Incidence, outcomes, and risk factors of Heparin-induced thrombocytopenia in patients undergoing primary and revision knee arthroplasty

Abstract

Background

See More

Heparin-induced thrombocytopenia (HIT) is a serious complication associated with heparin use in orthopedic surgery. However, its incidence and risk factors in total knee arthroplasty (TKA) and revision TKA (RTKA) remain unclear. This study aimed to evaluate how preoperative comorbidities, hospital characteristics, and patient demographics influence the incidence of HIT in patients who underwent TKA and RTKA. Differences in postoperative complications, mortality, hospital length of stay, HIT-related costs, and changes in HIT risk following revision surgery were examined.

Methods

This retrospective study examined data from the National Inpatient Sample (NIS) on patients with TKA and RTKA from 2010 to 2019, categorizing them by the incidence of HIT. Demographics (race, sex, and age) and hospital (admission type, insurance, hospital size, teaching status, and region) details were analyzed. Mortality, comorbidities, and perioperative complications were assessed, and logistic regression analyses were performed to identify potential risk factors.

Results

Pulmonary circulatory disorders were strongly associated with HIT in both TKA (P < 0.01, OR = 3.43) and RTKA (P < 0.01, OR = 4.13) groups. Teaching hospitals were associated with lower odds of HIT in the TKA group (P = 0.01, OR = 0.62). Risk factors in RTKA included valvular heart disease (OR = 2.50, 95% CI 1.12–5.57). Common complications among HIT cases included deep vein thrombosis, acute myocardial infarction, and acute renal failure. Pulmonary embolism, postoperative pneumonia, procedural pain, and prosthetic joint infection were more common in TKA group, whereas dyspnea was more prevalent in RTKA group.

Conclusions

Certain preoperative comorbidities and baseline characteristics are associated with increased HIT risk following TKA. RTKA is associated with higher odds of HIT and a greater incidence of adverse clinical outcomes. These findings may support the need for improved risk stratification and postoperative planning to reduce complications and enhance recovery.

Background

Total knee arthroplasty (TKA), one of the most successful orthopedic procedures of the 20th century, can effectively alleviate the pain from end-stage knee disease, rebuild patients’ joint function, and improve their quality of life. The increasing number of primary TKA cases, expansion of hospitals performing these procedures, and disparities in postoperative care resources have collectively contributed to the rising number of revision knee surgeries. Postoperative complications such as deep vein thrombosis (DVT) and venous thromboembolism (VTE) may occur, resulting in prolonged hospital stays, higher hospitalization costs, and increased mortality [1]. According to American College of Chest Physicians (ACCP) guidelines [2], routine anticoagulation prophylaxis is recommended postoperatively to lower the risk of VTE. Aspirin may serve as an anticoagulant agent in patients undergoing orthopedic surgery; however, its efficacy is generally lower than that of low-molecular-weight heparin (LMWH) [3], which remains the preferred anticoagulant agent in certain cases. Although effective in preventing VTE, heparin-based anticoagulation carries specific risks.

Heparin-induced thrombocytopenia (HIT)—a potentially severe complication linked to heparin therapy—is associated with DVT, pulmonary embolism (PE), and ischemic stroke. This leads to increased use of healthcare resources, prolonged hospital stays, higher hospitalization costs, more complications, and increased mortality [4]. Identifying HIT risk factors and understanding their clinical impact in patients who underwent surgery are essential for prevention and management.

Despite extensive studies on HIT following primary TKA, evidence regarding revision TKA (RTKA) remains remarkably scarce. This knowledge gap is clinically important because (1) RTKAs inherently carry higher thrombotic risks caused by longer operative time and greater tissue trauma [5], (2) re-exposure to heparin during revision surgery can facilitate antibody-mediated platelet activation [6], and (3) contemporary anticoagulation protocols increasingly use direct oral anticoagulants (DOACs) in primary TKA, while heparin remains predominant in complex revisions. The distinct pathophysiology and management challenges necessitate separate risk stratification for these populations.

