Decreased Glycocalyx Shedding on Presentation in Hemorrhaging Geriatric Trauma Patients

Abstract

Background:

Plasma levels of syndecan-1 (Sdc-1), a biomarker of endothelial glycocalyx (EG) damage, correlate with worse outcomes in trauma patients. However, EG injury is not well-characterized in injured older adults (OA). The aims of this study were to characterize Sdc-1 shedding in OA trauma patients relative to younger adults (YA) and determine associations with putative regulators of EG sheddases.

Methods:

We performed a secondary analysis of data from the Pragmatic, Randomized Optimal Platelet and Plasma Ratios (PROPPR) trial, stratifying bluntly injured subjects into OA and YA groups based on upper age quartile (57 years). Plasma Sdc-1 levels were compared in OA and YA at hospital arrival through post-injury day 3, and the independent association between age and Sdc-1 level at arrival was determined after adjusting for differences in gender, shock index (SI) and pre-existing comorbidities. In a follow-up analysis, case-control matching was used to create populations of OA and YA with equivalent SI and injury severity scores (ISS). Levels of Sdc-1 were compared between these matched groups, and the relationships with candidate regulators of EG shedding were assessed.

Results:

Of 680 PROPPR subjects, 350 (51%) had blunt injuries, and 92 (26.3%) of these were OA. Plasma Sdc-1 levels at arrival, 2, and 6 hours were significantly lower in OA compared to YA (all p<0.05). After adjusting for sex, prexisting morbidities and SI, age was associated with decreased Sdc-1 levels at arrival. In the matched analyses, Sdc-1, HMGB1 and TIMP-2 levels were lower in OA compared to YA. Both HMGB1 and TIMP-2 significantly correlated with arrival Sdc-1 and were inversely associated with age.

Conclusion:

This study indicates that increased age is independently associated with decreased Sdc-1 levels among patients with blunt injuries. Suppressed plasma levels of sheddases in relation to diminished Sdc-1 shedding suggest that mechanisms regulating EG cleavage may be impaired in injured older adults. These findings provide novel insight into the age-dependent impact of injury on the vascular endothelium, which could have important implications for the clinical management of older adults following trauma.

Introduction

Injured older adults comprise nearly 25% of trauma admissions in the United States. As the population ages this proportion is predicted to rise to nearly 50% by the year 2060.(1) Despite many older adults experiencing a low-energy traumatic mechanism, the morbidity and mortality from blunt injury is significantly worse than their younger counterparts.(2) Injured older adults are at risk for an increased hospital length of stay (LOS), increased mortality, and a poor discharge disposition.(2) A subgroup of older adults that is much more likely to suffer from poor outcomes after trauma are those that are frail.(3)

Injury to the vascular and microvascular endothelial glycocalyx is an important contributor to endotheliopathy of trauma that drives multiorgan dysfunction and untoward outcomes.(4) However, endothelial glycocalyx (EG) injury and the ensuing endotheliopathy is not well-characterized in trauma patients with advanced age. In general, injured adults are noted to have increased plasma levels of glycocalyx components that are proportional with the severity of injury.(3) Blunt injuries, in particular, are associated with high levels of vascular endothelial activation and injury to the EG.(5) This damage is demonstrated by elevated plasma levels of syndecan-1 (Sdc-1), a heparan sulfate proteoglycan shed from the EG. Circulating levels of endothelial biomarkers such as Sdc-1 correlate with the severity of endotheliopathy and worse patient outcomes.(6) In addition, inflammatory mediators (e.g., tumor necrosis factor-alpha (TNFα), interleukin (IL)-1β, IL-6 and IL-8), danger-associated molecular patterns (e.g., high mobility group box 1 (HMGB1)), and protease inhibitors (e.g., tissue inhibitor of metalloproteinase (TIMP)-1, 2, 3 and 4) are associated with regulation of Sdc-1 shedding.(7, 8) Endotheliopathy and inflammatory regulators of EG shedding are not well described in injured older adults.

Better characterization of vascular endothelial injury and activation in older, severely injured trauma patients may advance our understanding of why certain patients do better than others. In non-trauma literature, an age-related decline in EG expression has been observed in older adults;(9) however, this pattern has not been studied in trauma patients. Thus, the aims of this study were to (a) characterize EG injury in older adult trauma patients relative to younger adults and (b) determine associations between advanced age and putative regulators of glycocalyx sheddases.

