age at the time of diagnosis of ALL, ALL risk category, treatment protocol, presence and location of a previous DVT, type and duration of antithrombotic therapy (if applicable), location of the CVL, and other concurrent disorders in the past or present. In addition, if patients were diagnosed with PTS, patients were approached for consent to undergo ultrasound screening on the side that PTS was diagnosed. Ultrasounds were only performed in patients who consented to the procedure.

PTS scoring system

All patients were assessed for PTS in the upper body using a Pediatric PTS Score, which has been described in detail elsewhere [18]. In brief, the scale consisted of the following. (i) Symptoms reported by patient, parent, or proxy: pain or abnormal use, and swelling (not scored as increased limb circumference) scored as either absent (0) or present (1). (ii) Signs found by the examiner: limb circumference increased >3% compared with contralateral side, pitting edema, venous collaterals on skin as defined by obvious dilated unilateral superficial chest wall veins, pigmentation of skin, and tenderness on palpation scored on a dichotomous scale, and varicosities and head edema on a three-point Likert scale scored as absent (0), moderate (1), or severe (2) [21]. Any obvious dilation that was unilateral was considered significant. The presence of ulceration automatically resulted in a severe score. All signs and symptoms (except for ulceration) were equally weighted with a maximum score of 12. The classification of PTS was described as mild (score of 1 to 3), moderate (score of 4 to 8) or severe (score >8). Assessors of PTS were blinded to the location of the DVT.

Statistical analysis

Stata 9.2 (Stata Corp., College Station, TX, USA) was used to conduct the statistical analysis. The incidence of PTS in both groups is given as a point estimate with 95% confidence interval (95% CI). Demographic parameters were summarized by group and PTS status by mean (standard deviation), median (range) or relative frequency as applicable. The t-test for independent samples, Mann–Whitney U-test, and Fisher’s exact test were used to test for differences between patients with and without PTS. A P-value <0.05 was considered statistically significant. The likelihood ratio with 95% CI for development of PTS (LR+) with ≥50% obstruction on initial ultrasound in the PARKAA group was calculated using the standard formula [22].


Survivors of ALL with previously objectively diagnosed asymptomatic DVT (PARKAA group)

Thirty patients with objectively diagnosed DVT during the PARKAA study were eligible for inclusion into the follow-up study [4,20]. Of the 30 patients, one had a sinovenous thrombosis and was not screened for PTS, one centre did not participate (= 9), two patients had died, two refused to consent, and three patients were lost to follow-up, leaving 13 who had an assessment for PTS at their treatment centre. No difference in age, gender, ALL risk or AT treatment was seen in the patients from the non-participating centre when compared with the rest of the cohort (data not shown). In the cohort of 13 patients, there were four males (31%); median age at PTS assessment was 11.9 years (8.9–24.6), median age at diagnosis of ALL was 4.4 years (2.0–17.2). Patients had been followed since study entry for an average of 7.3 years (SD 0.6). While participating in the PARKAA study five patients were randomized to, and received, AT treatment, whereas the other eight were in the non-treated arm. There was no apparent difference between participation in the current study between the two arms of the PARKAA study.

Seven out of 13 patients (54%; 95% CI, 25–81%) showed signs compatible with the diagnosis of PTS. The mean PTS was 1.9 ± 1.2 (STD), with a range of 1–4. The patient with the score of 4 complained of pain in the affected arm when writing. There was no difference in incidence of PTS between patients treated with AT (60%) or not treated with AT (50%) while participating in the PARKAA study. Six patients had mild PTS; one had moderate PTS. All patients with PTS had collaterals on examination and increased limb circumference was seen in three out of seven. At the time of the DVT, all events were detected on screening with radiographic tests at study exit and no child was diagnosed with symptoms prior to study exit. In two patients, visible superficial collaterals were noted at the time of exit testing and in five the DVT were completely asymptomatic. Degree of occlusion was reported in 10 patients: >75% (= 5); 50–75% (= 1); 26–49% (= 2); <25% (= 2). Three of the 13 children had been treated with low molecular weight heparin for their DVT for 2, 4 and 6 months, respectively; two of these children developed PTS. The likelihood ratio for development of PTS (LR+) with ≥50% obstruction on initial radiographic testing was 3.3 (95% CI, 0.6; 18.9).

Follow-up ultrasounds performed at the time of PTS assessment in 6 out of 13 children showed evidence of a previous thrombosis with collateral formation in three children, all of whom had PTS. Ultrasound was normal in three children, one with PTS and two who did not have PTS. The child with PTS and normal ultrasound had a PTS score of 4 and reported aching in the arm when writing. Demographic and clinical characteristics by PTS status are shown in Table 1. There was no statistically significant difference between the patients with and without PTS.