Studies have indicated that factors such as sex, race, hospital location, and comorbidities may influence HIT rates after TKA and RTKA, however, comprehensive analyses are scarce, and the association between HIT and specific comorbidities remains unclear [7, 8]. Research on RTKA is limited. To assess the incidence of HIT, risk factors, and outcomes across the two surgical categories, the National Inpatient Sample (NIS) was analyzed. This study investigated the incidence of HIT in TKA and RTKA, with particular attention to the observation that revision procedures—often performed after primary TKA failure—appear to be associated with a higher burden of HIT-related complications. A better understanding of these associations may help guide clinical decision-making and improve perioperative care.

Methods

Data source

This retrospective cohort study utilized the NIS database from 2010 to 2019. This study adhered to the RECORD guidelines for observational research using routinely collected health data. This study used data from the NIS database, a comprehensive US database of hospital admissions managed by the Agency for Healthcare Research and Quality. It represents approximately 20% of annual hospitalization samples from > 1,000 hospitals and collects data on demographics, hospital characteristics, length of stay (LOS), costs, and diagnostic and procedural codes based on the International Classification of Diseases 9 and 10 Clinical Modification (ICD-9-CM and ICD-10-CM, respectively) standards. Ethical approval was not required because the data are anonymized and publicly accessible.

Data collection

HIT cases were identified using ICD-9-CM or ICD-10-CM codes. The study included adult patients (aged ≥ 18 years) who underwent TKA with a hospitalization of ≥ 24-h or RTKA for mechanical failure, infection, or instability (excluding same-day surgeries), all with documented heparin/LMWH exposure during hospitalization between January 1, 2010 and December 31, 2019. Patients aged < 18 years who had prior thrombosis, conditions contributing to thrombocytopenia, and missing data for some research variables were excluded. Eligible patients were divided into two groups based on the postoperative incidence of HIT. The primary outcome was the risk factors for patients with HIT, including demographics (age, sex, and race), hospital features (bed size, teaching status, location, and region), comorbidities (anemia, diabetes, hypertension, chronic pulmonary disease, renal failure, and other complications, quantified using the Charlson comorbidity index and Elixhauser comorbidity index), and surgical characteristics (primary or revision procedures). Age was analyzed as continuous and categorical variables (≤ 40, 41–50, 51–60, 61–70, and ≥ 71 years); sex was categorized as male or female; and race included White, Black, Hispanic, Asian or Pacific Islander, Native American, and others. Hospital bed size was classified as small (< 300 beds), medium (300–499 beds), and large (≥ 500 beds). Teaching status was divided into teaching and nonteaching hospitals, location was categorized as urban or rural, and regions included Northeast, Midwest, South, and West. The secondary outcomes included the incidence of HIT-associated acute venous (DVT and PE) or arterial (myocardial infarction, renal artery embolism, and other arterial thromboembolism) thrombosis, mortality, LOS, total hospital charges, and disposition. Comorbidities and perioperative complications were extracted using ICD codes. (Fig. 1).

Fig. 1

Flow diagram of the study a: National Inpatient Sample Database; b: International Classification of Diseases-9/10; c: Total Knee Arthroplasty; d: Revision Total Knee Arthroplasty e: Heparin-Induced Thrombocytopenia; f:Venous Thromboembolism

Data analysis

Statistical analyses were performed using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA). Normally distributed data are shown as the means ± standard deviations and were compared using an independent sample t-test. The Mann‒Whitney U test was used to analyze non-normally distributed data. Categorical variables were analyzed using the chi-square tests. Logistic regression was used to identify the risk factors of HIT, and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Significance was set at P< 0.05 [9]. Given the rarity of certain exposures (e.g., nonbleeding ulcers), point estimates of ORs should be interpreted with caution owing to wide CIs.

Results

Incidence of HIT after TKA and RTKA

Between 2010 and 2019, a total of 1,253,697 patients underwent TKA and 127,606 patients underwent RTKA. Among these, 137 and 47 patients developed HIT, corresponding to 10-year incidence rates of 0.01% and 0.03%, respectively. Over the study period, the incidence of HIT in the TKA group showed a slight decline from 0.02% in 2010 to approximately 0.006% in 2019. In the RTKA group, the incidence fluctuated, with notable nadirs in 2015 and 2017, followed by an upward trend after 2017 (Fig. 2; Table 1).

