Temporal changes in DVT risk after electrical injury

Abstract

Introduction

Previous work has utilized the National Burn Repository to examine deep venous thrombosis (DVT) after electrical injury. However, these studies were limited and could not examine when DVT occurs after electrical injury. In addition, the utility of risk assessment models for DVT risk stratification has not been examined in this patient population.

Methods

We performed a retrospective chart review of electrically injured patients at a single, ACS-verified burn center over a nine year period. Risk factors were identified and used to calculate Caprini scores at baseline and time of discharge. Outcomes of interest included symptomatic DVT or pulmonary embolus (PE) and time to DVT or PE.

Results

A total of 77 electrically injured patients were identified. DVT incidence was 6.5%. Patients with DVT had significantly higher TBSA (27.8% vs. 3.8%), mean number of operations (4.8 vs. 0.3), central venous catheter insertion (100% vs. 5.3%), ventilator days (16.2 vs. 0.3), ICU days (24.4 vs. 0.9), and mean change in Caprini score (18.6 vs. 1.3) during hospitalization. Baseline Caprini scores were low and DVT events occurred only after multiple risk factors were present; the average time-to-event was hospital day 17. Among patients with Caprini score >8, DVT incidence increased to 62%.

Conclusions

In our single center experience, the Caprini score was able to quantify DVT risk after electrical injury. In our series of 77 patients, the overall incidence of DVT was 6.5%. However, among patients whose Caprini score reached >8 during hospitalization, DVT incidence increased to 62%.

Introduction

Electrical injuries represent a small proportion of the patients referred to burn centers ^1-3. Severe electrical injury is commonly associated with concurrent injury to multiple organ systems; cardiac, neurologic, renal, and opthamalogic complications are common ⁴. Electrically injured patients have many recognized risk factors for DVT, including long ICU stays, central venous access, infections, concomitant burn injury, multiple operations, and prolonged immobility, among others ^5-12. Multiple case reports identifying a temporal relationship between electrical injury and venous thrombosis have been published ^13-19.

We have previously used the National Burn Repository (NBR) to examine DVT in electrically injured patients. Overall, DVT was a rare event among electrically injured patients who required hospital admission (0.9%). However, patients with more severe electrical injury, as indicated by the need for fasciotomy upon admission, were significantly more likely (7.6%) to have DVT events ⁸. We hypothesized that direct electrical injury to vascular structures may predispose patients to DVT, and that as a result DVT events would occur soon after injury. Unfortunately, the NBR dataset does not contain data on hospital day when DVT was diagnosed. While the NBR contains data on some recognized risk factors for DVT, many risk factors are either inconsistently documented (e.g. central venous catheter insertion) or not catalogued in the database (e.g. personal or family history of DVT). Thus, we were unable to rigorously examine our hypotheses using the NBR.

We present a retrospective chart review of patients with electrical injury admitted to an ACS-verified burn center. Our specific aims were 1) to examine time-to-event phenomena for DVT in electrically injured patients, and 2) to use the Caprini Risk Assessment Model to examine temporal changes in DVT risk after admission for electrical injury.

Methods

Our institution's Trauma-Burn Center maintains a prospective database of all patients with thermal or electrical injury. This database was used to identify all patients admitted for care after electrical injury between May of 2001 and April of 2010.

Retrospective chart review was performed using our institution's electronic medical record. Independent variables of interest included age, gender, ethnicity, mechanism of injury, body mass index (BMI), presence of medical comorbidities, total body surface area (TBSA) burned, presence of inhalation injury, hospital days, ventilator days, intensive care unit (ICU) days, injury type (flash, arc, or direct), high vs. low voltage injury, number of operative procedures, use of chemoprophylaxis, and need for fasciotomy or escharotomy.