Table 1.   Demographic and clinical characteristics of PARKAA patients with radiographically diagnosed DVT with and without PTS given as mean (SD) or relative frequency (n) as applicable
  PTS (= 7)No PTS (= 6)
  1. PTS, post-thrombotic syndrome; U/S, ultrasound; ALL, acute lymphoblastic leukemia.

Age at PTS assessment (years)12.9 (5.6)17.2 (5.7)
Male gender17% (= 1)50% (= 3)
Age at diagnosis (years)5.7 (5.4)9.8 (6.0)
Length of follow-up7.2 (0.6)7.5 (0.5)
High risk ALL29% (= 2)67% (= 4)

Non-Selected ALL survivors (CCS group)

During the study period from September 2005 to June 2006, a total of 43 eligible patients were seen in the Childhood Cancer Survivor outpatient clinic. Two patients refused consent/assent, yielding a total study population of 41 patients. Sixty-one per cent (= 25) were males. At the time of diagnosis of ALL, median age of the patients was 3.0 years (3 months to 12.6 years). Twenty-eight per cent of patients were high-risk precursor B ALL, 67% were standard risk precursor B ALL and 5% were T-cell ALL. All patients were treated according to a modified POG protocol. The mean age at PTS assessment was 12.8 years (SD 4.6), mean length of follow-up since diagnosis of ALL was 9.5 years (SD 4.0).

Post-thrombotic syndrome was diagnosed in 10 (24%; 95% CI, 11–38) patients. The mean PTS was 1.8 ± 1.2 (STD), with a range of 1–4. Nine patients had mild and one had moderate PTS. Collaterals (= 10; 100%) and increased arm circumference (= 5; 50%) were the most frequent symptoms. Eight patients had a CVL placed on the side of their PTS and thrombosis had been previously diagnosed on the side of the PTS in the other two patients. Five patients with PTS had follow-up ultrasounds performed: three patients had evidence of a previous thrombosis, one ultrasound showed no evidence of thrombosis, and one could not be interpreted.

Demographic and clinical characteristics by PTS status are shown in Table 2. There was no statistically significant difference between the patients with and without PTS.

Table 2.   Demographic and clinical characteristics of unselected survivors of childhood ALL with and without post-thrombotic syndrome given as mean (SD) or relative frequency (n) as applicable
  PTS (= 10)No PTS (= 32)
  1. PTS, post-thrombotic syndrome; ALL, acute lymphoblastic leukemia.

Age at PTS assessment (years)11.4 (6.1)13.2 (4.0)
Male gender60% (n = 6)61% (n = 19)
Age at diagnosis (years)3.4 (4.4)3.4 (3.0)
Length of follow-up (years)7.9 (4.1)10.2 (4.0)
High-risk ALL40% (n = 4)24% (n = 7)


The PARKAA study established a novel cohort of pediatric patients in whom asymptomatic DVT was detected on screening with radiographic tests. The PARKAA investigators continue to follow these patients as survivors of ALL. The cohort provided a unique opportunity to follow-up asymptomatic DVT in children long term. Therefore, the current study is the first to do a long-term follow-up for PTS development in children who were screened for asymptomatic DVT using radiographic tests. The results show that there is a high incidence of PTS of 54%. The incidence of PTS in asymptomatic DVT is comparable with that found for symptomatic childhood DVT (63%) [18]. These results were corroborated in a group of non-selected survivors of childhood ALL showing PTS is relatively common (24%).

Recent studies have found PTS in 33 to 70% of children suffering from symptomatic DVT [18,23,24]. However, a significant proportion of DVTs in children are asymptomatic. Compared with the acute presentation of DVT in adults, DVTs in children are frequently gradual in development, permitting an extensive collateral system to form, thereby reducing acute symptoms of DVT such as limb or neck swelling. Three studies that used radiographic screening for DVT in children found considerably higher incidences of DVT than based on clinical suspicion only [4,7,25]. The clinical significance and long-term outcome of these asymptomatic, radiographically identified DVTs has not yet been studied. However, there is no biological rationale to assume that acute (such as pulmonary embolism or CVL-related sepsis) and long-term complications (such as PTS or loss of venous access) of DVT are substantially less frequent with asymptomatic as compared with symptomatic DVT [26–29]. Wille-Jorgensen et al., in a meta-analysis of seven studies of adult post-surgical patients, showed that the relative risk for developing PTS in patients with asymptomatic DVT was 1.6 (95% CI, 1.24–2.02) as compared with patients without objectively verified DVT [30]. A recent study investigating the long-term outcome of radiographically diagnosed, asymptomatic DVT in adults following hip surgery found PTS in 24% of the patients after 18 months of follow-up, which is comparable with the incidence of PTS following a first symptomatic DVT [31,32]. As shown by the results from the current study, the incidence of PTS in children with asymptomatic DVT is similar to that of symptomatic DVT. These findings are not surprising as the DVTs in the PARKAA patients were associated with a degree of obstruction of 50% or more in the majority of patients, indicating compromise of venous blood flow despite the absence of symptoms. Patients in the PARKAA study with PTS were 3.3 times more likely to have an obstruction of ≥50% of the vessel diameter on initial diagnosis (LR+ 3.3) when compared with patients without PTS. However, the small patient numbers limit the interpretation of this result.