Methods

Study Population and Design

We performed a secondary analysis of deidentified data from the IRB-approved Pragmatic, Randomized Optimal Platelet and Plasma Ratios (PROPPR) trial dataset.(10) The original study enrolled injured adult patients requiring at least one unit of blood prior to arriving at the hospital or within the first hour of arrival and who were expected to require massive transfusion of blood products. Enrolled patients were randomized to receive either a 1:1:1 or 1:1:2 ratio of platelets, plasma, and packed red blood cells (pRBCs).(10) The PROPPR study was conducted with approval from the US Food and Drug Administration (NCT01545232), the Department of Defense, and in accordance with all local institutional review board committees for the Protection of Human Subjects under exception from informed consent.

In our secondary analysis, only subjects with blunt injuries were included given the high degree of anticipated injury to the vascular endothelium relative to subjects who suffered penetrating injuries. Subjects were stratified as younger adults (YA) or older adults (OA) based on the upper interquartile range for age. Additionally, YA and OA groups with equivalent injury severity score (ISS) and shock index (SI) were evaluated to control for the influences of injury and shock severity on Sdc-1 shedding after trauma.

Data Collection and Measurements

Subject demographic data and injury characteristics, including comorbidities, ISS, mechanism of injury, type of blunt injury, volumes of crystalloid, colloid, and blood product administered during resuscitation, and admission vital signs were collected prospectively during the PROPPR trial. Additionally, standard laboratory studies used to clinically evaluate coagulation and hemorrhagic shock were recorded for each patient. These values include activated partial thromboplastin time (aPTT), prothrombin time (PT), international normalized ratio (INR), base excess, and SI. Time to hemostasis and critical admission threshold (CAT) were included in the original dataset. Outcome data were also included in the original trial, and included number of ventilator-, ICU-, and hospital-free days as well as development of acute lung injury or acute respiratory distress syndrome (ALI/ARDS), acute kidney injury (AKI), multi-organ failure (MOF), thrombotic event and mortality at 24 hours, 72 hours, and 30 days. At designated time points in the first 72 hours of admission, citrated whole blood was collected from each subject. Samples were collected at 0, 2, 4, 6, 12, 24, 48, and 72 hours. Each sample of whole blood was centrifuged at 2,500 RCF for 20 minutes to obtain platelet-poor plasma, aliquoted, and stored at −80°C. Enzyme-linked immunosorbent assays (ELISA) were utilized to quantify levels of Sdc-1, HMGB1, TNFα, IL-1β, IL-6, IL-8, and TIMP-1, −2, −3, and −4, as described in prior PROPPR manuscripts.

Statistical Analysis

Descriptive statistics were conducted using frequencies and percentages for categorical data. Continuous data were summarized with means and standard deviations. To assess for differences in demographic variables between the two age groups, bivariate analyses consisted of the Likelihood Ratio chi-square for categorical variables and independent t-tests for continuous variables. Assumptions for both statistical tests were evaluated. The chi-squares’ expected frequencies for all categorical variables were met. For the independent t-tests, normality assumptions were tested with a normal quantile plot and the Anderson-Darling test. Continuous variables for both age groups indicated normality violations. Given that the sample size for both groups was greater than 30, the central limit theorem was invoked. Levene’s test was used to evaluate for equality of variance. For variables with equal variances, independent, pooled t-tests were used; t-tests assuming unequal variances were performed for variables without equal variance. Pearson’s rank correlation coefficient was used to evaluate relationships between plasma biomarker levels. An ordinary least squares (OLS) multivariate regression analysis was performed to determine independent associations between biomarker levels at the time of arrival to the hospital and injury and clinical characteristics. The OLS assumptions of normality, linearity, and homogeneity of variance were assessed with a scatterplot of the standardized residuals plotted against the standardized predicted values. In addition, the standardized residuals were plotted against each of the six predictor variables. Normality of the residuals was further assessed with a normal quantiles plot and the Anderson-Darling Test. Results from the normality of residuals assessments indicated that this assumption was violated; however, the scatterplot of residuals against the predicted values indicated this assumption was met. Therefore, we concluded that OLS assumptions were met. Case-control matching was performed to create OA and YA populations with equivalent SI and ISS at arrival, using a match tolerance of 0.1 for SI and 5 for ISS. Statistical analyses were conducted using SPSS Statistics, version 29 (IBM, Armonk, NY) and GraphPad Prism 9 (GraphPad Software, San Diego, CA). All analyses were two-sided with an alpha level of 0.05. P values <0.05 were considered significant.

