Variable Efficacy of Tranexamic Acid in Geriatric Hip Fractures Treated With Arthroplasty Versus Cephalomedullary Nails

Abstract

Introduction:

Geriatric hip fractures are associated with high rates of disability and mortality. Many of these patients require perioperative and postoperative allogeneic blood transfusions, which carry several noteworthy risks. A growing body of literature supports the efficacy of tranexamic acid (TXA) in geriatric hip fractures, without sufficient data examining which subgroups are likely to benefit the most.

Methods:

In this study, we sought to evaluate whether TXA was associated with reduced blood loss and transfusions in a geriatric population undergoing hip fracture fixation at our institution during a 2-year period. The first year's data were collected in a retrospective fashion before the introduction of a quality control initiative encouraging TXA administration for all geriatric hip fractures. The second year's data were collected prospectively. A subgroup analysis was conducted for patients who underwent arthroplasties.

Results:

Among the pooled cohort of patients undergoing surgery, TXA showed no benefit over control subjects for reducing blood loss or transfusion requirements. However, the subgroup of patients undergoing arthroplasty procedures showed a notable decrease in total blood loss and total units transfused during hospitalization.

Discussion:

These results suggest that TXA may be most beneficial when targeted to arthroplasties performed for geriatric hip fractures.

Hip fractures among geriatric patients are a leading cause of disability and mortality, with high rates of postoperative complications and prolonged hospital stays. Perioperative blood loss is a notable concern, and preoperative anemia is an independent risk factor of increased mortality and impaired functional status after surgery.^1-3 A growing body of evidence has shown tranexamic acid (TXA) to be effective in decreasing perioperative blood loss and transfusion requirements in hip and knee arthroplasty. Tranexamic acid is also known to reduce mortality in trauma patients who present with acute blood loss.⁴ There is markedly less research on the utilization and efficacy of TXA in the geriatric hip fracture patient population.

Tranexamic acid is an antifibrinolytic that blocks the lysine-binding site of plasminogen and is now routinely used in arthroplasty because of a reduction in postoperative transfusions.⁵ Blood products are a quantifiable expense in both their material cost per unit and the testing and processing of large quantities. The mean dollar amount paid per blood product unit in the United States was $207 in 2017.⁶ Additional hospital resources are dedicated to collecting, preparing, and administering transfusions.⁷ Furthermore, the need for transfusion is associated with increased lengths of stay and likelihood of intensive care unit admission.⁸

Allogeneic blood transfusions have direct risks for patients as well. Blood transfusions have been associated with a 35% greater risk of serious bacterial infection in general and 52% greater risk of pneumonia after hip fracture surgery.⁹ A recent meta-analysis found that allogeneic blood transfusions markedly increased the risk of postoperative surgical site infection among total hip and knee arthroplasty groups, from 1.74% to 2.88%.¹⁰ Other transfusion-related adverse events include hemolytic reactions, circulatory overload, transfusion-transmitted infection, and transfusion-related acute long injury.¹¹ Each year, fatalities occur in which transfusion is directly implicated.¹²

Existing data on TXA use in hip fracture surgery are confounded by variability in study design and the grouping of all hip fracture surgery types. Most studies support a general utility and application for hip fractures. One recent meta-analysis found that patients with hip fracture who received TXA had a 46% lower risk of requiring a postoperative transfusion, without a notable increase in thromboembolic events.¹³ Other studies have confirmed these findings.^14,15 Yet, patients with pertrochanteric and femoral neck fractures are often combined into one cohort, despite the inherent differences in pathology and treatment algorithms. Metrics for efficacy range from estimated surgical blood loss, corrected formulaic blood loss, hemoglobin drop, percent transfusion, or even number of units transfused. Yet, limited research has been devoted to differentiating whether all geriatric hip fractures should be treated with TXA or only select patients. In this study, we sought to answer the question of whether TXA use reduces total blood loss in all patients with geriatric hip fracture, or whether there were subgroups who might benefit most.

Methods

This study was conducted using our institution's hip fracture quality control registry and included patients presenting for orthopaedic evaluation at our level I trauma center over a 2-year period from November 2018 to November 2020. Inclusion criteria were patients aged 65 years or older who sustained acute intracapsular or extracapsular hip fractures treated with surgical fixation. Malignancy-associated pathologic fractures, periprosthetic fractures, and chronic fractures more than 2 weeks old were excluded.

