The Continued Rise of Venous Thromboembolism Across US Children’s Hospitals Free

This retrospective multicenter cohort study used the PHIS deidentified administrative database, which contains comprehensive data from not-for-profit tertiary care pediatric hospitals, representing about 25% of pediatric centers in the United States. Participating institutions represent diverse geographic locations, bed numbers, and average daily census and are affiliated with the Children’s Hospital Association (Overland Park, Kansas). The data warehouse is managed by Truven Health Analytics (Ann Arbor, Michigan), and assurance of data quality and reliability is a joint effort between the Children’s Hospital Association and participating hospitals. Because all data in the PHIS dataset are deidentified and collected for administrative purposes, this study did not qualify as human subject research and is considered exempt by the Nationwide Children’s Hospital Institutional Review Board.

Forty-nine hospitals contributed to the PHIS database between January 1, 2008, and December 31, 2019. The International Classification of Disease, Ninth Revision, and Tenth Revision (ICD-9-CM and ICD-10-CM) codes were used to identify all inpatient admissions with a diagnostic code for deep vein thrombosis or pulmonary embolism in their list of up to 41 ICD codes during this time frame (Supplemental Table 4).

We recorded age at admission, as well as sex, race, ethnicity, insurance type, and geographic region for all admissions. We classified age into the following categories: <1 month, 1 month to <1 year, 1 to 5 years, 6 to 12 years, and 13 to 17 years to compare VTE rates in neonates, infants, children, and adolescents. We used a previously published classification scheme that uses ICD-9-CM and ICD-10-CM codes to classify pediatric chronic conditions into neuromuscular, cardiovascular, respiratory, renal, gastrointestinal, hematologic and immunodeficiency, metabolic, other congenital or genetic defects, and malignancy.²  Pharmacy billing data determined anticoagulant use, and we examined the use of enoxaparin, warfarin, and direct oral anticoagulants.

Data were summarized using descriptive statistics, frequency and percentage for categorical variables, and median and interquartile range for continuous variables. The χ² test was used for dichotomous and categorical variables (age group, gender, race, ethnicity, disposition, and chronic condition) and the Mann-Whitney test was used to compare quantitative variables (lengths of stay). Logistic regression models were used to assess risk factors for VTE, such as the presence of CVC. The rate of VTE was calculated by dividing the number of VTE admissions by the total number of hospital admissions. Poisson regression was used to model the rate of VTE in each age group and the use of anticoagulants over time. SAS software, version 9.4 (SAS Institute, Cary, NC) was used for all statistical analyses.

During the 12-year study period, there were 52 401 hospital admissions among 39 713 pediatric patients with a diagnosis of VTE (Table 1). Among unique patients with VTE, 54.2% were male, the median age at admission was 6.1 years, and almost one-third (31.3%) were in the adolescent age group (13–17 years). The most common anatomic sites for VTE included the upper and lower extremities (35% and 28%, respectively) (Table 2). Pulmonary embolism accounted for 9% of VTE admissions.

The VTE admission rate increased from 46 VTE cases per 10 000 admissions in 2008 to 106 VTE cases per 10 000 admissions in 2019, a 130% increase (P < .0001) in VTE events. While the increase in VTE rate was statistically significant (P < .0001) for each age group studied, the largest increases were seen in adolescents and children 1 month to <1 year of age (Fig 1).

The age distribution of VTE admissions in our analysis was the typically reported bimodal age distribution, with VTE most commonly seen in the infant and adolescent populations (Fig 2). While most VTE events occurred in patients <12 months of age (32% of all VTE admissions), the relative risk of VTE was highest in adolescent admissions, particularly in patients 15 to 17 years of age. A VTE diagnosis was reported in these older adolescents in >1 out of every 100 inpatient admissions.

When comparing hospital admissions with and without a VTE diagnosis (Table 1), those with VTE tended to include older patients (median age 5.2 years compared to 3.8 years) with a longer length of stay (median of 30 days compared to 3 days). Admissions with a VTE diagnosis were less likely to originate from the emergency department and more likely to include a surgical procedure. VTE admissions were more likely to include time in the NICU and substantially more likely to include time in the pediatric ICU.

