The Journal of Arthroplasty 40 (2025) 119e126

Contents lists available at ScienceDirect

The Journal of Arthroplasty

journal homepage: www.arthroplastyjournal.org

Primary Hip

Blood Transfusion in the Age of Tranexamic Acid: Who Needs a

Type and Screen Before Total Hip Arthroplasty?
Muhammad A. Haider, Spencer A. Ward, MD, Vinaya Rajahraman, MD,
Joshua C. Rozell, MD, William Macaulay, MD, Ran Schwarzkopf, MD,
Matthew Hepinstall, MD *
Department of Orthopedic Surgery, NYU Langone Health, New York, New York

a r t i c l e i n f o a b s t r a c t

Article history: Background: Modern surgical protocols, particularly the use of tranexamic acid (TXA), have reduced, but
Received 23 January 2024 not eliminated, blood transfusions surrounding total hip arthroplasty (THA). Identifying patients at risk
Received in revised form for transfusion remains important for risk reduction and to determine type and screen testing.
17 June 2024
Methods: We reviewed 6,405 patients who underwent primary, unilateral THA between January 2014
Accepted 19 June 2024
Available online 22 June 2024
and January 2023 at a single academic institution, received TXA, and had preoperative hemoglobin (Hgb)
values. We compared demographics, baseline Hgb levels, and surgical details between patients who were
and were not transfused. Data were analyzed utilizing multivariate regression and receiver operating
Keywords:
total hip arthroplasty
characteristic curve analysis.
transfusion Results: The overall perioperative and intraoperative transfusion rates were 3.4 and 1.0%, respectively.
type and screen Patients who were older, women, and American Society of Anesthesiologists class >II demonstrated an
tranexamic acid increased risk of transfusion. Risk of transfusion demonstrated an inverse correlation with preoperative
hemoglobin Hgb levels, a bimodal association with body mass index, and a direct correlation with age, surgical time,
perioperative management and estimated blood loss on multivariate analysis. The receiver operating characteristic analysis
demonstrated a preoperative Hgb cutoff of 12 g/dL for predicting any transfusion. Above the threshold of
12 g/dL, total and intraoperative transfusions were rare, with rates of 1.7 and 0.3%, respectively. Total and
intraoperative transfusion rates with Hgb between 11 and 12 g/dL were 14.3 and 4.6%, respectively.
Below 11 g/dL, total and intraoperative transfusion rates were 27.5 and 10.1%, respectively.
Conclusions: In the age of TXA, blood transfusion is rare in THA when preoperative Hgb is >12 g/dL,
challenging the need for universal type and screening. Conversely, patients who have Hgb < 11.0 g/dL,
remain at substantial risk for transfusion. Between Hgb 11 and 12 g/dL, patient age, sex, body mass index,
American Society of Anesthesiologists classification, anticipated estimated blood loss, and surgical time
may help predict transfusion risk and the need for a perioperative type and screen.
Level of Evidence: III.
© 2024 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and
similar technologies.

Total hip arthroplasty (THA) is one of the most successful function in end-stage osteoarthritis of the hip [1]. The demand for
orthopaedic surgeries, routinely relieving pain and restoring THA is expected to explode, with certain estimates predicting a
469% increase in volume by the year 2060 [1e3]. Despite the
success of THA, it does not come without risks. Perioperative
bleeding can be extensive, resulting in the risk of blood transfusion
One or more of the authors of this paper have disclosed potential or pertinent [4,5]. Studies before routine tranexamic acid (TXA) use have
conflicts of interest, which may include receipt of payment, either direct or indirect,
reported a mean drop in hemoglobin (Hgb) of 4 grams per deciliter
institutional support, or association with an entity in the biomedical field which
may be perceived to have potential conflict of interest with this work. For full following THA [6]. Contemporary studies, with preoperative
disclosure statements refer to https://doi.org/10.1016/j.arth.2024.06.053. administration of TXA, estimate an average blood loss of 1,188 to
* Address correspondence to: Matthew Hepinstall, MD, Department of Ortho- 1,651 mL perioperatively, representing up to 20% of a patient’s total
pedic Surgery, NYU Langone Health, NYU Langone Orthopedic Hospital, 301 E 17th
blood volume [7,8]. Blood loss is the most frequent adverse event
St, New York, NY 10003.

https://doi.org/10.1016/j.arth.2024.06.053
0883-5403/© 2024 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
120 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126