Beginning in November of 2019, a quality improvement (QI) pilot project was approved by our institution's Pharmacy and Therapeutics Committee to improve preoperative screening of patients with hip fracture with the goal of ensuring timely perioperative TXA administration among patients who were thought to be appropriate candidates. The preliminary goal of this study was to increase our institution's utilization of TXA based on the existing data described above while allowing us to investigate subgroups of patients for whom we should prioritize TXA administration. Contraindications to TXA administration at our institution included history of blood clot, pulmonary embolism, stroke, transient ischemic attack, or allergy to the medication. The assessment of contraindications and ultimate decision on administering TXA were left to the anesthesiologist's discretion. Patients included in this study among the “no TXA” cohort included historical control subjects presenting in the 12 months preceding November 2019 who may have been candidates but did not receive TXA, as well as those patients presenting after November 2019 who either were not candidates for TXA or were candidates for TXA but still did not receive TXA because of incomplete adoption of TXA administration practices. When administered, TXA was dosed as a mixture of 1,000 mg TXA-NaCl in 100 mL normal saline, infused over 10 minutes. Infusion was begun before skin incision.

Hip fracture registry data were used to compare patients who received TXA with those who did not. These data included basic demographic information, pertinent medical history, preoperative blood counts, injury classification, type of anesthesia administered, and type of surgical treatment received. Postoperative outcomes were also assessed, including delirium, transfusion requirement, blood loss, length of stay, discharge disposition, and 30-day mortality. Blood loss was calculated as described according to well-established formulas by Gross and Nadler, which have been applied in much of the existing literature.^16,17 In brief, patient blood volume was calculated (male: 0.3669×[height]³ + 0.03219×[mass] + 0.6041, or female: 0.3561×[height]³ + 0.03308×[mass] + 0.1833) and used to determine total blood loss, comparing preoperative hematocrit with the lowest subsequent hematocrit observed during the admission: total blood loss = PBV×([preoperative hematocrit] − [lowest hematocrit]) ÷ (([preoperative hematocrit] + [lowest hematocrit]) ÷ 2).

Statistical analysis was conducted using R v4.0.0. Continuous data were compared using the Student t test, and categorical data were compared using the Fisher exact test. Significance was set at P < 0.05, with P < 0.10 considered a trend.

This study was under the purview of an existing protocol approved by our institutional review board. No external funding sources were used.

Results

A total of 218 patients in the hip fracture registry were included in this study, with a mean age of 80.48 years. Patient demographics and perioperative factors among the “TXA” and “no TXA” cohorts are listed in Table 1. No difference was observed in age or sex between patients receiving TXA and those who did not. However, patients receiving TXA were more likely to have presented initially with altered mental status (31.5% vs. 12.4%; P = 0.002). Conversely, patients receiving TXA also tended to have lower (better) American Society of Anesthesiologists physical status classification scores, although this was only a trend (P = 0.062). Remaining perioperative factors including patient functional status, fracture type, time from admission to surgery, preoperative blood levels, and type of anesthesia were not markedly different between groups.

Table 1.

Demographics (All Patients)

Demographics	No TXA		TXA		P
	n = 145	% or (SD)	n = 73	% or (SD)
Age	80.4	8.9	80.64	9.5	0.855
Female	94	64.8%	42	57.5%	0.304
Male	51	35.2%	31	42.5%
Pertinent medical history
Baseline dementia	37	25.5%	26	35.6%	0.154
History of stroke	30	20.7%	11	15.1%	0.362
History of VTE or PE	16	11%	4	5.5%	0.220
History of stent	27	18.6%	7	9.6%	0.113
AMS on presentation	18	12.4%	23	31.5%	0.002
ASA classification
ASA 1	0	0%	0	0%	0.062
ASA 2	19	13.1%	19	26%
ASA 3	107	73.8%	47	64.4%
ASA 4	19	13.1%	7	9.6%
Functional status
Level 1	84	57.9%	46	63%	0.249
Level 2	42	29%	14	19.2%
Level 3	19	13.1%	13	17.8%
Fracture type
Femoral neck	75	51.7%	43	58.9%	0.432
Intertrochanteric	67	46.2%	30	41.1%
Subtrochanteric	3	2.1%	0	0%
Perioperative care factors
Time from admission to surgery (hr)	18.91	27.67	16.3	10.8	0.320
Preoperative hemoglobin	11.69	1.85	12.13	1.9	0.107
Preoperative hematocrit	34.86	5.36	35.85	5.42	0.202
Surgical treatment
Cephalomedullary nail	63	43.4%	28	38.4%	0.718
Hemiarthroplasty	52	35.9%	31	42.5%
Closed reduction percutaneous pinning	18	12.4%	6	8.2%
Sliding hip screw	6	4.1%	4	5.5%
Total hip arthroplasty	6	4.1%	4	5.5%
Type of anesthesia administered
General	127	87.6%	65	89%	0.194
Spinal	12	8.3%	8	11%
Regional	6	4.1%	0	0%

ASA = American Society of Anesthesiologists, TXA = tranexamic acid

Postoperative outcomes were also compared between patients who received TXA and those who did not (Table 2). Among postoperative outcomes—including postoperative delirium, transfusion requirement, blood loss, length of stay, discharge disposition, and 30-day mortality—only discharge disposition was significantly different between groups (P = 0.002), with a greater proportion of TXA patients being discharged home compared with patients who did not receive TXA (19.2% vs. 9%).