Results

Unmatched Analysis

In the initial PROPRR study, 680 patients were enrolled. Patients with penetrating injury were excluded, leaving 350 bluntly injured patients for inclusion in our secondary analysis. Patients were classified as YA or OA based on a statistical delineation of the upper quartile for age (57 years) for the entire population of bluntly injured subjects. Notably, there were significantly more women in the OA group compared to the YA group. The older group also had significantly more comorbidities present at hospital arrival (Supplemental Table 1).

No differences were observed between YA and OA groups with respect to ISS, number of severe head injuries (determined by an abbreviated injury score for the head ≥3), time to hemostasis, and volume of crystalloids and blood products administered (Supplemental Table 1). Similar percentages of YA and OA sustained low energy blunt injuries (i.e., falls, assault, machinery) (Supplemental Table 2). Likewise, higher energy blunt injuries involving motor vehicle collisions (MVC) were also similar between YA and OA with the exception of MVCs involving pedestrians. The OA group contained more MVC pedestrian injuries compared to the YA group (30% vs 16%, respectively, p <0.01) (Supplemental Table 2).

Despite the overall similarities in injury characteristics, patients in the OA group had a significantly decreased SI at admission, driven principally by a lower mean pulse rate in the older group (Supplemental Table 1). Subjects in the older group also had significantly decreased platelet count and hemoglobin concentrations at admission compared to the younger group. Outcomes between the two populations significantly varied, with older patients having significantly fewer ventilator-, ICU-, and hospital-free days in addition to an overall increased 30-day mortality rate (Supplemental Table 1).

Levels of Sdc-1 for both age groups appeared to follow the same general trends over the 72 hours following admission. However, older patients had significantly decreased levels of plasma Sdc-1 at arrival, 2 hours, and 6 hours (Supplemental Figure 1).

OLS multivariable regression analysis was performed to evaluate the relationship between clinical factors that differed between YA and OA and arrival levels of Sdc-1. After controlling for gender and the presence of comorbidities, decreased age and increased SI were significantly associated with higher Sdc-1 levels in injured patients at hospital arrival (Supplemental Table 3).

Matched Analysis

To further determine the association of age with circulating biomarkers of EG injury, plasma Sdc-1 levels were compared in matched populations of injured OA and YA based on ISS and SI levels at admission. The resulting groups differed by age but had similar injury severity and were in similar states of hemorrhagic shock (Table 1). Like the unmatched populations, the OA group contained more women and had a significantly higher incidence of cardiovascular disease and diabetes mellitus.

Table 1.

Demographics, injury and admission characteristics, initial management, and outcomes of matched younger and older trauma patient populations according to injury severity score and shock index at arrival.