The percentage of patients with VTE with an underlying complex chronic condition (CCC) was 78%, with a median of 3 CCCs per patient (range: 1–11) (Table 3). Patients without VTE had a much lower proportion (27%) of CCCs. The most commonly associated CCCs were cardiovascular (38% of VTE patients) and gastrointestinal (28%), followed by neurologic/neuromuscular (18%), metabolic (17%), and renal or urologic conditions (15%). Also, approximately one-third of VTE patients were technology-dependent, defined by PHIS as patients with devices such as gastrostomy tubes, tracheostomy tubes, pacemakers, and cerebrospinal fluid shunts. There was a significantly higher proportion of patients with no CCC in patients 15 to 17 years of age compared to patients <15 years of age, 39% vs 17%, respectively.

Enoxaparin continued to be the most frequently prescribed anticoagulant in VTE admissions (Fig 3). The use of warfarin continued to decline, from use in 9% of VTE admissions in 2008 to 4% of VTE admissions in 2019. The use of direct oral anticoagulants (DOAC) was first recorded in 2014 (<0.5%) and had increased to use in 3% of VTE admissions by 2019. Due to the inability in PHIS to distinguish between unfractionated heparin dispensed to flush or maintain catheter patency, as opposed to the treatment of VTE, we were unable to determine the proportion of patients with VTE who received this agent.

The rate of VTE in hospitalized pediatric patients continues to increase from a 70% increase reported from 2001 to 2007 to the 130% increase from 2008 to 2019.¹  Overall, there has been a 211% increase in the VTE rate in hospitalized children since 2001. This is despite ongoing efforts to prevent VTE.³  The risk of VTE is spread throughout all age groups but consistent with prior findings it is the highest in infants and adolescents.

This study is not able to ascertain whether increased VTE rates represent a true increase in VTE occurrence or rather just improved detection, but it seems unlikely that an ongoing, almost 20-year trend is due to increased VTE awareness alone. Rather, this continued increase in VTE is more likely due to increasing medical complexity of patients and ongoing VTE risk factors with hospitalization. In the PHIS 2001–2007 VTE analysis, the majority (62%) of unique patients with at least 1 VTE admission had an underlying complex chronic condition (CCC).¹  The percentage of patients with a CCC was higher in our current analysis (78%; P < .0001), with a median of 3 CCCs per patient. In addition, up to one-third of patients were technology dependent. A study evaluating the Healthcare Cost and Utilization Project Database also identified an increase in the proportion of admissions for children with CCCs from 2010 to 2016.⁴  Unfortunately, rigorous epidemiologic studies to look at the rates of CVC placement or use of screening imaging studies over time have not been conducted. Notably, rates of childhood obesity have also continued to increase from 1999 to 2016, which may also have contributed to the increase.⁵ 

The national network of over 130 US pediatric hospitals, Solutions for Patient Safety⁶ , has identified hospital acquired VTE as a significant harm in hospitalized pediatric patients. Unfortunately, due to a lack of proven strategies to prevent CVC associated VTE, their efforts have been limited to the prevention of non-CVC associated thrombosis in children >12 years which represents only a small portion of VTE cases. In fact, CVC associated VTE accounts for 80% of all hospital acquired VTE while non-CVC VTE in children >12 years of age occurs in only 10% to 15%.⁷ 

Debate remains regarding the efficacy of anticoagulation to prevent CVC-VTE. Several studies using prophylactic dosing strategies for primary prophylaxis with CVC placement have not demonstrated efficacy.^8–10  The 2012 CHEST guidelines recommend against the use of systemic thromboprophylaxis in children with CVCs and the more recent ASH guidelines do not address this topic.^11,12  Recent data from a phase 2b clinical trial provides new evidence regarding efficacy of primary prophylactic enoxaparin in critically ill children >1 year of age when initiated within 24 hours of CVC placement, and these findings will be further investigated in a phase 3 study.^13,14 

This study confirms the well-known bimodal peak of VTE occurrence by age, with the highest rates of VTE found in the adolescent age group. This may represent children living longer with complex medical conditions combined with unique additional risk factors for thrombosis with age. Risk factors more common in adolescents can include onset of puberty, use of exogenous hormones, and history of prior VTE. Adolescents were less likely to have an underling CCC, suggesting that many of these VTE were associated with the risk factors described above, as well as acute events such as trauma and surgical challenges, and confirming what is generally observed in clinical practice.