Table 2.

Postoperative Outcomes (All Patients)

Outcomes	No TXA		TXA		P
	Mean or n	% or (SD)	Mean or n	% or (SD)
Delirium	37	25.5%	25	34.7%	0.201
Required transfusion	65	44.8%	29	39.7%	0.562
Units of blood transfused
Postoperative day 0	0.15	0.52	0.07	0.38	0.183
Postoperative day 1	0.28	0.65	0.27	0.67	0.984
Postoperative day 2	0.29	0.64	0.19	0.54	0.241
Postoperative day 3	0.29	0.64	0.19	0.54	0.944
Entire hospital stay	1.03	1.41	0.84	1.22	0.284
Total blood loss (by POD1 Hct)	1.09	0.91	1.18	0.88	0.483
Total blood loss (by POD3 Hct)	1.55	1.08	1.59	0.93	0.832
Total blood loss (by lowest Hct)	1.76	1.08	1.82	1.04	0.713
Length of stay (d)	7.01	6.32	6.21	5.17	0.314
Discharge disposition
SNF	108	74.5%	52	71.2%	0.002
Home	13	9%	14	19.2%
Rehabilitation	21	14.5%	2	2.7%
Expired	2	1.4%	2	2.7%
Hospice	0	0%	2	2.7%
Another hospital	0	0%	1	1.4%
Long-term acute care	1	0.7%	0	0%
30-day mortality	5	3.4%	4	5.5%	0.487

ASA = American Society of Anesthesiologists, TXA = tranexamic acid

To determine whether the type of surgical procedure affects the utility of TXA administration, a subgroup analysis was conducted among the 93 patients who underwent hemiarthroplasty or total hip arthroplasty (Table 3). In this combined arthroplasty subgroup, there were no differences in age (P = 0.912) or sex (P = 0.387) or in history of stroke (P = 0.631), VTE or PE (P = 0.525), or stent (P = 0.299) between patients who received TXA and those who did not. Patients in the arthroplasty subgroup who received TXA did tend to have better physical status with somewhat lower American Society of Anesthesiologists scores than those who did not receive TXA (P = 0.011). In the arthroplasty subgroup, there was no difference in preoperative hemoglobin (P = 0.502) or hematocrit (P = 0.551). However, patients in the arthroplasty subgroup who received TXA had fewer units of blood transfused on postoperative day 3 (P = 0.024) and throughout the entire hospital stay (P = 0.042) than those who did not receive TXA. A similar trend was observed for the odds of requiring transfusion at some point during hospitalization, with a trend toward lower odds of requiring transfusion among patients who received TXA (31.4% vs. 50%, P = 0.089).

Table 3.

Subgroup Analysis of Patients Who Underwent Hemiarthroplasty or Total Hip Arthroplasty After Hip Fracture

Surgical Treatment	No TXA		TXA		P
	n = 58	% or (SD)	n = 35	% or (SD)
Hemiarthroplasty	52	89.7%	31	88.6%	>0.999
Total hip arthroplasty	6	10.3%	4	11.4%
ASA classification
ASA 1	0	0%	0	0%	0.011
ASA 2	4	6.9%	9	25.7%
ASA 3	45	77.6%	25	71.4%
ASA 4	9	15.5%	1	2.9%
Perioperative care factors
Time from admission to surgery (hr)	25.64	41.45	18.24	13.2	0.212
Preoperative hemoglobin	11.62	1.79	11.88	1.85	0.502
Preoperative hematocrit	34.52	5.22	35.2	5.38	0.551
Postoperative outcomes
Delirium	20	34.5%	14	40%	0.659
Required transfusion	29	50%	11	31.4%	0.089
Units of blood transfused
Postoperative day 0	0.21	0.61	0.06	0.24	0.100
Postoperative day 1	0.17	0.5	0.23	0.6	0.643
Postoperative day 2	0.29	0.65	0.14	0.49	0.211
Postoperative day 3	0.29	0.65	0.14	0.49	0.024
Entire hospital stay	1.19	1.33	0.63	1.21	0.041
Total blood loss (L, by lowest Hct)	2.07	1.07	1.63	0.95	0.042
30-day mortality	4	6.9%	2	5.7%	>0.999