during surgery, and extensive intraoperative blood loss is an Data Collection

established risk factor for higher rates of mortality, postoperative
complications, and increased usage of health care resources [9,10]. Demographic variables were collected using the electronic
Furthermore, when patients who have substantial blood loss un- medical record (EMR, Epic Systems Corporation, Verona, Wiscon-
dergo transfusions, they incur additional transfusion-associated sin) and included age, body mass index (BMI), sex, smoking status,
risks, such as infection, acute hemolytic reactions, anaphylactic race, and the American Society of Anesthesiologists (ASA) classifi-
reactions, and in extreme cases, death [11]. Transfusions can also cation. In addition, perioperative data were collected regarding
delay patient mobilization and increase hospital length of stay operative time, surgical approach, length of hospital stay, estimated
(LOS), with consequent increases in health care costs to both the blood loss (EBL), and whether a patient received either an intra-
patient and hospital system [12,13]. Given the risks and costs operative or postoperative transfusion. Estimated blood loss (EBL)
associated with perioperative blood loss and transfusions, it is was collected from anesthesia notes from the institutional
imperative for both surgical teams and hospital systems to employ electronic medical record. Drains are not routinely used at our
efforts to preemptively identify and optimize susceptible patients. institution following primary elective THA and have generally not
Historically, blood transfusions occurred at much higher rates, been used over the study period. However, our method of data
resulting in routine type and screen testing [14]. However, due to capture did not allow us to capture any rare drain use.
advances in surgical technique, the implementation of TXA, and
changes in transfusion protocol, perioperative transfusion rates Data Analyses
have dropped precipitously over the past 20 years [15]. Now that
transfusions are relatively rare, universal type and screen testing Data were analyzed using Statistical Packages for the Social
surrounding routine arthroplasty procedures has been questioned Sciences (SPSS) version 25.0 (International Business Machines
[14]. By performing informed type and screens on a case-by-case Corporation [IBM], Armonk, NY). Continuous variables were tested
basis, providers can minimize costs and inconvenience to patients for normality. Chi-square analyses were utilized for nominal cate-
while ensuring necessary screening is performed in a timely gorical variables, and independent t-tests were used for continuous
manner. variables, with significance set at P value of less than .05. Mean
Furthermore, many institutions have re-evaluated blood values were reported with ranges for continuous variables, and
transfusion triggers and considered or adopted updated blood categorical variable counts were reported with percentages within
management protocols, designed to minimize unnecessary trans- each respective cohort. Multiple independent binary logistic
fusions and costs, while maximizing patient safety and outcomes regression analyses were utilized to identify risk factors for all
[16e18]. Optimizing blood management protocols, including transfusions and intraoperative transfusions, respectively.
reductions in unnecessary type and screen testing and avoidance of Multiple receiver operating characteristic (ROC) curve analyses
transfusions, have been estimated to save approximately $200 per were then completed utilizing risk variables of interest in
patient [14]. Given that annual THA and total knee arthroplasty conjunction with Youden’s J statistic to identify an optimal cutoff
volume is projected to reach 2 million cases by 2030 [3], these preoperative Hgb value at which a type and screen would be most
optimized protocols can result in hundreds of millions of dollars in beneficial [19]. Optimal cutoff values were identified for both all
cost savings each year. transfusions as well as intraoperative transfusions for the entire
The aims of the present study were to: (1) determine modern cohort, for men, and for women, respectively.
intraoperative and postoperative transfusion rates for patients
undergoing primary THA who had routine use of TXA; (2) identify Results
demographic and surgical risk factors for perioperative transfusion;
and (3) determine factors to guide judicious use of type and screen Patient Population
testing both preoperatively and postoperatively.
Our study population included 6,405 patients who underwent
primary THA between January 2014 and January 2023. The overall
Methods and Materials perioperative, intraoperative, and postoperative transfusion rates
at our institution were 3.4, 1.0, and 2.7%, respectively. On Chi-square
Study Design and Population analysis, patients who received transfusions were more likely to be
older, women of a higher ASA class, and have lower preoperative
We conducted a retrospective cohort study of 16,912 patients Hgb values. Mean preoperative Hgb values were lower in patients
who underwent primary elective, unilateral THA at a single receiving postoperative (11.5 g/dL) and intraoperative (10.9 g/dL)
tertiary care academic institution from January 1, 2014, to January transfusions when compared to patients who were not transfused
31, 2023. Institutional Review Board approval was obtained before (13.5 g/dL, P < .001 for both comparisons). There was no difference
data collection. Inclusion criteria included TXA given before or in smoking status between cohorts. Demographic differences can
during surgery, Hgb values collected before surgery, and patients be found in Tables 1 and 2.
undergoing primary elective unilateral THA who had a diagnosis of
osteoarthritis. Patients were excluded if they were <18 years of Perioperative Characteristics
age (n ¼ 16), did not receive TXA (n ¼ 2,872), did not have
available preoperative Hgb values (n ¼ 7,185), or underwent Mean operative time was higher in both patients receiving
surgery for fracture (n ¼ 421), malignancy (n ¼ 2), or revision intraoperative (176.4 min) and postoperative (142.2 min) trans-
surgery (n ¼ 11). The standard protocol of TXA administration at fusions when compared to those not transfused (107.2 min, P < .001
our institution includes 1 gram of TXA administered intravenously for both comparisons). Likewise, mean EBL was significantly greater
(IV) before skin incision and during wound closure, for a total of 2 for patients requiring intraoperative (898 mL) and postoperative
gram of TXA. In cases where IV TXA is not possible due to (469 mL) transfusions when compared to patients requiring no
contraindications, 3 grams of topical TXA are placed in the wound. transfusion (298 mL, P < .001 for both comparisons). Furthermore,
This institutional protocol has remained unchanged over the study patients requiring either an intraoperative (6.0 days) or post-
period. operative (5.9 days) transfusion demonstrated significantly longer
 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126 121

Table 1 Table 2
Demographic Differences in Patients Receiving Postoperative Transfusion Versus No Demographic Differences in Patients Receiving Intraoperative Transfusion Versus No
Transfusion Following THA. Transfusion Following THA.

Patient Variable Postoperative No Transfusion P Patient Variable Intraoperative No Transfusion P

Transfusion (n ¼ 176) (n ¼ 6,166) Value Transfusion (n ¼ 63) (n ¼ 6,166) Value

Age (y, range, SD) 68 (21 to 95, 15.2) 63. (20 to 97, 11.8) <.001 Age (y, range, SD) 70 (20 to 98, 15.2) 63 (20 to 97, 11.8) <.001
Sex (Female %) 138 (78.4) 3,483 (56.5) <.001 Sex (Female %) 44 (69.8) 3,483 (56.5) .033
BMI (kg/m2, range, SD) 27.8 (15.4 to 50.6, 6.1) 29.5 (14.3 to 61.3, <.001 BMI (kg/m2, range, SD) 28.7 (18.5 to 48.1, 6.6) 29.5 (14.3 to 61.3, .276
6.0) 6.0)
Race (%) .405 Race (%) .552
White 121 (68.6) 4,405 (71.4) White 48 (76.2) 4,405 (71.4)
Black 31 (17.6) 881 (14.3) Black 10 (15.9) 881 (14.3)
Asian 6 (3.4) 142 (2.3) Asian 1 (1.6) 142 (2.3)
Other 18 (10.2) 738 (12.0) Other 4 (6.3) 738 (12.0)
Smoking Status (%) .351 Smoking Status (%) .938
Never 100 (56.8) 3,213 (52.1) Never 35 (55.6) 3,213 (52.1)
Former 67 (38.1) 2,419 (39.2) Former 23 (36.5) 2,419 (39.2)
Current 9 (5.1) 521 (8.4) Current 5 (7.9) 521 (8.4)
Unknown 0 (0.0) 13 (0.2) Unknown 0 (0.0) 13 (0.2)
ASA (%) <.001 ASA (%) <.001
I 6 (3.4) 442 (7.2) I 2 (3.2) 442 (7.2)
II 63 (35.8) 3,699 (60.0) II 17 (27.0) 3,699 (60.0)
III 94 (53.4) 1922 (31.2) III 37 (58.7) 1,922 (31.2)
IV 13 (7.4) 103 (1.7) IV 7 (11.1) 103 (1.7)
Surgical Approach (%) <.001 Surgical Approach (%) <.001
Posterior 132 (75.0) 3,762 (61.0) Posterior 53 (84.1) 3,762 (61.0)
Direct Anterior 38 (21.6) 1,972 (32.0) Direct Anterior 7 (11.1) 1,972 (32.0)
Modified Lateral 6 (3.4) 432 (7.0) Modified Lateral 3 (4.8) 432 (7.0)
Preoperative Hgb (g/dL, 11.5 (6.7 to 17.5, 1.8) 13.5 (7.5 to 19.2, <.001 Preoperative Hgb (g/dL, 10.9 (6.8 to 15.6, 2.1) 13.5 (7.5 to 19.2, <.001
range, SD) 1.5) range, SD) 1.5)
Hgb < 12 g/dL (%) 98 (55.7) 861 (14.0) <.001 Hgb < 12 g/dL (%) 46 (73.0) 861 (14.0) <.001
Hgb < 11 g/dL (%) 61 (34.7) 260 (4.2) <.001 Hgb < 11 g/dL (%) 35 (55.6) 260 (4.2) <.001