Variable	Matched Younger (< 57 years) (n = 82)	Matched Older (≥ 57 years) (n = 82)
Demographics
Age (years)	34±12	68±10***
Male, n (%)	58 (71)	49 (60)
Race
White	67 (82)	62 (76)
Black	10 (12)	7 (9)
Asian	2 (2)	9 (11)*
Other	5 (6)	13 (16)*
Unknown	2 (2)	2(2)
Hispanic/Latino Ethnicity	20 (24)	8 (10)*
Pre-existing Conditions & Comorbidities
CVD/Hypertension, n (%)	5 (6)	30 (37)***
Diabetes, n (%)	1 (1)	14 (17)***
Renal Disease, n (%)	0 (0)	1 (1)
Liver Disease, n (%)	4 (5)	25 (31)***
Cancer/Immunosuppression, n (%)	1 (1)	4 (5)
Injury Characteristics
Injury Severity Score	33±15	32±15
AIS Head ≥ 3, n (%)	27 (33)	36 (44)
Shock Index	1.0±0.4	1.0±0.4
CAT, n (%)	72 (88)	66 (81)
Time to Hemostasis (min)	216±147	218±211
Physiological Variables at Admission
Pulse (bpm)	110±30	97±23**
SBP (mmHg)	111±29	101±30*
Base Excess (mmol/L)	−8.1±5.9	−7.9±6.3
INR	1.5±0.5	1.5±0.6
aPTT (sec)	39±27	38±18
Hemoglobin (g/dL)	11.7±2.7	12.7±14.7**
Platelet Count (103/μL)	221±85	193±80*
White Blood Cell Count (103/μL)	14.6±7.7	13.9±6.0
Management
Pre-hospital Crystalloid (L)	0.7±0.7	0.7±0.9
Crystalloid, first 24h (L)	7.2±4.1	7.4±5.6
Units of RBC, first 24h	13±14	13±11
Units of Plasma, first 24h	9±10	9±9
Units of Platelets, first 24h	2±2	2±2
Units of Cryoprecipitate, first 24h	5±9	4±8
Anti-fibrinolytics, n (%)	17 (21)	23 (28)
Outcomes
Ventilator-free days	15±12	10±12*
ICU-free days	12±11	8±10*
Hospital-free days	6±8	4±8**
Thrombotic Event, n (%)	13 (16)	16 (20)
AKI, n (%)	18 (22)	23 (28)
ALI/ARDS, n (%)	22 (27)	23 (28)
Multiorgan Failure, n (%)	7 (9)	7 (9)
24-hour mortality, n (%)	15 (18)	17 (21)
72-hour mortality, n (%)	20 (24)	23 (28)
30-day mortality, n (%)	26 (32)	37 (45)

^*P <0.05;

^**P<0.01;

^***P<0.001

Continuous variables are reported as mean ± standard deviation, and categorical variables are reported number (percent).

AIS, abbreviated injury scale; AKI, acute kidney injury; ALI, acute lung injury; aPTT, activated partial thromboplastin time; ARDS, acute respiratory distress syndrome; CAT, critical administration threshold; ICU, intensive care unit; INR, international normalized ratio; MA, maximal amplitude; RBC, red blood cell; SBP, systolic blood pressure; TEG, thromboelastography.

The incidence of severe head injury was similar between age groups, and there were no significant differences in the types of low or high energy blunt injuries sustained by OA and YA (Table 2). However, the OA group was again noted to have a significantly lower pulse rate and systolic blood pressure at admission, as well as significantly lower platelet counts and hemoglobin concentration (Table 1). No significant differences in coagulation were detected based on thromboelastography measures at admission (Supplemental Table 4). There were no significant differences in fluid resuscitation or blood products administered between groups. Older patients had significantly fewer ventilator-, ICU-, and hospital-free days (Table 1).

Table 2.

Characterization of blunt injuries in younger and older adult trauma patients.

Type of Blunt Injury	Younger (< 57 years) (n = 82)	Older (≥ 57 years) (n = 82)
Fall	9 (11)	12 (15)
Assault	0	1 (1)
Machinery	0	2 (2)
Motorcycle	9 (11)	8 (10)
Bicycle	0	0
Motor Vehicle Collision
Bicycle vs vehicle	2 (2)	1 (1)
Motorcycle vs vehicle	9 (11)	4 (5)
Pedestrian vs vehicle	17 (21)	26 (32)
Occupant of vehicle	33 (40)	27 (33)
Unknown	1 (1)	0
Other	9 (11)	8 (10)

Similar to the unmatched analysis, OA had overall less Sdc-1 shedding at arrival and 6 hours after admission compared to ISS- and SI-matched YA (Figure 1).

Figure 1.

Time course of plasma syndecan-1 levels from matched younger and older adult trauma cohorts. Syndecan-1 levels in older trauma patients are significantly lower than younger trauma patients at admission and 6 hours following hospitalization, though the overall trend of syndecan-1 levels between age groups parallel each other. Lines on dot plot indicate means and standard deviation. *p<0.05 between age groups following independent t-test.

Multivariate linear regression performed using the matched populations confirmed that increasing age was significantly associated with decreasing levels of Sdc-1 after controlling for differences in sex or presence of comorbidity (Table 3).

Table 3.

Multivariable linear regression model evaluating factors associated with the syndecan-1 levels at hospital arrival.