The Caprini score ¹¹ has been validated as a predictor of venous thromboembolism (VTE) risk in medical inpatients ²⁰, general surgery patients ²¹, and plastic surgery patients ^{22, 23} (Figure 1). A Caprini score was calculated for each patient based on risk factors present at baseline and an aggregate score was maintained as risk factors were acquired during hospitalization. Of note, each inpatient operation meeting criteria for “major operation” added two points to the aggregate score. For patients with VTE events, a Caprini score at the time of VTE was documented.

Figure 1.

The Caprini Risk Assessment Model.

Dependent variables included symptomatic DVT or PE. DVT was diagnosed using lower or upper extremity duplex ultrasound. PE was diagnosed using pulmonary embolism protocol computed tomography. PE diagnosed at autopsy were considered to be symptomatic when the appropriate clinical scenario was present (ie PEA arrest leading to sudden death). Among those patients with DVT or PE, time to event was identified as an additional dependent variable.

Medical chart review was performed for 90 days after electrical injury or until time of last followup. Length of followup was tracked as an additional independent variable.

Data were analyzed using the Stata11 statistical package (StataCorp LP, College Station, Texas). Bivariate statistics were used to compare patients with and without VTE using the Fisher's exact test for dichotomous variables and the two tailed t-test for continuous variables. p ≤ .05 was considered significant. Stratified analysis was performed based on aggregate Caprini score at discharge. Multivariable regression was not performed due to a paucity of outcome events.

This study was approved by the our institution's Institutional Review Board.

Results

We identified 77 patients with electrical injury admitted for care at our institution between May 2001 and April 2010. Patient demographics are described in Table 1. No patient in this series had a personal or family history of VTE prior to admission.

Table 1.

Demographics of 77 electrically injured patients treated over a nine year period.

	Patients without DVT (n=72)	Patients with DVT (n=5)
Mean age	37	38
Male, %	92	100
Mean BMI	27.1	29.9
High voltage (>1000 volts) injury, %	44	80
Indirect (flash) injury, %	21	40
Mean baseline Caprini score	1.0	1.0

The overall rate of DVT was 6.5% (5 of 77 patients). Two patients with DVT also had pulmonary embolus. Thus, the overall rate of pulmonary embolus was 2.6% and of venous thromboembolism was 6.5%. All DVT and PE events occurred during the patient's initial inpatient stay. One patient with known DVT died after PEA arrest. PE was diagnosed at autopsy.

Risk factors for DVT are examined in Table 2. DVT after discharge have been shown to be common ^24-26. Thus, 15 patients with less than 20 days followup were excluded from this analysis. When compared to patients without DVT, patients with DVT had significantly higher TBSA, frequency of central venous catheter insertion, number of operations, number of ventilator days, and number of ICU days. Additionally, the mean change in Caprini score during hospitalization was 14-fold higher in patients who developed DVT (18.6 vs. 1.3, p<0.001).

Table 2.

Bivariate statistics examining risk factors for DVT among patients with at least 20 days of followup.

	Patients without DVT (n=57)	Patients with DVT (n=5)	p value
TBSA, %	3.8	27.8	0.001
Central line, %	5.3	100	<0.001
Mean number of operations	0.3	4.8	<0.001
Mean ventilator days	0.3	16.2	<0.001
Mean ICU days	0.9	24.4	<0.001
Mean change in Caprini score during hospitalization	1.3	18.6	<0.001

Five patients required surgical extremity decompression after electrical injury. Three patients who required operative decompression did not have venous thromboembolism. One patient had fasciotomy of the upper arm, forearm, and hand as well as upper arm escharotomy on hospital day one and a second patient required single digit escharotomy on hospital day zero. The third patient without DVT required hand and forearm fasciotomy on hospital day one and subsequently required an upper extremity amputation on hospital day 12. Of the two patients who required operative decompression and had DVT, one had fasciotomy of the forearm and hand in addition to upper arm escharotomy on hospital day two. This patient developed an iliofemoral DVT on hospital day 28. The second patient required forearm and hand fasciotomy on hospital day one and developed a calf vein DVT on hospital day 24.