Bolded P-values are statistically significant (P < .05). Bolded P-values are statistically significant (P < .05).
THA, total hip arthroplasty; SD, standard deviation; BMI, body mass index; ASA, THA, total hip arthroplasty; SD, standard deviation; BMI, body mass index; ASA,
American Society of Anesthesiologists; Hgb, hemoglobin. American Society of Anesthesiologists; Hgb, hemoglobin.

LOS when compared to patients not requiring a transfusion (1.8 with either intraoperative or postoperative transfusion risk,
days, P < .001 for both comparisons). In addition, rates of discharge respectively (Table 5).
to a skilled nursing facility were higher in patients receiving For intraoperative transfusion risk, age (RR 1.06 [1.03 to 1.08], P
intraoperative (42.9%) or postoperative (30.1%) transfusions < .001), women (RR 2.50 [1.07 to 2.74], P < .001), ASA > 2 (RR 4.68
compared to patients not transfused (7.5%, P < .001 for both com- [2.73 to 8.02], P < .001), and increasing operative time (RR 1.019
parisons). A full account of perioperative variables is located in [1.013 to 1.026], P < .001) were associated with an increased risk of
Table 3. A subanalysis of operative time and EBL for patients who transfusion. Increasing preoperative Hgb levels (RR 0.41 [0.36 to
had preoperative Hgb between 11 and 12 can be found in Table 4. 0.48], P < .001) were protective against transfusion risk. Body mass
On univariate analysis, both intraoperative (81.0) and postoperative index (BMI) was not associated with the risk of intraoperative
(75.0) transfusion groups demonstrated higher rates of posterior transfusion (Table 5).
surgical approach use when compared to patients not transfused
(61.0%, P < .001 for both comparisons).
Cutoff Value Determination
Multivariate Risk Analysis
Following ROC curve analysis, Youden’s J statistic identified an
Upon performing a subanalysis for total transfusion risk, using Hgb as the value of 12 g/dL to be the value at which the risk for
multiple independent binary logistic regression, it was identified transfusion increases significantly (Figures 2 and 3). Accordingly,
that age (relative risk [RR] 1.04 [1.02 to 1.05], P < .001), women using binary logistic regression, an Hgb value of <12 g/dL was
(RR 2.39 [1.61 to 3.55], P < .001), ASA > 2 (RR 2.26 [1.59 to 3.21], associated with higher risk of all and intraoperative transfusions,
P < .001), and increasing operative time (RR 1.011 [1.007 to respectively (RR 8.8 and 15.6, P < .001) (Table 5). The risk for total
1.015], P < .001) were associated with increased risk of any and intraoperative transfusions significantly increased below an
transfusion. Of note, increasing BMI (RR 0.94 [0.92 to 0.96], P < Hgb value of 11 g/dL (RR 14.94 and 24.19, P < .001). Transfusion
.001) and increasing preoperative Hgb levels (RR 0.46 [0.43 to rates between preoperative Hgb values of 11 g/dL and 12 g/dL were
0.50], P < .001) were protective from transfusion on regression modest at 14.3 and 4.6%, respectively. Below 11 g/dL, total and
analysis (Table 5). However, when transfusion risk was stratified intraoperative transfusion rates were high at 27.5 and 10.1%,
by BMI, it became clear that transfusion risk and BMI share a respectively. Above 12 g/dL, total and intraoperative transfusion
bimodal relationship (Figure 1). Increasing BMI was protective rates were low (1.7 and 0.3%, respectively).
against perioperative transfusion up to the range of BMI’s from Similarly, following ROC curve analysis, an operative time of
30.0 to 34.9, which was associated with the lowest rate of 112 minutes was identified as an optimal cutoff value for the
transfusion. Above 35, increasing BMI is correlated with prediction of transfusion using Youden’s J statistic (Figure 4). Total
increasing transfusion risk (Figure 1). On multivariate analysis, transfusion rates above and below this value were determined to be
posterior (P ¼ .938, .166), anterior (P ¼ .128, .258), and lateral 6.2 and 1.6%, respectively. Intraoperative transfusion rates above
(P ¼ .488, .544) surgical approaches did not significantly correlate and below this value were 2.0 and 0.6%, respectively.
122 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126

Table 3
Perioperative Characteristics in Patients Receiving Intraoperative Transfusion Versus Postoperative Transfusion Versus No Transfusion.