	Adjusted R2 = 0.076
	B (95% CI)	P value*
Age	−3.3 (−5.3 to −1.3)	0.002
Female Gender	12.6 (−65.2 to 90.3)	0.750
Preexisting Condition/Comorbidity	48.8 (−50.8 to 148.5)	0.334

Regulators of Syndecan-1 Shedding

In matched YA and OA patients, only plasma levels of HMBG1 and TIMP-2 at hospital arrival were significantly different (Figure 2A,B) while no differences were observed in plasma levels of TNFα, IL-1β, IL-6, IL-8, TIMP-1, TIMP-3, or TIMP-4 (data not shown). Plasma levels of HMBG1 and TIMP-2 significantly correlated with plasma Sdc-1 levels at hospital arrival (Figure 2C,D).

Figure 2.

Plasma HMGB1 and TIMP-2 levels at arrival in matched younger and older adult trauma cohorts and their relationship with plasma Sdc-1 levels. (A) HMGB1 and (B) TIMP-2 levels are lower in older adult trauma patients compared to younger trauma patients with equal injury and shock severity. Lines on dot plot indicate means and standard deviation. *p<0.05 and ***p<0.0001 between age groups following independent t-test. Sdc-1 levels at arrival correlate with (C) HMGB1 and (B) TIMP-2 levels **p<0.001 based on Pearson’s correlation test.

Multivariate linear regression analysis of the matched YA and OA groups demonstrated that increasing age was independently associated with decreased HMGB1 levels (Table 4) and TIMP-2 levels (Table 5) at admission after controlling for sex and presence of comorbidities.

Table 4.

Multivariable linear regression model evaluating factors associated with the HMBG1 levels at hospital arrival.

	Adjusted R2 = 0.162
	B (95% CI)	P value*
Age	−0.69 (−1.0 to −0.35)	<0.001
Female Gender	−9.67 (−22.7 to 3.4)	0.145
Preexisting Condition/Comorbidity	−2.05 (−18.8 to 14.7)	0.809

Table 5.

Multivariable linear regression model evaluating factors associated with the TIMP2 levels at hospital arrival.

	Adjusted R2 = 0.081
	B (95% CI)	P value*
Age	−0.13 (−0.2 to −0.04)	0.008
Female Gender	−1.41 (−5.2 to 2.4)	0.466
Preexisting Condition/Comorbidity	−0.81 (−5.7 to 4.1)	0.744

Discussion

The findings of our study suggest that age is independently associated with decreased circulating levels of Sdc-1 among patients with blunt injuries. On admission and approximately 6 hours after hospitalization, Sdc-1 levels were observed to be significantly lower in older adults compared to younger adults. Putative regulators of EG shedding were also significantly associated with age and correlated with plasma Sdc-1 levels. Suppressed plasma levels of HMGB1 and TIMP-2 in the setting of diminished Sdc-1 shedding suggest that mechanisms regulating EG cleavage may also be impaired in older adult trauma patients, although further mechanistic studies are warranted to more definitively evaluate these associations.

These findings are novel given that literature reporting age-related differences in trauma-mediated EG shedding is currently sparse. There is evidence that there is an age-dependent variation in response to stress hormones and fibrinolysis in the setting of multitrauma. Johansson et al. (11) found that traumatically injured older adults, though presenting with higher plasma levels of noradrenaline, demonstrated lower plasma levels of adrenaline, as well as modulated platelet and leukocyte counts after injury.(11) Moore and colleagues (12) found that age was associated with a higher incidence of a fibrinolysis shutdown phenotype following polytrauma.(12) Furthermore, it is well established that trauma-induced innate and adaptive immune responses can differ considerably with advancing age.(13) Despite these recognized linkages between aging and putative injury response mechanisms, the impact of aging on vascular dysfunction after trauma remains relatively elusive.

In non-trauma literature, an age-associated decline in EG expression has been noted in OA with cardiovascular disease.(9,14) This thinner EG is found to change microvascular flow distribution during exercise and rest.(14) Results from these studies, in light of the evidence of a relatively blunted hormonal response and our finding of reduced EG shedding in response to trauma, may point to a relatively immunosuppressed state upon presentation compared to YA suffering similar traumatic injuries. While more work is needed to more robustly characterize the impact of advanced age on EG health and trauma-mediated endothelial activation/EG shedding, consideration of these factors is highly germane to the management of traumatically injured adults with advanced age.

