Post-thrombotic syndrome in children: a systematic review of frequency of occurrence, validity of outcome measures, and prognostic factors

Abstract

Background Post-thrombotic syndrome is a manifestation of chronic venous insufficiency following deep venous thrombosis. This systematic review was conducted to critically evaluate pediatric evidence on frequency of occurrence, validity of outcome measures, and prognostic indicators of post-thrombotic syndrome.Design and Methods A comprehensive literature search of original reports revealed 19 eligible studies, totaling 977 patients with upper/lower extremity deep venous thrombosis. Calculated weighted mean frequency of post-thrombotic syndrome was 26% (95% confidence interval: 23–28%) overall, and differed significantly by prospective/non-prospective analysis and use/non-use of a standardized outcome measure.Results Standardized post-thrombotic syndrome outcome measures included an adaptation of the Villalta scale, the Clinical-Etiologic-Anatomic-Pathologic classification, and the Manco-Johnson instrument. Data on validity were reported only for the Manco-Johnson instrument. No publications on post-thrombotic syndrome-related quality of life outcomes were identified. Candidate prognostic factors for post-thrombotic syndrome in prospective studies included use/non-use of thrombolysis and plasma levels of factor VIII activity and D-dimer.Conclusions Given that affected children must endure chronic sequelae for many decades, it is imperative that future collaborative pediatric prospective cohort studies and trials assess as key objectives and outcomes the incidence, severity, prognostic indicators, and health impact of post-thrombotic syndrome, using validated measures.

Introduction

Post-thrombotic syndrome (PTS) is a syndrome of chronic venous insufficiency following deep venous thrombosis (DVT), affecting both adults and children. Physical findings of PTS can include edema, pain, dilated superficial collateral veins, stasis dermatitis, and ulceration involving the affected limb. As edema and pain are often present acutely in patients with DVT, particularly in the setting of complete veno-occlusion, assessment beyond the acute and subacute periods following DVT onset is important for definitive diagnosis of PTS.

The pathophysiology of PTS is thought to ultimately involve venous hypertension.1 Venous hypertension, in turn, may result from venous valvular reflux, thrombotic veno-occlusion, and other causes of impaired venous return (e.g. right heart failure). While the correlation of PTS by clinical history or physical examination with valvular reflux by vascular imaging is rather poor, and while venous hypertension without reflux can be sufficient to cause PTS, PTS is nevertheless believed in many instances to be related to valvular reflux. Valvular reflux may be caused by valvular injury following acute thrombotic veno-occlusion (i.e. burst valve) or by vascular inflammation resulting in valvular fibrosis and insufficiency. Whether venous valvular insufficiency may in some cases be reversible remains poorly understood.

While considerable research effort has been devoted to the problem of PTS following DVT in adults in recent years, it has received less attention in pediatrics. However, since affected children will endure sequelae for many decades, there is a critical need to establish high-quality evidence on the burden of PTS in children with DVT. This systematic review was performed to critically evaluate evidence from the pediatric literature on the frequency of occurrence of PTS following DVT in children, the validity of pediatric PTS outcome measures, and prognostic indicators of PTS in children.

Design and Methods

We searched MEDLINE, EMBASE, and The Cochrane Library from 1960 (or earliest date for databases not extending back to 1960) through September 2009 (inclusive). MeSH terms and search strategy employed were as follows: “post-thrombotic syndrome OR post-phlebitic syndrome” AND “children OR pediatrics”. Reference lists of articles identified by the search were also reviewed for inclusion of additional relevant reports. Languages were limited to English, French, German, Italian, and Spanish. Study selection, categorization and data extraction were performed in duplicate by two independent reviewers (NG and MD). Publications were categorized as follows: single case reports; narrative reviews; commentaries; case series, cross-sectional studies; case-control studies; registries and cohort studies; and clinical trials. Single case reports, narrative reviews, commentaries, and conference abstracts were excluded from subsequent review. In case of reports on subsets of larger published series, only the most comprehensive series was included. In the event of conflicting study design classification or decision regarding inclusion/exclusion, consensus was achieved through discussion.