Patient Variable Intraoperative Transfusion (n ¼ 63) Postoperative Transfusion (n ¼ 176) No Transfusion (n ¼ 6,166) P Value

Surgical Time (min, range, SD) 176.4 (74 to 480, 75.6) 142.2 (44 to 379, 56.3) 107.2 (50 to 324, 31.7) <.001
Estimated Blood Loss (mL, SD) 898.4 ± 624.8 468.8 ± 290.9 298.0 ± 134.6 <.001
Length of Stay (d, range, SD) 6.0 (1 to 2, 6.7) 5.9 (1 to 55, 5.5) 1.8 (0 to 21, 1.5) <.001
Discharge Disposition <.001
Home 31 (49.2) 106 (51.9) 5,613 (91.0)
SNF 27 (42.9) 53 (30.1) 463 (7.5)
ARF 5 (7.9) 16 (9.1) 83 (1.3)
Other 0 (0.0) 1 (0.6) 7 (0.1)

Bolded P-values are statistically significant (P < .05).

SD, standard deviation; SNF, skilled nursing facility; ARF, acute rehabilitation facility.

Discussion Patients in this study who had lower preoperative Hgb values
were significantly more likely to require perioperative transfusions.
The main findings of the present study were as follows: (1) The This intuitive finding has been well-documented in previous liter-
overall perioperative, intraoperative, and postoperative transfusion ature [7,20,24e27]. Patients who are anemic preoperatively are less
rates for THA were 3.4, 1.0, and 2.7%, respectively, at a single large able to tolerate perioperative blood loss and thus more likely to
academic institution with routine use of TXA. (2) Patients who had receive transfusions. The open question is: what preoperative Hgb
lower preoperative Hgb levels were more likely to require trans- threshold predicts the need for perioperative transfusion and,
fusions; patients who had Hgb values less than 11 g/dL and 12 g/dL subsequently, indicates a need for preoperative, intraoperative, or
had overall transfusion rates of 27.5 and 14.3%, respectively. (3) postoperative type and screen testing? This question has been
Patients requiring perioperative transfusion were significantly investigated previously; however, these studies were performed
older, more commonly women, had lower BMIs, and were more before the routine use of TXA [26,28]. The present study found that
commonly ASA class > II, compared to patients not requiring a a preoperative Hgb below 11.0 g/dL is a strong independent
transfusion. These trends held true on multivariate analysis, predictor for perioperative transfusion (P < .001), with 27.5% of
accounting for the confounding effect of preoperative Hgb. (4) patients who had a preoperative Hgb < 11 g/dL requiring intra-
Patients requiring a transfusion were significantly more likely to operative or postoperative transfusion. Preoperative Hgb below
have a prolonged surgical time (>112 minutes) and increased EBL. 12.0 g/dL was an additional significant predictor for perioperative
(5) Patients requiring a perioperative transfusion had an increased transfusion (P < .001), with 16.7% of patients who had a preop Hgb
mean LOS and a decreased rate of discharge to home. < 12 g/dL requiring a transfusion. This is compared to a 1.7% overall
The overall perioperative transfusion rate for THA was found to transfusion rate for patients with Hgb >12 g/dL. Data from before
be 3.4, with 1.0% of these being intraoperative transfusions. routine use of TXA revealed a mean drop in Hgb of approximately 4
Transfusion rates vary widely across institutions and geographic g/dL following THA [6]. With a postoperative transfusion threshold
locations and have changed drastically in recent years. Over the of approximately 7 g/dL, patients who have preoperative Hgb less
past 2 decades, studies have reported varying allogenic transfusion than 11 g/dL would routinely require transfusion. The present study
rates between 10 and 70% for routine primary THAs and TKAs suggests that patients who had a preoperative Hgb < 11 g/dL still
[20e22]. Transfusion rates have steadily declined over the past benefit from a preoperative type and screen, as these patients
decade. In 2017, Bedard et al. analyzed over 69,000 THA cases and continue to be at high risk for requiring either intraoperative or
reported perioperative THA transfusion rates of ~9% [23]. Contrast postoperative transfusion. The transfusion threshold of 7 g/dL
this rate to the study by Antoniou et al. in 2014, where they found commonly utilized at our institution is more restrictive than some
an overall transfusion rate of 22.2% after analyzing over 9,000 THA’s prior reports in the orthopaedic literature [29] and is informed by
[24]. The present study’s overall perioperative transfusion rate, at an evolving understanding of blood management based on exten-
3.7%, remains substantially below the rates cited in the aforemen- sive literature over the past 2 decades [30]. Transfusion triggers and
tioned studies. This may reflect improvements in surgical thresholds may continue to change and evolve as advances in
techniques, patient optimization, and perioperative care, including perioperative optimization, surgical techniques and hemostasis are
improved blood product stewardship. Institution-specific policies achieved in the coming years.
and procedures likely play a role in the relatively low transfusion Patients requiring perioperative transfusions in this study were
rates demonstrated in the present study. Despite these low rates, significantly older than patients not requiring transfusions; this is
there remains room for improvement. Studies have reported that consistent with previous literature [24,31]. Antoniou et al. (2014)
with improved blood management protocols, decreasing invasive- analyzed 9,362 THAs and found that, on average, transfused
ness of procedures, and improved preoperative optimization, patients were 4 years older compared to nontransfused patients, P
transfusion rates following THA could potentially be reduced to < .0001 [24]. Less restrictive transfusion criteria may be used with
near zero in the coming years [15]. advanced age, as older arthroplasty patients are less tolerant of

Table 4
Subanalysis of Surgical Time and EBL in Patients Receiving Intraoperative Versus Postoperative Versus No Transfusion With Preoperative Hgb Between 11 and 12 g/dL.

Patient Variable Intraoperative Transfusion (n ¼ 12) Postoperative Transfusion (n ¼ 40) No Transfusion (n ¼ 679) P Value

Surgical Time (min, range, SD) 165.1 (104 to 224, 38.3) 143.4 (72 to 361, 66.2) 109.8 (42 to 286, 30.5) <.001
EBL (mL, SD) 316.7 ± 76.4 325.0 ± 138.5 297.6 ± 144.1 .845

Bolded P-values are statistically significant (P < .05).

SD, standard deviation, EBL, estimated blood loss; Hgb, hemoglobin.
 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126 123

Table 5
Multiple Independent Binary Logistic Regression for Intraoperative and Post-
operative Transfusion Risk.