Interestingly, we observed positive correlations between Sdc-1 and HMGB1 in addition to Sdc-1 and TIMP-2. Similar to Sdc-1, these proposed regulators of EG damage were inversely associated with age. Though further work is needed to confirm the causal relationship between these regulators and EG injury in the setting of trauma, non-trauma literature appears to support a mechanistic linkage. Yu et al.(15) demonstrated that inhibition of HMGB1 in the setting of lupus nephritis reduced EG shedding and maintained integrity of glomerular endothelial cells. The TIMP family of metalloprotease inhibitors have the ability to bind heparan sulfate and other molecules of the EG.(16) However, each molecule’s precise target(s) remains to be fully clarified. TIMP-3 has been shown to bind and cleave the glycocalyx ectodomain more so than TIMP-2.(16) However, TIMP-2 has also been demonstrated to regulate matrix metalloproteinase 2 (MMP-2) activity, which has a role in the structural integrity of the EG.(17,18) Much work remains to understand the roles of the individual sheddases and their contribution to glycocalyx shedding as it relates to age in the setting of trauma.

Understanding the injured older adult response to trauma is vital as the proportion of older trauma patients are increasing and have worse outcomes than their younger adult counterparts.(2) A key consideration in older adult trauma patients is whether this immunosuppressed state or aberrant response to injury is uniform across the entire older adult cohort or specific to the frail subset of geriatric trauma patients. Over the last decade our knowledge regarding poor outcomes in the frail subset of injured older adults has increased considerably with the work of Joseph et al.(19,20) It is now well established that frail older adults are subject to greater morbidity and mortality from traumatic injury and require more attention during triage, inpatient admission, with perioperative care, and post-discharge follow-up.(19–23) Whether a patient’s frailty status corresponds to an atypical endothelial response to trauma remains an important topic for further investigation.

The plasma biomarker differences between the OA and YA groups highlight the unique individual phenotypes that may exist following injury. Differentiation of phenotypic responses to injury is a relatively new area of investigation that has the potential to inform a more personalized (and theoretically more beneficial) treatment approach in the management of traumatized adults. Wu et al. (24) analyzed a multiomic dataset from the Prehospital Air Medical Plasma (PAMPer) trial and identified two distinct phenotypes of adults who suffered polytrauma that manifested disparate responses to plasma resuscitation. Recently, Thau et al. (25) also demonstrated two different trauma-induced endotypes in the PROPPR study population with differential treatment responses to 1:1:1 versus 1:1:2 component therapy. These studies provide scientific premise for the development of phenotypically directed resuscitation strategies, and has implications for the treatment of injured OA based on our study’s current findings that the vascular response to trauma differs between OA and YA.

There are important limitations to this study given its retrospective nature. As opposed to prospective studies, retrospective analyses offer inferior level of evidence compared with prospective studies and is subject to confounding as there may be unmeasured risk factors not captured by this database. For example, medications used to manage pre-existing morbidities in OA may have influenced the physiologic response to injury that, in turn, impacted mechanisms of EG shedding. From our study we can only determine an association and not causation based on the data gathered at the time of the study. Finally, the smaller sample size of older subjects limited the number of cofounding variables adjusted for in our subanalyses.

More work is needed to determine if the age-associated decline in glycocalyx shedding noted in this study is present to a similar degree in frail versus non-frail older adults as well as patients sustaining blunt versus penetrating mechanisms. In addition, whether the EG response to resuscitation in older adults is different from younger adults remains to be studied.

Conclusions

In our study, hemorrhaging, bluntly injured older trauma patients had decreased circulating levels of Sdc-1 and EG sheddases upon admission compared to a similarly injured younger population. Further studies are warranted to explore the age-dependent mechanisms driving decreased glycocalyx shedding in OA and to elucidate the contribution of decreased Sdc-1 shedding on geriatric trauma outcomes. Ultimately, a deeper mechanistic understanding of the impact of age on the vascular response to injury could help inform tailored therapeutic approaches to improve the clinical management of elderly trauma patients.