Fig. 2

The incidence of HIT after TKA and RTKA from 2010–2019. HIT: heparin-induced thrombocytopenia; TKA: total knee arthroplasty; RTKA: total knee arthropl

Table 1 Incidence of HIT after total knee arthroplasty and revision knee arthroplasty from 2010 to 2019

Demographic characteristics

In the TKA group, patients with HIT were older on average than those in the control group (68.04 vs. 66.29 years, P = 0.04), and a greater proportion were aged ≥ 71 years (46.71% vs. 34.45%, P = 0.03). Results of the multivariate analysis suggested that age ≥ 61 years was associated with lower odds of HIT (OR = 0.56, 95% CI 0.37–0.85, P = 0.01),This counterintuitive result should not be interpreted as a true protective effect. It is more likely explained by differences in anticoagulation strategies, monitoring intensity, and coding practices, and therefore warrants cautious interpretation. Sex differences were observed in univariate analysis but were not significant in multivariate analysis (OR = 1.04, 95% CI 0.73–1.49, P = 0.83). In the RTKA group, age, sex, and race were not significantly associated with HIT (Table 2).

Table 2 Patient characteristics and outcomes after TKA and RTKA (2010–2019)

Hospital characteristics

For TKA, discharge against medical advice was significantly associated with HIT (P < 0.01; OR = 15.32, 95% CI 1.80–130.46), though the wide confidence interval suggests that the estimate is imprecise and should be considered with caution. Conversely, treatment in a teaching hospital was associated with a lower incidence of HIT (P < 0.01, OR = 0.62, 95% CI 0.42–0.90) (Fig. 3a, Table 3a). In RTKA, admission to hospitals in the southern region was associated with a higher incidence of HIT (P < 0.01, OR = 5.44, 95% CI 1.60–18.56). No significant differences were found by hospital location (urban vs. rural), teaching status, or discharge disposition (Fig. 3b, Table 3b).

Fig. 3

a. Forest plot of adjusted odds ratios for risk factors associated with HIT after total knee arthroplasty (2010–2019) b. Forest plot of adjusted odds ratios for risk factors associated with HIT after revision total knee arthroplasty (2010–2019)

Table 3 Risk factors associated with HIT in patients undergoing TKA and RTKA (2010–2019)

Risk factors for HIT in TKA and RTKA

Logistic regression analysis was used to explore factors associated with HIT. Among all examined comorbidities, pulmonary circulatory disorders showed the strongest association with HIT in the TKA group (OR = 3.42, 95% CI 1.80–6.53, P < 0.01) (Fig. 3a, Table 3a). For RTKA, nonbleeding ulcers were significantly associated with HIT (OR = 19.05, 95% CI 2.06–175.85, P = 0.01); however, the wide confidence interval indicates an imprecise estimate and thus finding should be considered with caution. Pulmonary circulatory disorders (OR = 4.13, 95% CI 1.68–10.17, P < 0.01), tumors (OR = 7.22, 95% CI 2.30–22.62, P < 0.01), and valvular heart disease (OR = 2.50, 95% CI 1.12–5.57, P = 0.02) were also associated with HIT (Fig. 3b, Table 3b). In the TKA cohort, depression was inversely associated with HIT (OR = 0.53, 95% CI 0.29–0.97, P = 0.04). Given potential confounding, coding limitations, and the rarity of events, this association should be regarded as exploratory rather than causal.

