Incidence of venous thromboembolism following head and neck surgery

Venous thromboembolism (VTE), comprising both deep venous thrombosis (DVT) and pulmonary embolism (PE), is a common postoperative complication, especially among those with cancer among [1]. PE, in particular, is recognized as an important cause of potentially preventable death among postoperative patients [2]. Not to be underestimated, DVT is associated with high morbidity and mortality [3, 4] with potential of severe complications of post-thrombotic syndrome and associated high healthcare cost and quality of life burden [5, 6]. Prevention as well as early detection and treatment VTE can reduce the associated morbidity and mortality [7]. However, in the surgical population, especially, the benefit of pharmacological thromboprophylaxis (i.e., anticoagulation) must be weighed against the risk of bleeding.

The frequency of postoperative VTE in the head and neck surgery population varies widely from 0 to 26% [8‐11]. Accordingly, assessment of the risks and benefits of thromboprophylactic measures for head and neck surgery patients and determination of guideline recommendations have been challenging. A meta-analysis from 2018 including five studies focused on head and neck patients, all with cancer, described a VTE incidence of 5% for patients undergoing resection [12]. Conversely, in another meta-analysis from 2018 including 29 studies studying patients with and without cancer, an overall VTE risk of 1.5% was observed [13]. Stratification by other relevant factors, such as anatomical location and age, has not performed in the majority of such studies. To support the clinical decision-making on thromboprophylaxis, the American College of Chest Physicians (ACCP) published evidence-based guidelines for the prevention of VTE in surgical patients [7]. The guidelines tailor recommendations to individual surgical specialties and also recommend use of the Caprini risk score for selected surgical specialities [14]. Unfortunately, the ACCP guidelines do not include specific recommendations for head and neck surgery. Numerous obstacles hinder guidelines for prevention of VTE in head and neck surgery: broad heterogeneity of procedure type, variable risk among those undergoing surgery for malignant and non-malignant disease, and lack of specific risk stratification tools in the otolaryngology population.

To provide more clarity regarding the heterogeneous risk of otolaryngology patients, we assessed the frequency of VTE in patients undergoing head and neck surgery, stratified according to cancer status, using contemporary data from a nationwide Danish cohort.

This study was an observational cohort study analysing Danish nationwide administrative registry data from year 2010–2018.

The study population comprised 116,963 patients who has undergone head and neck surgery between year 2010 to 2018 (Fig. 1). Baseline characteristics are presented in Table 1. Of these, 12,083 patients (10%) had cancer (47.2% female; mean age 66.2 years), and 104,870 patients (90%) did not have cancer (48.6% female; mean age, 51.2 years).

Patients with cancer primarily had an endoscopic procedure performed (71.2% of the patients with cancer vs. 24.4% of the patients without cancer) whereas patients without cancer primarily had surgery on mouth/throat performed (41.8% vs 22.8%, respectively). A greater proportion of cancer patients had hypertension (33.0% vs. 19.3%) and were recently immobilized (20.1% vs. 5.2%). However, a greater proportion of those without cancer had recent trauma or fracture (2.6% with cancer vs 9.7% without cancer). During follow-up, 13.7% of the patients with cancer died compared with 4.0% of patients without cancer.

A total of 444 patients experienced a VTE event during 3-month follow-up. At 3-month follow-up, VTE rates per 100 person-years were 5.08 for patients with cancer and 1.30 for patients without cancer (Table 2). Figure 2 shows the cumulative incidence proportions of VTE for head and neck patients undergoing surgery and for matched controls. At 3-month follow-up, the absolute VTE risk was 1.2% for patients with cancer 0.3% for those without (Table 2). For the matched controls, a total number of 248 had VTE after 3 months corresponding to a risk of < 0.1%.

During year 2021 on the Department of Otolaryngology, Head and Neck Surgery and Audiology, Aalborg University Hospital, 21 patients out of 1384 hospitalised patients (1.5%) received prophylactic doses of dalteparin or tinzaparin.

In this large nationwide cohort of patients undergoing head and neck surgery, we observed that few patients suffered VTE after head and neck surgery. However, the highest risk of VTE was found among postoperative patients with cancer. Older patients and those undergoing endoscopic head and neck surgery demonstrated higher rates of VTE.

Comparison of studies investigating VTE incidence after head and neck surgery has been challenging due to several epidemiologic differences. First, the follow-up period has varied widely, from VTE evaluated only until discharge to assessment out to 6 weeks after discharge [12, 23]. Second, complexity and length of head and neck surgery can vary greatly [8, 10, 24]. In addition, some studies have chosen to include both symptomatic and asymptomatic VTE while others have focused on only symptomatic events [25]. Some studies include both patients with and without cancers, while others have been confined to patients either with or without malignancy [11, 12, 23]. Statistically, the issue of death as a competing risk for VTE has rarely been considered, e.g. by using the Aalen-Johansen estimator, assuming death as competing risk [26]. Finally, some studies use medical thromboprophylaxis others have relied upon mechanical prophylaxis [27]. These differences from population to methodology have limited the ability to obtain consistent estimates of VTE risk.