ASA = American Society of Anesthesiologists, TXA = tranexamic acid

Discussion

This study sought to investigate which subgroup of patients should be targeted during a QI protocol designed to increase utilization of TXA. Our results suggest that TXA may have increased utility when used for geriatric hip fractures that are treated with arthroplasties. In this cohort, TXA administration led to a reduction in blood loss and transfusion requirements. This difference fades when including all procedures commonly used to treat hip fractures (ie, arthroplasty, cephalomedullary nails, percutaneous fixation, and sliding hip screw constructs). Given that approximately one-third of patients expire within 1 year of hip fracture surgery, as well as the known association between transfusions and mortality, these findings have widespread applications.^2,18 These data are also consistent with previous studies showing a reduction in transfusions when TXA is used in arthroplasty procedures for hip fractures.¹⁹ Our findings suggest that previous positive findings for all pooled patients with hip fracture may have been influenced by the proportion of arthroplasties performed.¹³

These findings are consistent with much of the existing literature while simultaneously helping to clarify the circumstances under which TXA may be most indicated. One randomized controlled trial of TXA use in extracapsular hip fractures, all of whom were treated with short intramedullary nails, demonstrated a notable reduction in total blood loss but did not reach significance in reducing transfusion rates.²⁰ Pooled data on hip fractures in general have shown a reduction in the proportion of patients requiring blood transfusions, but with variability in sample size and study design, as well as inconsistent statistical significance despite a clear clinical trend.²¹ Although a recent large-scale prospective analysis validated the administration of TXA in reducing postoperative transfusion requirements without increased complications, that study did not conduct subgroup analysis on arthroplasty patients.²² Another randomized trial evaluating the same question corroborated the encouraging results of TXA in the combined hip fracture cohort for transfusion rates, but statistical significance dissipated in subgroup analysis.²³ It appears that variability in study designs and choice of metrics for blood loss versus transfusion requirements greatly affect which findings reach threshold for significance.

As a result, the widespread applications of TXA remain an area of ongoing exploration. In addition to the notable evidence supporting TXA in primary arthroplasty, more recent data demonstrate a notable reduction in blood loss and transfusion requirements for the treatment of periprosthetic hip fractures requiring revision surgery as well.²⁴ Future areas of exploration include optimal dosing regimens. There is some research that suggests multiple postoperative doses of TXA may further reduce blood loss.²⁵ Other data have shown that one dose of TXA is equally effective as two doses for decreasing blood transfusions after TKA and THA.²⁶ Some have investigated a possible synergistic effect of oral and intravenous TXA compared with intravenous alone, but data do not support a difference in reducing transfusions for primary arthroplasty.²⁷ There is also evidence that combining intravenous and topical TXA intraoperatively may lower total blood loss but does not markedly reduce transfusion requirements compared with intravenous TXA alone.²⁸

There are several weaknesses of our study. Owing to its observational nature, we could not determine whether TXA was withheld based on anesthesiologist's discretion or because of inconsistencies in hip fracture pathway orders. The lack of documentation regarding justification for withholding also led to difficulty tracking compliance with our QI program. Not all providers were aware of the updated protocol, given the large number of anesthesiologists and surgeons at our institution. We are under the assumption that inconsistent administration was more provider-dependent than patient-dependent. Also of note, no patients in our study developed symptomatic VTEs within 30 days of surgery. Possible explanations include missed diagnosis, care received elsewhere, lack of follow-up, and asymptomatic events. However, this result is consistent with baseline low rates of symptomatic DVTs in the 30 days after hip fracture surgery and that TXA use does not seem to affect this rate. Furthermore, we found that although the total units of blood transfused reached statistical significance in our arthroplasty subgroup, the binary need for transfusion did not. It is unclear whether this distinction is relevant for the more notable long-term outcomes such as mortality and complications.

In conclusion, the use of TXA in orthopaedic procedures has grown in popularity, in part because of increasing awareness of the role of perioperative anemia in postoperative outcomes. While TXA has been codified in the arthroplasty literature as a valuable tool in reducing the need for postoperative transfusions, existing data on geriatric hip fractures are limited by variability in study design and inconsistent metrics for blood loss. In this study, we demonstrate that TXA markedly reduces blood loss and postoperative transfusion requirements in patients undergoing arthroplasty procedures, but not other surgical fixation for geriatric hip fractures. Previous research supporting the benefits of TXA for all hip fractures may have been disproportionately affected by rates of arthroplasty procedures. Future investigation is needed to explore whether administration mode or timing may further improve TXA's efficacy. In particular, topical or sequential intravenous doses may strengthen the treatment effects. Furthermore, not all hospitals will have unlimited access to this medication. Our results both validate the continued use of TXA in patients with geriatric hip fractures undergoing arthroplasty and provide a subgroup for focus in targeted administration protocols.