Effect of HIT on outcomes

After TKA, the incidence of DVT (12.41% vs. 0.31%, P < 0.01), PE and infarction (3.65% vs. 0.13%, P < 0.01), acute myocardial infarction (1.64% vs. 0.12%, P < 0.01), acute renal failure (12.41% vs. 1.95%, P < 0.01), pneumonia (2.92% vs. 0.34%, P < 0.01), postoperative pain (8.01% vs. 4.11%, P < 0.01), and periprosthetic joint infection (PJI; 2.92% vs. 0.51%, P < 0.01) may increase in patients who develop HIT (Table 4). Following RTKA, the incidence of DVT (14.89% vs. 0.70%, P < 0.01), acute myocardial infarction (2.13% vs. 0.29%, P = 0.02), dyspnea (6.38% vs. 0.34%, P < 0.01), and renal failure (36.17% vs. 6.61%, P < 0.01) were observed at higher rates following HIT diagnosis (Table 4). The mortality (4.45% vs. 0.04%, P < 0.01) and hospitalization costs (112,318.39 vs. 110,552.50, P < 0.01) in the TKA group were significantly greater than those of general patients, and the Charlson index (4.32 vs. 3.74, P < 0.01) and Elixhauser comorbidity index (5.45 vs. −0.09, P < 0.01) were significantly greater (Table 2). The LOS was significantly longer (11.26 vs. 4.23, P < 0.01) and the associated hospitalization costs were markedly greater (169,141.83 vs. 92,266.85, P < 0.01) in the RTKA group than those in standard patients. Additionally, the Charlson comorbidity index (2.31 vs. 1.75, P < 0.01) and Elixhauser comorbidity index (7.70 vs. 4.82, P < 0.01) were significantly high in the RTKA group (Table 2).

Table 4 Complications outcomes of postoperative HIT patients undergoing TKA and RTKA

Discussion

With ongoing updates to antithrombotic treatment guidelines [10], this study builds on previous research [11]to further assess the risk and prognosis of HIT in patients who underwent TKA between 2010 and 2019, with a specific focus on those who underwent RTKA. According to the ninth edition of the ACCP guidelines [2], the recommendations for HIT treatment and prevention have significantly decreased its incidence among patients with TKA. However, over the past decade, the incidence of HIT in the RTKA group (0.3%) has remained notably higher than that in the TKA group (0.1%), whereas new recommendations have added bivalirudin as a treatment option for HIT and optimized antibody testing protocols. These updates do not fundamentally alter existing strategies for HIT prevention in patients who underwent TKA. These observed differences suggest that HIT risk stratification need to account for the surgical type. This disparity persists despite a broader adoption of newer anticoagulants, as patients who underwent RTKA often require heparin owing to (a) heparin’s reversibility favors its use over DOACs in RTKA with a higher bleeding risk, (b) the need for bridging therapy in patients with prior mechanical complications, and (c) surgeon preferences in complex reconstructions. The identification of RTKA-specific risk factors provides actionable targets for preoperative optimization. Although both procedures can have severe adverse outcomes with concurrent HIT, research on HIT risk factors in RTKAs is limited. Our study seeks to fill that gap.

The observed inverse association between age ≥ 71 years and HIT risk is unlikely to reflect a genuine biological protective effect. Several mechanisms may account for this finding. Older or frailer patients are more often managed with non-heparin alternatives such as aspirin or direct oral anticoagulants, or with lower-dose and shorter-duration heparin regimens, thereby reducing effective exposure. In addition, elderly patients frequently undergo surgery at teaching hospitals or within enhanced recovery pathways, which provide more intensive monitoring and individualized perioperative care [12]. Such practice differences may reduce the likelihood of HIT being triggered or recognized. Furthermore, reliance on ICD coding may bias results: in older patients, postoperative thrombocytopenia is more often attributed to infection, bone marrow suppression, or other comorbidities rather than formally coded as HIT, leading to potential under-ascertainment [13]. Finally, given the very low event rate of HIT, regression estimates are vulnerable to sparse-data bias, which can produce artifactual inverse associations. Taken together, these considerations suggest that the apparent “protective effect” of advanced age is more likely driven by treatment patterns, diagnostic practices, and residual confounding rather than by age itself. Although age does not correlate with outcomes in patients undergoing RTKA, the RTKA group is, on average, younger than the first-time surgery group (65.04 vs. 68.04 years). In contrast to the long-term risks of bone resorption and prosthetic loosening associated with prolonged prosthesis use, younger patients more often undergo surgical treatment for PJI or periprosthetic fractures [14]. These patients differ from those undergoing primary surgery in that, in addition to the bone defects resulting from the initial procedure, the increased complexity of bone and soft tissue structures complicates the repair process [15, 16]. Furthermore, heparin re-exposure can trigger the antibody activation [17], presenting significant challenges related to HIT during the second surgery. From 2011 to 2019, the observed incidence of HIT declined in TKA and RTKA, which may reflect evolving anticoagulation practices, including the broader adoption of alternative agents such as DOACs and aspirin in some centers. Despite this overall decrease, HIT remains a clinically relevant concern, particularly in revision procedures and among younger patients. This finding highlights the need for standardized, personalized anticoagulation monitoring following RTKA, with special attention to younger patients to prevent overlooked HIT diagnoses.