The five identified DVT events occurred on hospital day 6, 12, 17, 24 and 28. No early or immediate DVT events were noted. Using baseline risk factors, all patients who developed DVT were risk stratified as “low” (Caprini scores of 0, 1, 1, 1, and 2). These patients gradually accumulated risk factors, as evidenced by increased Caprini score, during their hospitalization (Figure 2); 80% were classified as “super high-risk” ^{27, 28}, with a Caprini score >8, at the time of the DVT event. Three of the five received uninterrupted unfractionated or low-molecular weight heparin prophylaxis for the duration of their hospitalization. One patient received chemoprophylaxis for 50% of the days prior to DVT diagnosis. One patient received no chemoprophylaxis.

Figure 2.

Change in Caprini score over hospitalization.

Discussion

We have previously hypothesized that DVT occurs immediately or soon after electrical injury due to direct electrical injury to vascular structures. Our retrospective chart review does not support this hypothesis. Instead, our data supports that patients with severe electrical injury acquire risk factors during the course of their hospitalization, including prolonged immobility, central venous catheter insertion, major infections, and multiple operations (Figure 2); a “tipping point” appears to be reached at a Caprini score of eight. Among patients whose aggregate risk reached the “super high-risk” level of Caprini score >8 ^{27, 28}, 62% (5 of 8) developed a DVT. Although the Caprini score was initially designed as a pre-operative risk stratification tool ^{29, 30}, we have demonstrated its utility to track changes in VTE risk during hospitalization.

The Caprini RAM is a widely utilized risk assessment model (RAM). The RAM, in some cases with minor modifications, has recently been validated in diverse patient populations, including medical inpatients ²⁰, general, vascular, and urology surgery patients ²¹, and patients having post-bariatric body contouring ³¹ or plastic and reconstructive surgery ^{22, 23}. Previous publications ^{2, 8} have used the National Burn Repository to examine venous thromboembolism after thermal or electrical injury. With a sample size numbering thousands of patients, these studies were well suited to examine VTE incidence. However, the NBR does not contain several key variables for VTE risk stratification and thus these studies were unable to validate the Caprini or other existing RAMs.

Our data from electrically injured patients support previous publications which examine VTE risk factors after thermal injury. These publications identified increased TBSA ^{8-10, 32}, number of ICU days ^{8, 9}, number of ventilator days ⁸, total number of operations ⁸, and central venous catheter ⁶ as risk factors for VTE.

Our study has several limitations. We do not provide a discussion on the physics of electrical current or overall management of electrical injuries. These have been well reviewed elsewhere ^{4, 33-35}. Our study is prone to chronology bias ³⁶, where secular trends in medical care may change the standard of medical care. At our institution, we have routinely provided daily low-molecular weight heparin chemoprophylaxis to electrically injured patients since 2004; this could alter the natural history of DVT and change observed incidence. In our series, DVT events occurred in both patients who received no chemoprophylaxis and those who received uninterrupted daily chemoprophylaxis starting on hospital day one.

DVT is a rare event and our patient population is small. Thus, our observed DVT incidence of 6.5% may not represent the true incidence. Of note, a study of DVT after electrical injury using the NBR, which included over 1400 patients, demonstrated DVT incidence of 0.9% ². As only five DVT events were observed in our case series, we were unable to perform multivariable logistic regression. Thus, we cannot comment on independent risk factors for DVT after electrical injury.

Conclusion

Using baseline risk factors, DVT risk among our series of electrically injured patients was low. Patients with severe electrical injury quickly acquire multiple risk factors for DVT, such as prolonged immobility, central venous access, and multiple operations. The Caprini score is able to quantify DVT risk after electrical injury. In our series, the overall incidence of DVT was 6.5%. However, among patients whose Caprini score reached >8 during hospitalization, DVT incidence increased to 62%.