Interestingly, there was a dip in VTE rates appreciated in the 2008 to 2009 timeframe, when the data switches from the original to the new analysis.¹  Specifically, the estimated 2007 rate was 58 VTE cases per 10 000 hospital admission versus the 2008 to 2009 rate, which was 46 VTE cases per 10 000 admissions. This finding could be due to a variety of factors, including: (1) change in the number of hospitals and data quality in PHIS over time, (2) patients admitted in 2007 and discharged in 2008 would not be included in our new analysis, (3) possible changes in health care utilization during the economic recession of 2008 to 2009 in the United States, and (4) potential differences in data abstraction, despite attempts to maintain consistency.

Trends in anticoagulant use demonstrates that enoxaparin continues to predominate in hospitalized pediatric patients. Interestingly, in 2019 there was a reduction in enoxaparin use, which corresponded with an increase in the use of DOAC s, as pediatric data on DOACs has become available. Our study likely underestimates the actual rate of DOAC use for pediatric/adolescent VTE as patients may have received enoxaparin inpatient and then transitioned to a DOAC as an outpatient.

This study has several limitations. Use of ICD codes for VTE diagnosis has not been carefully validated in PHIS, although other studies suggest that this is an acceptable method, despite potential inaccuracies.¹⁵  Approximately 18% of unique patients had >1 admission with a diagnosis of VTE. This was similar to the proportion reported by Raffini et al.¹  Although we are unable to verify that these were all recurrent events and it is possible that the previous VTE episode was included in the discharge codes for readmission, we included these additional VTE admissions in the analysis for several reasons. First, these additional VTE admissions were included in the 2009 study, and we wanted to replicate these methods as much as possible. Second, as shown in the 2009 study, the characteristics of these patients (proportion, older age, and association with malignancy) are similar to those reported in recurrent VTE in other cohorts. Lastly, even if some of the subsequent admissions were not recurrent events, inclusion of this diagnosis code likely reflects VTE as an active or ongoing problem, more accurately reflecting the disease burden of VTE in hospitalized children. Importantly, we were unable to differentiate between hospital acquired VTE versus admission for primary VTE. Collection of several relevant VTE risk factors, including family history, immobility, dehydration, recent estrogen use, and obesity was limited. Comparison of some baseline characteristics in admissions with VTE to without VTE (Table 1) may have been statistically significant but not clinically significant (eg race, payer type, or region) given the large sample size. As this is not a population based epidemiologic study, we were unable to estimate the rate of VTE in all children. Despite these limitations, we believe the results remain meaningful in highlighting current trends in pediatric thrombosis.

VTE in hospitalized patients continues to be an increasing harm despite current prevention efforts. Without targeted and effective interventions for CVC associated VTE, these complications will continue. Morbidity from VTE can be significant and includes pain, pulmonary embolism, paradoxical emboli and stroke, loss of venous access, infection, increased hospital cost and length of stay, long term post thrombotic syndrome and in some cases, death.^16,17  From 2001 to 2019 there has been a 211% increase in VTE in hospitalized pediatric patients. These findings support the need for clinical trials focusing on VTE risk stratification and discovery of more effective prevention strategies.

CCC

complex chronic conditions

CVC

central venous catheter

DOAC

direct oral anticoagulants

ICD-9-CM

International Classification of Disease, Ninth Revision

ICD-10-CM

International Classification of Disease, Tenth Revision

PHIS

pediatric health information system

VTE

venous thromboembolism