Patient Variable RR for P RR for P

Intraoperative Value Postoperative Value
Transfusion [95% CI] Transfusion [95%
CI]

Age (y) 1.06 (1.03 to 1.08) <.001 1.04 (1.02 to 1.05) <.001
BMI (kg/m2) 0.99 (0.95 to 1.03) .702 0.94 (0.92 to 0.96) <.001
Sex (Female) 2.50 (1.07 to 2.74) <.001 2.39 (1.61 to 3.55) <.001
ASA > 2 4.68 (2.73 to 8.02) <.001 2.26 (1.59 to 3.21) <.001
Surgical ime (min) 1.019 (1.013 to <.001 1.011 (1.007 to <.001
1.026) 1.015)
Surgical Time 12.05 (4.38 to <.001 2.85 (2.26 to 3.59) <.001
>100 min 33.20)
Surgical Time >110 6.88 (3.59 to 13.19) <.001 3.29 (2.66 to 4.07) <.001
min
Surgical Time >120 7.88 (4.41 to 14.08) <.001 4.18 (3.54 to 4.94) <.001
min
Preoperative 0.41 (0.36 to 0.48) <.001 0.46 (0.43 to 0.50) <.001
Hemoglobin
Level >12 g/dL
Hemoglobin <12g/ 15.56 (8.83 to <.001 8.80 (7.15 to 10.84) <.001
Fig. 2. Transfusion ROC analysis for overall cohort. AUC, area under the curve; ROC,
dL 27.17)
receiver-operating characteristic. AUC ¼ 80.0%, 95% CI 76.8 to 83.1%. CI, confidence
Hemoglobin <11g/ 24.19 (14.58 to <.001 14.94 (11.89 to <.001
interval.
dL 40.14) 18.77)
Surgical Approach
Posterior 1.05 (0.31 to 3.56) .938 0.81 (0.59 to 1.09) .166
Direct Anterior 0.33 (0.08 to 1.39) .128 1.20 (0.87 to 1.66) .258 The present study found a bimodal distribution of all transfusion
Modified Lateral 0.66 (0.21 to 2.12) .488 1.19 (0.68 to 2.09) .544 risk when stratified by BMI (Figure 1); increasing BMI was
Bolded P-values are statistically significant (P < .05). protective up to a limit (35), after which increasing BMI began to
RR, relative risk; CI, confidence interval; BMI, body mass index; ASA, American increase transfusion risk (Figure 1). This finding may help resolve a
Society of Anesthesiologists. controversy in the literature; prior studies have shown increasing
BMI to be either a protective factor against transfusion [36,37] or a
risk factor for transfusion [25,26]. Both are conceivable. Increased
extended postoperative anemia as a result of decreased hemato- basal blood volume in heavier patients could make them more
poietic activity and platelet function [32]. tolerant of surgical blood loss and resistant to postoperative
Similarly, the finding that women are at increased risk for anemia, thus decreasing the likelihood of a perioperative
transfusion has also been well-documented in the literature transfusion [27]. On the other hand, patients who had higher BMIs
[32,33]. Women tend to have lower preoperative Hgb values when (>30) may have experienced increased operative blood loss,
compared to men; in addition to the impact of menstrual blood loss predisposing them to transfusion [38]. A bimodal distribution of
[34], women do not experience the stimulatory effect of androgens risk may help explain why, depending on the mean BMI of a study
on red cell production [35]. Thus, women tend to be less resistant to cohort, BMI could be classified as a risk or protective factor for
perioperative blood loss and anemia and generally end up requiring perioperative transfusion.
higher rates of transfusions. Interestingly, women remained pre-
dictive of transfusion risk on multivariate analysis after controlling
for preoperative Hgb levels in the present study. This suggests that
factors other than preoperative Hgb may also predispose women to
a higher risk of perioperative transfusion.

Fig. 3. Intraoperative transfusion ROC analysis for overall cohort. AUC, area under the
Fig. 1. Distribution of the incidence of transfusion by BMI classification. BMI, body curve; ROC, receiver-operating characteristic. Overall AUC ¼ 83.4%, 95% CI 77.1 to
mass index. 89.7%. CI, confidence interval.
124 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126

Fig. 4. Overall and Intraoperative Transfusion ROC analysis by operative time in minutes. A. Any Transfusion: AUC ¼ 74.9%, 95% CI 71.6 to 78.2%; intraoperative transfusion: AUC ¼
83.1%, 95% CI 77.9 to 88.4%. ROC, receiver operating characteristic; AUC, area under the curve; CI, confidence interval.