In this study, discharge against medical advice was positively associated with HIT in the TKA group. However, given the small number of events and wide CIs, this association is significantly unstable and should be interpreted with caution. This finding may be confirmed by studies with larger sample sizes. Teaching hospitals serve as a protective factor for TKA group by offering advanced treatment options, including personalized care and platelet count monitoring. This approach facilitates more precise management of anticoagulant therapy and reduces the incidence of HIT [18, 19]. The prevalence of chronic conditions such as obesity, diabetes, and hypertension is higher in the southern United States than in other regions [20], potentially increasing the risk of bleeding in patients with these comorbidities. Although heparin therapy can effectively control postoperative bleeding, it also poses an increased risk of HIT [21]. Therefore, the risks and benefits of heparin use must be carefully assessed in postoperative care.

In the revision cohort, comorbidities were significantly associated with increased HIT risk. Multivariate analysis revealed that increased number of comorbidities was a risk factor, suggesting that certain conditions, which do not significantly increase the risk during the initial TKA, are associated with HIT following revision surgery. Although some risk factors exhibited particularly high effect sizes, the consistent increase in thrombotic comorbidities across TKA and RTKA groups supports the fundamental pathophysiology of HIT. These conditions include tumors, nonhemorrhagic ulcers, and valvular heart disease. Pulmonary circulation disorders may be risk factors for HIT after two surgical procedures, such as chronic obstructive pulmonary disease (COPD) [22]. COPD has been identified as a risk factor for HIT following cardiac surgery [23]. Conversely, in pulmonary arterial hypertension, heparin therapy can improve hemodynamics, lower pulmonary artery pressure, and alleviate patient symptoms [24]. Even if heparin is discontinued preoperatively, prior exposure may still lead to antibody formation, increasing the likelihood of postoperative HIT [25]. A retrospective study revealed that adult patients receiving continuous IV heparin before cardiac surgery had a tenfold higher risk of postoperative HIT [26]. Therefore, patients with pulmonary circulation disorders should undertake additional measures to reduce the risk of postoperative HIT. The complexity of RTKA is usually greater than that of TKA, particularly when dealing addressing bone defects and joint instability. RTKA may require the use of bone grafts or specific repair components to address joint failure caused by infection or other complications. In addition, RTKA typically results in higher rates of re-repair and longer recovery periods, making postoperative management more complex and increasing complication risks. This study indicates that valvular heart disease is a potential risk factor in RTKA. Such patients are more likely to have had heparin exposure during valve replacement surgery, cardiopulmonary bypass, and postoperative unfractionated heparin therapy. In patients with cancer, platelet counts may not accurately reflect HIT because of factors such as chemotherapy effects, bone marrow infiltration, myeloproliferative neoplasms, disseminated intravascular coagulation, and thrombotic microvascular disease. Moreover, LMWH is a primary anticoagulant in patients with cancer, and warfarin is a poor choice in this situation, making HIT management more challenging. At a comprehensive cancer center, the annual incidence of HIT was 0.24%, with approximately 0.57 HIT cases occurring per 1000 patients with cancer using heparin [27]. The link between cancer and HIT needs further investigation. This study also revealed an association between nonbleeding ulcers and HIT in the RTKA group; however, the number of events was extremely small and the confidence interval was wide. Thus, this finding is significantly unstable and should be interpreted with caution. Such ulcers are often related to diabetic foot complications or prior warfarin therapy. Chronic ulcers can create a hypercoagulable state, and bacterial biofilms can hinder healing, highlighting the importance of infection control [28,29,30]. The inverse association between depression and HIT is unlikely to reflect a true biological protective effect. More plausible explanations involve differences in clinical management. Patients with depression frequently use selective serotonin reuptake inhibitors (SSRIs) or serotonin–norepinephrine reuptake inhibitors (SNRIs), which are associated with increased bleeding risk. This may lead clinicians to adopt more cautious anticoagulation strategies, including avoidance of heparin or closer monitoring when heparin is administered [31]. In addition, depressed patients often receive multidisciplinary care and more frequent medical follow-up, which may indirectly alter perioperative anticoagulation practices [32]. Finally, because HIT is a rare event, regression estimates for small subgroups are vulnerable to instability and sparse-data bias. Taken together, these factors suggest that the apparent protective association is more likely due to clinical practice patterns and residual confounding rather than a causal effect of depression.