The following data were extracted from each eligible publication, as available: study methodology; age range; selection criteria; catheter-related DVT cases; cancer-associated DVT cases; VTE sites; frequency of PTS occurrence; method(s) employed for PTS outcome measurement; validation of PTS instrument; prognostic factors for PTS development; and quality of life (QOL) outcomes. Mean weighted frequency of PTS was calculated as previously described.2

Results

Overall search results

A flow chart of search results and their distribution by category of publication is provided in Figure 1, showing both included3–21 and excluded22–72 reports. Nineteen original reports met eligibility criteria (Table 1); no systematic reviews were identified. A total of 1,387 VTE patients were reported, including 1,084 with DVT affecting venous outflow from the upper or lower extremities (UE/LE DVT). The percentages of children who had LE DVT, UE DVT, and both were 55%, 44%, and 1%, respectively. Loss to follow up was 8% overall. In total, 997 patients were assessed for PTS.

Figure 1.Flow chart of search results and their distribution by category of publication, indicating both included and excluded reports.

In evaluating PTS, five studies8,13,14,18,20 used a pediatric modification8 of the Villalta scoring system for PTS in adults,73 two studies11,16 employed the Manco-Johnson instrument,51 and one study10 utilized the Clinical-Etiologic-Anatomic-Pathophysiologic (CEAP) classification system.74 Components of each of these outcome measures are summarized in Figure 2. The remaining 11 studies3–7,9,12,15,17,19,21 did not indicate the use of a standardized outcome measure.

Frequency of post-thrombotic syndrome occurrence

As shown in Table 1, frequency of PTS occurrence varied widely from 0 to 70% across studies. The earliest and largest data on PTS in pediatric DVT were provided by the Canadian Registry of Venous Thromboembolic Complications22 and the Canadian Childhood Thrombophilia Registry;5 however, these data preceded the standardization of PTS assessment in children. Findings with the use of a standardized PTS scoring system were first reported in 2003 in a retrospective analysis by the Childhood Thrombophilia Program at the Hospital for Sick Children8 in 153 children with a history of UE/LE DVT. This study employed a pediatric adaptation of the Villalta PTS scoring system,75 and defined a striking frequency of 63% for the occurrence of PTS in children with a history of limb DVT, at an average follow-up duration of 16 months post-event. By comparison, a prospective cohort including 52 unselected cases of acute UE/LE DVT found a PTS incidence of 33% at 1–2 years post-event, using the Manco-Johnson instrument.11

Table 1.Study characteristics of original reports on post-thrombotic syndrome (PTS) in children that met eligibility criteria for the present systematic review. In total 1,387 children with VTE were reported. Mean weighted frequency of PTS is calculated as the cumulative number of subjects with PTS divided by the cumulative number of subjects with DVT analyzed for PTS, and expressed as percent.

Calculated weighted mean frequency of PTS was 26% (95% confidence interval [CI]: 23–28%) overall, and differed significantly for prospective versus non-prospective analyses (17% [95% CI: 14–20%] vs. 43% [38–48%], respectively; P<0.001) and use/non-use of a standardized outcome measure (41% [36–46%) vs. 14% [11–17%]; P<0.001 (Table 2). There was no correlation between follow-up duration and PTS frequency across studies (r=− 0.015; P=0.97). Subgroup analysis for frequency of PTS in cancer-associated and catheter-associated DVT was not feasible, due to a paucity of studies separately reporting outcomes in these populations. In addition, PTS severity was graded only in studies employing the modified Villalta instrument,8,14,18,20 and functional significance of PTS was reported only in a single study, using the Manco-Johnson instrument.16

Figure 2.Components of standardized outcome measures for PTS employed in pediatric studies, including the modified Villalta scoring system, the Manco-Johnson instrument, and Basic CEAP. (A) Modified Villalta Scale.8,75 (B) Manco-Johnson instrument.16,51 (C) Basic Clinical-Etiologic-Anatomic-Pathophysiologic (CEAP) classification of chronic lower extremity venous disease,74 as used in van Ommen et al.10