An increased risk of perioperative transfusion in patients who increased risk of deep infections, venous thromboses, and pul-
had higher ASA classes has previously been well-documented [24]. monary emboli in patients who received a transfusion when
Antoniou et al. (2014) reported an odds ratio for transfusion of 1.3 compared to patients who did not receive a transfusion [42]. It is
(P < .001) for patients who have ASA class > II. Other studies have indeed likely that these complications contribute to prolonged LOS
found that patients who are ASA class IV are 14.71 times more likely and may result in the utilization of higher levels of care following
to receive a transfusion, compared to patients who are ASA class I discharge.
[25]. The presence of chronic systemic disease predisposes to It was previously routine practice to obtain type and screen
anemia and is associated with increased bleeding risk in patients testing before primary THA, due to historically high transfusion
who have higher ASA classes [39]. Patients who have higher ASA rates before the adoption of TXA [14]. However, since transfusions
class are also more vulnerable to complications of anemia and thus following primary THA have become less common in modern
more likely to benefit from transfusion at a given Hgb value. practice, universal type and screen practices represent a source of
Increased EBL and surgical time have consistently been associ- inefficiency for many patients at low risk of transfusion. This in-
ated with an increased risk of transfusion in prior literature [40]. efficiency and increased cost particularly burden patients traveling
The present study found an approximately 4 times greater risk for from afar, who may require an extra preoperative visit to undergo
all-transfusions when surgical time exceeded 120 minutes in THA screening before the day of surgery; unlike routine bloodwork,
(P < .001). On ROC analysis of operative time, the present study preoperative type and screen testing is typically restricted to the
found an area under the curve of 83.1% for intraoperative trans- specific blood bank that would issue blood in the case of a peri-
fusions, and Youden’s J statistical analysis found an optimal surgical operative transfusion [43]. By eliminating universal type and screen
time cutoff of 112 minutes for predicting transfusion risk. These testing, studies have estimated an average of $200 per patient can
findings are consistent with the study by Parvizi et al. (2019), where be saved, in addition to any direct or indirect costs incurred by
they found that for every 15 minutes of operative time, blood loss patients to setup and travel to extra testing appointments [14,43]. If
increases on average by 211.5 mL [40]. As surgical time and a patient deemed low-risk for transfusion based on preoperative
consequent blood loss increase, so does a patient’s risk of post- factors is later determined to be higher-risk for a postoperative
operative anemia and subsequent transfusion. transfusion based on surgical duration or EBL, then a type and
Patients requiring transfusion in the present study demon- screen could be drawn with a complete blood count during surgery
strated longer hospital stays compared to patients not requiring a or in the postanesthesia care unit. This avoids the inefficiencies and
transfusion; this is in accordance with prior literature [41,42]. logistical issues of arranging extra preoperative appointments, or
Similar to the present study, Rodriguez et al. (2022) found that drawing blood in the preoperative holding area, which can delay
transfused patients demonstrated a significantly longer mean LOS and disrupt case scheduling.
when compared to patients not requiring a transfusion, 4.8 versus Based on the current study, patients who have a Hgb < 11 g/dL
3.5 days, respectively (P < .01) [41]. The finding that transfused before primary elective THA stand to benefit from treatment of
patients have lower rates of discharge to home compared to pa- anemia before surgery, as higher preoperative Hgb levels are
tients who did not receive transfusions is also consistent with associated with a substantially decreased risk of blood
previous literature [42]. Barsoum et al. (2014) conducted a transfusion. If Hgb levels cannot be increased above 11 g/dL, our
nationwide analysis of over 2 million THAs from 2001 to 2009 and data suggest it remains prudent to collect a type and screen
found that, on average, patients who receive a perioperative before primary elective THA. Patients who have preoperative Hgb
transfusion are 28% more likely to be discharged to an inpatient between 11 and 12 g/dL were at modest risk of requiring trans-
facility when compared to patients who did not receive a trans- fusion at 14.3 and 4.6% for overall and intraoperative transfusion
fusion [42]. Barsoum et al. attribute the increased LOS and lower rates, respectively. The conservative approach would be to
rate of discharge to home, in part, to the higher rates of medical continue preoperative type and screen testing in these patients.
and surgical complications experienced by patients receiving Alternatively, patient-specific factors such as age, sex, ASA class,
perioperative transfusions. The authors demonstrated an and BMI could potentially help inform judicious testing in these
 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126 125