Regardless of surgery type, thrombosis-related diseases after HIT are expected to increase, resulting in higher hospitalization costs, prolonged hospital stays, and increased mortality [8, 33]. TKA and RTKA have notably similar postoperative complications. Following TKA surgery, HIT can develop, prompting clinicians to restrict patient activity and initiate anticoagulation therapy, which may increase the risk of pneumonia. A study investigating the effect of postoperative physical activity on the incidence of pneumonia in patients with esophageal cancer revealed that the average number of steps taken between postoperative days 8–10 serves as a critical predictor for postoperative pneumonia development. Particularly, patients who recorded < 1,494 steps per day were more likely to develop postoperative pneumonia, underscoring the significant influence of early postoperative physical activity levels on pneumonia occurrence [34]. Although RTKA does not increase the risk of PE, it may increase the likelihood of respiratory conditions such as dyspnea. TKA increases the risk of infection and postoperative pain; thus, clinicians should be prepared to ensure adequate monitoring and timely detection of HIT.

The analysis highlights several considerations for clinical practice. Patients who ultimately require revision procedures, particularly those with pulmonary circulatory disorders or valvular heart disease, appear to be more vulnerable to HIT. In these high-risk settings, closer postoperative surveillance—including routine platelet monitoring and timely adjustment of anticoagulation regimens—may help reduce the likelihood of delayed recognition and related complications. Equally important, efforts to optimize primary TKA and minimize the need for revision surgery may represent a critical strategy to lower the overall incidence of HIT and associated adverse outcomes.

This study has several limitations, including potential errors and biases inherent in the NIS database that could affect the accuracy of the results. This study uses ICD coding to identify the following potential biases in HIT: (1) Heterogeneity of diagnostic criteria: different institutions may use different diagnostic criteria (e.g., 4 T score, PF4 antibody detection, or functional testing), which may lead to diagnostic error classification bias. However, institutionally validated ICD-10 codes can be integrated with other data sources to improve accuracy in identifying postoperative complications [35]. (2) There may be false positives all together: non-immune HIT (HIT type I). (3) False negative inclusion: some cases with serological confirmation but without ICD documentation may be missed owing to time window limitations, such as some HIT occurring 5–10 days after exposure to heparin [36]. Given that NIS does not contain outpatient data, we cannot track the two surgeries. Moreover, this study represents an initial exploratory analysis of retrospectively collected data, which primarily reveals associations rather than causal relationships. Thus, further prospective studies with more granular clinical data are warranted to confirm the present findings and clarify potential causal mechanisms. Despite these limitations, NIS data can still effectively reflect the epidemiological characteristics of HIT confirmed during hospitalization, providing important insights for hospital quality assessment and acute management.

Conclusions

Certain preoperative comorbidities and baseline characteristics were associated with the occurrence of HIT after TKA. RTKA was additionally linked to less favorable clinical outcomes. These observations underscore the importance of early identification of high‑risk patients and may help guide strategies aimed at reducing the likelihood of revision procedures and related complications.