Validity of outcome measures

Published validation data for pediatric PTS outcome assessment are limited to the Manco-Johnson instrument.16 This instrument (Figure 3) combines the basic CEAP classification system74 for physical assessment of signs of chronic venous insufficiency in combination with the Wong-Baker faces scale73 for evaluation of pain symptoms. Validity and reliability testing of the Manco-Johnson instrument employed a derivation cohort/validation cohort approach.16 In a derivation cohort consisting of 78 healthy children without prior DVT or recent leg injury, the upper limit of normal values for contralateral leg circumference difference was determined to be 1.0 cm at both the mid-thigh and mid-calf. No physical stigmata of PTS were found. Additionally, pain that limits aerobic exercise (e.g. sports, recreational activities, age-appropriate play), affects activities of daily living, or that occurs at rest was absent in this population. Subsequently, in an independent validation cohort consisting of 45 children with and without prior DVT, inter-rater reliability for the physical and functional components of the Manco-Johnson instrument (i.e. signs and pain symptoms of PTS) ranged between 91 and 100%. In spite of prior observations of some natural variation in contralateral arm circumference among healthy children,68 further validation of the Manco-Johnson instrument for the upper extremities has yielded similar results to those established in the lower limbs.76

Additional evidence for validity of a PTS outcome measure is provided by the demonstration of an adverse QOL among affected individuals. However, no pediatric studies to date have reported QOL in relationship to PTS findings, and no venous disease-specific pediatric QOL instrument has yet been developed and validated.

Table 2.Comparison of mean weighted post-thrombotic syndrome frequency by analytic design of studies and standardization of outcome assessment.

Prognostic factors for development of post-thrombotic syndrome

Over the past several years, knowledge has accumulated regarding prognostic factors for the development of PTS in adults and children. As presented in Table 3, prognostic factors from prospective studies include: patients’ characteristics; plasma/serum biomarkers relating to thrombophilia, coagulation activation, or the inflammatory response; and treatment factors. In pediatric thromboembolism, the risk of a composite outcome of recurrent VTE (rare), persistence of thrombosis despite a 3–6 month course of anticoagulation, or the development of PTS (using the Manco-Johnson instrument) is increased when FVIII activity and/or D-dimer concentration are elevated at diagnosis (OR 6.1, 95% CI: 2.1–17.7) or following 3–6 months of anticoagulation (OR 4.7, 95% CI: 1.8–12.6).10 Specifically, dual marker elevation at presentation increases the risk of adverse outcome from 50% (pre-test probability) to 86% (post-test probability).83 Thrombolysis may decrease PTS risk in pediatric patients with completely veno-occlusive proximal leg DVT in whom acute factor VIII and D-dimer levels both exceed prognostic thresholds (OR 0.018, 95% CI: <0.001–0.483);16 a priori PTS risk appears to be high in such children when conventional anticoagulation is employed.

Non-prospective studies have suggested additional potential prognostic factors for PTS in children. In the cross-sectional study of Kuhle et al.,6 multiple logistic regression analysis revealed that lack of thrombus resolution was associated with a statistically significant 4-fold increase in odds of PTS. With each additional venous segment of involvement by thrombus and each additional year of follow up, odds for PTS increased significantly by 2- and 1.2-fold. In addition, descriptive data from the registry report of Monagle et al.5 showed a skewed distribution of PTS across the pediatric age range, with infants and children of pre-school age more frequently affected than older children. Lastly, Sharathkumar and Pipe18 demonstrated in a retrospective series that initiation of anticoagulant treatment greater than 48 h following diagnosis of DVT and a history of recurrent thrombosis were each significantly more frequent among patients affected, versus unaffected, by PTS.

Figure 3.Manco-Johnson instrument.

Table 3.Prognostic factors for development of post-thrombotic syndrome following deep venous thrombosis in adults and children, from prospective studies.

Discussion

The present systematic review of the literature revealed that PTS is common following UE/LE DVT in children, with an overall weighted mean frequency of PTS occurrence of 26% (95% CI: 23–28%) among a total of nearly 1,000 UE/LE DVT patients studied. This frequency is in the range of those determined in adult studies, at 20–40%.84,85 The review also found considerable variability in the observed frequency of PTS across studies (ranging from less than 10% to approximately 70%), even when analysis of the literature is restricted to prospective analyses of unselected cases. Lastly, it identified a single pediatric PTS outcome measure (the Manco-Johnson instrument) for which validity has been shown.16 This instrument defined a cumulative incidence of PTS of 33% at 1–2 years following UE/LE DVT in children,11 which was considerably higher than the weighted mean frequency of PTS determined here among prospective analyses (17% [95% CI: 14–20%]), slightly lower than that for studies similarly employing a standardized outcome measure (41% [36–46%]), and quite consistent with the weighted mean frequency of PTS in the subset of studies that both employed prospective analysis and a standardized outcome measure (26% [20–33%]; cumulative n=174 subjects with UE/LE DVT) (data not shown).