patients, as might surgeon-specific metrics such as typical surgi- Ran Schwarzkopf: Writing e review & editing, Supervision,
cal times and blood loss. Furthermore, operative time and EBL can Methodology, Conceptualization. Matthew Hepinstall: Writing e
inform the need for blood counts and type and screen testing review & editing, Supervision, Resources, Methodology, Investiga-
during or immediately after surgery in all patients who do not tion, Conceptualization.
have a preoperative type and screen. Our results suggest a type
and screen be strongly considered if surgical time exceeds References
120 minutes. For patients who have preoperative Hgb > 12 g/dL,
overall transfusion rates were rare at 1.7%, with 0.3% of these [1] Satalich JR, Lombardo DJ, Newman S, Golladay GJ, Patel NK. Cementation in
total hip arthroplasty: history, principles, and technique. EFORT Open Rev
being intraoperative transfusions. Given the rare occurrence of 2022;7:747. https://doi.org/10.1530/EOR-22-0002.
transfusion in this group, our data suggest that universal type and [2] Shichman I, Roof M, Askew N, Nherera L, Rozell JC, Seyler TM, et al. Projections
screen are likely not necessary when preoperative Hgb is greater and epidemiology of primary hip and knee arthroplasty in medicare patients
to 2040-2060. JBJS Open Access 2023;8. https://doi.org/10.2106/
than 12 g/dL. This recommendation falls in accordance with the
JBJS.OA.22.00112.
study by Dexter et al. (2012), which analyzed over 160,000 elec- [3] Sloan M, Premkumar A, Sheth NP. Projected volume of primary total joint
tive cases, including over 5,000 primary THAs. They found that if arthroplasty in the U.S., 2014 to 2030. J Bone Joint Surg Am 2018;100:
the lower limit of the 95% confidence interval of transfusion rate 1455e60. https://doi.org/10.2106/JBJS.17.01617.
[4] MB H, JF M. Total blood loss in major joint arthroplasty. A comparison of
falls below 5.0%, which is the case in our study, then type and cemented and noncemented hip and knee operations. J Arthroplasty 1988;3:
screen testing were shown to be unnecessary [44]. S47e9. https://doi.org/10.1016/S0883-5403(88)80008-1.
This study has several potential limitations that are worth [5] Billote DB, Glisson SN, Green D, Wixson RL. A prospective, randomized study
of preoperative autologous donation for hip replacement surgery. J Bone Joint
noting. Retrospective studies cannot establish causality and are Surg Am 2002;84:1299e304. https://doi.org/10.2106/00004623-200208000-
prone to biases such as selection and observer biases. As our study 00002.
included patients operated on by several different surgeons and [6] Callaghan JJ, O’Rourke MR, Liu SS. Blood management: issues and options.
J Arthroplasty 2005;20:51e4. https://doi.org/10.1016/J.ARTH.2005.03.018.
surgical teams, variability in incision sizes, soft tissue dissection, [7] Donovan RL, Lostis E, Jones I, Whitehouse MR. Estimation of blood volume and
and hemostasis strategies may have introduced heterogeneity to blood loss in primary total hip and knee replacement: an analysis of formulae
our results. Furthermore, variable clinical judgment and decision- for perioperative calculations and their ability to predict length of stay and
blood transfusion requirements. J Orthop 2021;24:227. https://doi.org/
making may have introduced heterogeneity into transfusion 10.1016/J.JOR.2021.03.004.
practices across different physicians and surgeons. Our population [8] Cai L, Chen L, Zhao C, Wang Q, Kang P. Influencing factors of hidden blood loss
was selected from 1 high-volume center with high utilization of after primary total hip arthroplasty through the posterior approach: a retro-
spective study. BMC Musculoskelet Disord 2023;24:582. https://doi.org/
regional anesthesia and relatively restrictive transfusion practices,
10.1186/S12891-023-06716-Z.
and thus may not be representative of the population undergoing [9] Shah A, Palmer AJR, Klein AA. Strategies to minimize intraoperative blood loss
THA in other environments. Additionally, our study only analyzed during major surgery. Br J Surg 2020;107:e26e38. https://doi.org/10.1002/
primary elective unilateral THA. Thus, our results and recommen- BJS.11393.
[10] Sizer SC, Cherian JJ, Elmallah RDK, Pierce TP, Beaver WB, Mont MA. Predicting
dations cannot be applied to other procedure types such as revision blood loss in total knee and hip arthroplasty. Orthop Clin North Am 2015;46:
THA, bilateral THA, or THA for a diagnosis of fracture or malignancy. 445e59. https://doi.org/10.1016/J.OCL.2015.06.002.
These represent areas for further investigation and warrant the [11] Maxwell MJ, Wilson MJA. Complications of blood transfusion. Contin Educ
Anaesth Crit Care Pain 2006;6:225e9. https://doi.org/10.1093/BJACEACCP/
development of procedure specific transfusion guidelines. As the MKL053.
vast majority of patients within our study received IV TXA per our [12] Bower W, Jin L, Underwood MJ, Lam FY, Lai PB. Peri-operative blood trans-
institutional protocol, our results may not be generalizable to other fusion increases length of hospital stay and number of postoperative com-
plications in non-cardiac surgical patients. Hong Kong Med J 2010;16:116e36.
methods of TXA administration. [13] Spahn DR. Anemia and patient blood management in hip and knee surgery: a
systematic review of the literature. Anesthesiology 2010;113:482e95. https://
Conclusions doi.org/10.1097/ALN.0B013E3181E08E97.
[14] Christopher ZK, Bruce MR, Reynolds EG, Spangehl MJ, Bingham JS, Kraus MB.
Routine type and screens are unnecessary for primary total hip and knee
In the age of TXA, both intraoperative and overall transfusions arthroplasties at an academic hospital. Arthroplast Today 2020;6:941. https://
for patients undergoing THA are rare for patients who have a pre- doi.org/10.1016/J.ARTD.2020.10.006.
[15] Lindman IS, Carlsson LV. Extremely low transfusion rates: Contemporary
operative Hgb above 12 g/dL. Nevertheless, patients who have a
primary total hip and knee arthroplasties. J Arthroplasty 2018;33:51e4.
preoperative Hgb less than 11 g/dL remain at significant risk of https://doi.org/10.1016/J.ARTH.2017.07.034.
requiring either an intraoperative or postoperative transfusion. [16] Frank SM, Rothschild JA, Masear CG, Rivers RJ, Merritt WT, Savage WJ, et al.
These patients are likely to benefit from presurgical optimization of Optimizing preoperative blood ordering with data acquired from an anes-
thesia information management system. Anesthesiology 2013;118:1286e97.
their anemia to minimize transfusion risk and perioperative https://doi.org/10.1097/ALN.0B013E3182923DA0.
complications. If surgery is undertaken on these patients who do [17] Rinehart JB, Lee TC, Kaneshiro K, Tran MH, Sun C, Kain ZN. Perioperative blood
not have successful optimization, a preoperative type and screen ordering optimization process using information from an anesthesia infor-
mation management system. Transfusion 2016;56:938e45. https://doi.org/
should be obtained. For patients who have preoperative Hgb levels 10.1111/TRF.13492.
between 11 and 12 g/dL, preoperative type and screen testing [18] Woodrum CL, Wisniewski M, Triulzi DJ, Waters JH, Alarcon LH, Yazer MH. The
remain reasonable. If routine testing is not performed in this group, effects of a data driven maximum surgical blood ordering schedule on pre-
operative blood ordering practices. Hematology 2017;22:571e7. https://
patient factors such as age, sex, BMI, ASA class, and surgical factors doi.org/10.1080/10245332.2017.1318336.
such as surgical time and EBL may help inform judicious type and [19] Fluss R, Faraggi D, Reiser B. Estimation of the Youden Index and its associated
screen testing. cutoff point. Biom J 2005;47:458e72. https://doi.org/10.1002/BIMJ.200410135.
[20] Carling MS, Jeppsson A, Eriksson BI, Brisby H. Transfusions and blood loss in
total hip and knee arthroplasty: a prospective observational study. J Orthop
CRediT authorship contribution statement Surg Res 2015;10:48. https://doi.org/10.1186/S13018-015-0188-6.
[21] Rosencher N, Kerkkamp HEM, Macheras G, Munuera LM, Menichella G,
Barton DM, et al. Orthopedic Surgery Transfusion Hemoglobin European
Muhammad A. Haider: Writing e review & editing, Writing e
Overview (OSTHEO) study: blood management in elective knee and hip
original draft, Formal analysis, Data curation. Spencer A. Ward: arthroplasty in Europe. Transfusion 2003;43:459e69. https://doi.org/10.1046/
Writing e original draft, Formal analysis, Data curation. Vinaya J.1537-2995.2003.00348.X.
Rajahraman: Formal analysis, Data curation. Joshua C. Rozell: [22] Stanworth SJ, Cockburn HAC, Boralessa H, Contreras M. Which groups of pa-
tients are transfused? A study of red cell usage in London and southeast
Writing e review & editing, Supervision, Methodology. William England. Vox Sang 2002;83:352e7. https://doi.org/10.1046/J.1423-
Macaulay: Writing e review & editing, Supervision, Methodology. 0410.2002.00237.X.
126 M.A. Haider et al. / The Journal of Arthroplasty 40 (2025) 119e126