Data from the current study have shown that about 25% of non-selected survivors of childhood ALL show clinical signs of PTS, thereby confirming a significant incidence of undiagnosed asymptomatic DVT in this population. In a similar study, Journeycake et al. found mild PTS in 5/50 cancer survivors (10%; 95% CI, 3–22%) [33]. Van Ommen et al. reported an incidence of PTS of 50% in the lower system in children with a history of cardiac catheterization [19]. A comparison of the results of these studies is hampered by the fact that the risk of DVT (and hence that of PTS) differs considerably for a number of reasons (e.g. number and location of CVLs, duration of CVL placement, underlying disorders, treatment of DVT). Still, all incidences reported in the literature, albeit different, can be considered high. On the other hand, a recent study by Ruud et al. reported an incidence of clinically mild PTS of only 6% in 71 non-selected childhood cancer survivors using the same standardized scoring system [18,34]. The authors explained the low incidence by the fact that all patients in the study had their CVL placed in the jugular vein while our PTS scoring system may be more appropriate for use in upper and lower limb PTS.

The majority of children in the current study as well as those in previously published papers exhibited only mild PTS [18,19,33,34]. Frequently, presence of collateral veins was the only symptom. While mild PTS has no immediate consequences for the affected children, the findings are still important for a number of reasons. Firstly, presence of PTS may indicate compromise of venous circulation and loss of future venous access, a significant complication in children with underlying chronic disease requiring repeated venous access. Secondly, the incidence of PTS can be considered a minimum estimate of DVT incidence in pediatric patients with ALL. Thirdly, a previous silent thrombosis may put these children at an increased risk for a new DVT and they may benefit from anticoagulant prophylaxis in future high-risk situations. Finally, PTS in adults results in substantial burden to the healthcare system and reduction in quality of life [35,36].

The study has some limitations that need to be addressed. First, the numbers of patients enrolled were relatively small in both groups. However, the lower limits of both 95% confidence intervals indicate a high incidence. In addition, only 13/30 patients with asymptomatic thrombosis from the original PARKAA cohort were recruited, raising concerns about a bias in the patient sample. For the 17 patients that did not participate, one had a sinovenous thrombosis and one patient died, leaving 15 eligible patients. The primary reason for the other patients not being enrolled is that one centre did not participate, accounting for 9/15 missing patients. Therefore, the authors do not anticipate a bias in the assessed population. Secondly, there were only a few ultrasounds available in the CCS group confirming presence of a previous thrombosis. However, patients either had a CVL placed or a previously diagnosed DVT on the side of their PTS symptoms. Also, other differential diagnoses of venous insufficiency such as primary varicose veins or congenital venous abnormalities are rather uncommon. Therefore, PTS is the most likely explanation for swelling and collaterals found in this population. However, for clinical purposes, routine screening of children without a previously documented thrombosis is not recommended. Finally, most patients without PTS were not screened with ultrasound. An obvious control for the PARKAA cohort would be the children that were not diagnosed with thrombosis. There is one main challenge to that approach in that after participating in the PARKAA study, these children will have gone on to several years of chemotherapy as well as multiple central lines and other interventions. Therefore, the child may have had a DVT after being in the PARKAA study that would be detected on ultrasound, confounding the interpretation.

In summary, the current study found a high incidence of PTS in children with asymptomatic DVT, which is similar to that of symptomatic DVT. Further, a significant proportion of non-selected survivors of childhood ALL show signs and symptoms compatible with the diagnosis of PTS, indicating previously undiagnosed DVT in this population. Routine screening for DVT should be considered in children with ALL undergoing chemotherapy.


The authors would like to acknowledge and thank the staff of the Childhood Cancer Survivor Clinic at Stollery Children’s Hospital, Edmonton, for their invaluable support with the study. The study was funded by the Canadian Institutes of Health Research (Grant #77742) and the May McLeod Foundation. S. Kuhle was the recipient of the 2005/2006 Specialty/Subspecialty Fellowship Award at the Department of Pediatrics, Stollery Children’s Hospital, Edmonton.

Author contributions

S. Kuhle performed research, analyzed and interpreted statistical analysis, drafted the manuscript; M. Spavor, P. Massicotte, J. Halton, I. Cherrick, D. Dix, D. Mahoney, M. Bauman and S. Desai performed research; L. G. Mitchell designed research, performed research, analyzed and interpreted statistical analysis and drafted the manuscript.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.