The above findings are important in establishing key evidence on PTS in children. In 2008, the US Surgeon General’s Call-to-Action on Prevention of DVT and pulmonary embolism emphasized chronic venous insufficiency following DVT as an important priority for future investigation.86 Recently, the number of pediatric publications reporting on PTS has shown a marked increase, with over 50% of studies included here having been published in the past five years. In addition to prospective/non-prospective analysis and use/non-use of a standardized outcome measure, factors that could contribute to the variability in PTS frequencies reported across these studies include heterogeneity in disease severity among study populations, as well as differences in study population distributions with respect to putative modulators of PTS risk (e.g. patient age, body mass index, and activity level; lag time from symptom onset to antithrombotic therapy; DVT extent, degree of occlusion, and involvement of central vasculature; anticoagulant treatment intensity and duration; graduated compression stocking use and duration thereof). The fact that non-prospective analyses showed a significantly higher weighted mean frequency of PTS than prospective ones serves to emphasize the important potential for disease severity selection bias in retrospective studies. Conversely, our finding that PTS frequency was significantly higher in studies employing a standardized (vs. non-standardized) outcome measure suggests that the use of standardized outcome measures is critical for sensitivity in detecting PTS. While considerable differences existed in average follow-up duration across studies, no correlation was found between follow-up duration and PTS frequency.

Limitations of the present work include issues related to the original studies and to the systematic review itself. First, included studies were heterogeneous with respect to design, follow up, and PTS outcome measurement. Not all studies evaluated PTS as a primary objective, potentially limiting study quality. However, formal assessment of study quality was not feasible due to lack of sufficient detail provided by most reports. Nevertheless, categorization according to basic study design and characterization of the study population was undertaken during systematic review. Second, as noted earlier, time from DVT presentation to PTS assessment was not standardized across studies (nor typically within a study), leading to possible imprecision in composite measurement of PTS occurrence. Standardization of time to PTS assessment and serial assessment in long-term follow up would facilitate comparability of endpoints among future studies, and better define natural history. Thirdly, few studies investigated prognostic factors for PTS development. Consequently, only limited analysis could be performed in this review, and greater emphasis on investigation of prognostic factors for PTS must be underscored. Lastly, perhaps the greatest limitation in synthesizing evidence on PTS in children is the lack of a venous disease-specific pediatric QOL instrument by which to assess the functional significance of PTS. Until such a tool is developed and validated, the understanding of functional significance of PTS is limited to the use of the Manco-Johnson instrument.

In summary, this is the first systematic review of PTS in children. Among a total of 1,387 patients reported (including 997 with UE/LE DVT evaluated for PTS in follow up) across 19 eligible studies, the overall weighted mean frequency of PTS occurrence was 26%. However, validity and reliability in pediatric PTS measurement has been demonstrated only for the Manco-Johnson instrument, employed in two of the studies. Prognostic clinical factors and laboratory markers for development of PTS have recently begun to be prospectively investigated. Given that children must endure disease sequelae that adversely impact QOL for many decades, future collaborative prospective cohort studies and RCTs in the field of pediatric VTE should assess as key study objectives and outcomes the incidence, severity, prognostic indicators, and health impact of PTS, using validated measures.

Footnotes

• ↵* NAG and MPD contributed equally.
• Funding: NAG is funded in part by a Career Development Award from the National Institutes of Health, National Heart Lung and Blood Institute (1K23HL084055-01A1). SRK is supported by a National Research Scientist Award (chercheur national) of the Fonds de la recherche en santé du Québec. MC holds a Career Investigator Award from the Heart and Stroke Foundation of Ontario. UN-G is supported by a university research award (IZKF) and “Förderverein Schlaganfall und Trombosen e.V.” MJM-J is funded in part by a thrombosis/thrombophilia network award from the Centers for Disease Control and Prevention (U01DD00016).
• Authorship and Disclosures The information provided by the authors about contributions from persons listed as authors and in acknowledgments is available with the full text of this paper at www.haematologica.org.
• Financial and other disclosures provided by the authors using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are also available at www.haematologica.org.

• Received April 28, 2010.
• Revision received June 14, 2010.
• Accepted June 29, 2010.