[23] Bedard NA, Pugely AJ, Lux NR, Liu SS, Gao Y, Callaghan JJ. Recent trends in factors associated with greater risk of allogeneic blood transfusion n.d. Br J
blood utilization after primary hip and knee arthroplasty. J Arthroplasty Anaesth 2012;108:63e71. https://doi.org/10.1093/bja/aer326.
2017;32:724e7. https://doi.org/10.1016/J.ARTH.2016.09.026. [34] Gombotz H, Schreier G, Neubauer S, Kastner P, Hofmann A. Gender disparities
[24] Hart A, Khalil JA, Carli A, Huk O, Zukor D, Antoniou J. Blood transfusion in in red blood cell transfusion in elective surgery: a post hoc multicentre cohort
primary total hip and knee arthroplasty. Incidence, risk factors, and thirty-day study. BMJ Open 2016;6:12210. https://doi.org/10.1136/BMJOPEN-2016-
complication rates. J Bone Joint Surg Am 2014;96:1945e51. https://doi.org/ 012210.
10.2106/JBJS.N.00077. [35] Murphy WG. The sex difference in haemoglobin levels in adults d mecha-
[25] Guerin S, Collins C, Kapoor H, McClean I, Collins D. Blood transfusion nisms, causes, and consequences. Blood Rev 2014;28:41e7. https://doi.org/
requirement prediction in patients undergoing primary total hip and knee 10.1016/J.BLRE.2013.12.003.
arthroplasty. Transfus Med 2007;17:37e43. https://doi.org/10.1111/J.1365- [36] Cao G, Chen G, Yang X, Huang Q, Huang Z, Xu H, et al. Obesity does not in-
3148.2006.00698.X. crease blood loss or incidence of immediate postoperative complications
[26] Nuttall GA, Santrach PJ, Oliver WC, Horlocker TT, Shaughnessy WJ, during simultaneous total knee arthroplasty: a multicenter study. Knee
Cabanela ME, et al. The predictors of red cell transfusions in total hip 2020;27:963e9. https://doi.org/10.1016/J.KNEE.2020.01.012.
arthroplasties. Transfusion 1996;36:144e9. https://doi.org/10.1046/J.1537- [37] Frisch N, Wessell NM, Charters M, Peterson E, Cann B, Greenstein A, et al.
2995.1996.36296181927.X. Effect of body mass index on blood transfusion in total hip and knee
[27] Bashaireh K, Aljararhih O, Alawneh K. Impact of body mass index on hemo- arthroplasty. Orthopedics 2016;39:e844e9. https://doi.org/10.3928/
globin level and blood transfusion in total knee arthroplasty: a retrospective 01477447-20160509-04.
case control study. Ann Med Surg 2020;55:180. https://doi.org/10.1016/ [38] Noticewala MS, Nyce JD, Wang W, Geller JA, Macaulay W. Predicting need for
J.AMSU.2020.05.028. allogeneic transfusion after total knee arthroplasty. J Arthroplasty 2012;27:
[28] Boettner F, Altneu EI, Williams BA, Hepinstall M, Sculco TP. Nonanemic patients 961e7. https://doi.org/10.1016/J.ARTH.2011.10.008.
do not benefit from autologous blood donation before total hip replacement. [39] Daabiss M. American Society of Anaesthesiologists physical status classification.
HSS J 2010;6:66e70. https://doi.org/10.1007/S11420-009-9145-4. Indian J Anaesth 2011;55:111. https://doi.org/10.4103/0019-5049.79879.
[29] Mitchell MD, Betesh JS, Ahn J, Hume EL, Mehta S, Umscheid CA. Transfusion [40] Ross D, Erkocak O, Rasouli MR, Parvizi J. Operative time directly correlates
thresholds for major orthopedic surgery: a systematic review and meta-analysis. with blood loss and need for blood transfusion in total joint arthroplasty. Arch
J Arthroplasty 2017;32:3815e21. https://doi.org/10.1016/J.ARTH.2017.06.054. Bone Jt Surg 2019;7:229.
[30] Chaudhry YP, Macmahon A, Rao SS, Mekkawy KL, Toci GR, Oni JK, et al. Pre- [41] Iacobelli DS, Syku M, Abutalib Z, Berliner ZP, Joseph A, Cushner F, et al.
dictors and outcomes of postoperative hemoglobin of <8 g/dL in total joint Transfusion avoidance in severely anemic total hip and total knee arthro-
arthroplasty. J Bone Joint Surg Am 2022;104:166e71. https://doi.org/10.2106/ plasty patients: an analysis of risk. Arthroplast Today 2022;14:128e32.
JBJS.20.01766. https://doi.org/10.1016/J.ARTD.2022.01.033.
[31] Pedersen AB, Mehnert F, Overgaard S, Johnsen SP. Allogeneic blood trans- [42] Saleh A, Small T, Pillai ALPC, Schiltz NK, Klika AK, Barsoum WK. Allogenic
fusion and prognosis following total hip replacement: a population-based blood transfusion following total hip arthroplasty: results from the nation-
follow up study. BMC Musculoskelet Disord 2009;10:167. https://doi.org/ wide inpatient sample, 2000 to 2009. J Bone Joint Surg Am 2014;96:e155.
10.1186/1471-2474-10-167. https://doi.org/10.2106/JBJS.M.00825.
[32] Barr PJ, Donnelly M, Cardwell C, Alam SS, Morris K, Parker M, et al. Drivers of [43] Basavarajegowda A, Shastry S. Pretransfusion Testing. Treasure Island (FL):
transfusion decision making and quality of the evidence in orthopedic sur- StatPearls Publishing; 2023. StatPearls [Internet].
gery: a systematic review of the literature. Transfus Med Rev 2011;25: [44] Dexter F, Ledolter J, Davis E, Witkowski TA, Herman JH, Epstein RH. Systematic
304e316.e6. https://doi.org/10.1016/J.TMRV.2011.04.003. criteria for type and screen based on procedure’s probability of erythrocyte
[33] Walsh TS, Palmer J, Watson D, Biggin K, Seretny M, Davidson H, et al. Mul- transfusion. Anesthesiology 2012;116:768e78. https://doi.org/10.1097/
ticentre cohort study of red blood cell use for revision hip arthroplasty and ALN.0B013E31824A88F5.