A systematic review and meta-analysis including five prospective studies of head and neck cancer patients undergoing tumour resection, found a VTE incidence of 5% [12]. In three of the studies, the follow-up period was not reported, which challenges the interpretation of the estimated 5% risk. In these studies, pharmacological thromboprophylaxis was used routinely only in two of five studies. Hence, the study also concluded that administration of pharmacological thromboprophylaxis was inconsistent in head and neck cancer surgery indicating that current guidelines on thromboprophylaxis in cancer surgery have yet to be implemented in routine clinical practice. In another systematic review and meta-analysis including 23 studies with a total of 618,264 patients including patients with and without cancer, a VTE risk of 0.4% was estimated. Of the included studies, 12 studies evaluated VTE outcomes only until discharge, whereas all the other studies looked at outcomes up to 30 days after discharge, except for a single study evaluating VTE risk up to 6 weeks after discharge [23].

A high VTE risk of 26% was noted 30 days after surgery among 133 patients undergoing oral oncologic surgery with simultaneous reconstruction [9]. All patients had routine postoperative ultrasonography performed to assess for asymptomatic deep venous thrombosis. Of 35 patients with detected VTE, three patients had PE without note of whether these were symptomatic. Of note, surveillance ultrasonography is not routine clinical practice after surgery. Furthermore, the clinical significant and treatment of incidental or asymptomatic VTE is subject to fierce debate [28]. Hence, the clinical implication of this study is uncertain. None of these studies report results stratified on anatomical areas.

However, in the context of these prior studies and our current analysis, it appears that the overall incidence of VTE following head and neck surgery appears lower than that of most other surgical specialties [1]. This has important implications because the major concern regarding pharmacological thromboprophylaxis administration has consistently been bleeding. Bleeding within the neck compartment can be life threatening due to its proximity to the airway. Consequently, the concern of bleeding might contribute to low provider prescription of prophylactic anticoagulation and limited patient adherence to thromboprophylaxis protocols. Furthermore, several studies have showed that the benefit from thromboprophylaxis outweighs the risk of bleeding only in high risk patients [8, 25]. The meta-analysed incidence of bleeding complications was 0.9% [23]. The addition of medical thromboprophylaxis did not result in a significantly lower VTE incidence (odds ratio, 0.86;95% CI 0.48–1.52) but produced a higher risk of bleeding (odds ratio, 3.78; 95% CI 2.20–6.48)).

Some patients undergoing head and neck surgery will undergo re-operation within the following months after a first operation. For anatomical surgery types, the highest risk in our study was found for patients with cancer after an endoscopic procedure (1.3%). For some patients this estimate might reflect a higher risk due to re-operation during follow-up. On the other hand, many patients with cancer may shift to radiation therapy, associated with a relative low VTE risk, after an initial endoscopy instead of receiving more surgery [29]. Either way, an assessment of VTE risk may influence further therapy as well as prophylactic measures. To potentially help inform anticoagulant treatment strategies, future studies may focus on evaluating why VTE occurs and which patients may be specifically at risk of developing VTE.

For surgical patients in general, prognostic factors for VTE include general anaesthesia, prolonged surgical time, postoperative immobilisation, mechanical trauma to venous structures, stasis, and cancer, and reconstructive surgery. Among patients undergoing surgery, surgical duration has been shown to be directly proportional to risk of VTE [30]. For patients with cancer, the 2021 American Society of Hematology guideline recommends pharmacological rather than mechanical thromboprophylaxis for all cancer patients undergoing a surgical procedure at lower bleeding risk [31]. Yet, current practices in VTE prophylaxis for patients with and without cancer vary widely among the otolaryngology community [32, 33]. In a survey of otolaryngologists, 88% indicated that guidelines for the prevention of VTE would be useful [34]. Nevertheless, in a retrospective study evaluating use of thromboprophylaxis in oncologic patients undergoing head and neck surgery, only 568 (56%) out of 1018 received appropriate thromboprophylaxis [8].

Despite efforts in VTE prevention, international guidelines do not provide recommendations specifically for head and neck surgery. The ACCP has made recommendations for some individual surgical specialties, but not for head and neck surgery. For other surgical specialities, the Caprini risk assessment model is recommended [22]. The Caprini score has undergone limited validation in the otolaryngology patient population [25, 35, 36]. In a retrospective validation study of 2016 head and neck patients undergoing surgery, an overall VTE risk of 1.3% after 30 days was observed. The study concluded that only patients with a Caprini score > 8 had a VTE risk higher than 3%. This contrasts with the ACCP recommendations in which only patients with a Caprini score ≥ 3 (3% risk) are recommended to receive medical thromboprophylaxis [7]. Accordingly, in 2022 an adapted model of the Caprini score for otolaryngology patients was published, suggesting a higher score cut-off value (≥ 5) before recommending medical thromboprophylaxis [36]. In our study, patients with cancer aged 60–74 and patients with cancer undergoing an endoscopic procedure exhibited the highest VTE risk (1.3%). As such, the applicability of the Caprini score for patients undergoing head and neck surgery in our cohort may be questionable. To our knowledge, no formal validation study on otolaryngology patients assessing calibration and discriminative power of the Caprini model has been performed [37]. Hence, no international guidelines recommend routine use of the Caprini score for head and neck surgery patients.

Limitations of our study include those typically of large dataset analyses using diagnostic codes. VTE ICD-10 coded records may not reflect an actual VTE event because of inaccurate coding and misclassification. However, our approach minimizes this risk by requiring the VTE diagnosis code to be in combination with relevant imaging examinations ensuring a positive predictive value of 91% for the diagnosis [19]. Yet, given the register-based study type, we were not able to include asymptomatic VTE, and therefore, the true VTE risk is anticipated to be higher [38]. No formal study has been performed on thromboprophylaxis use in Danish otolaryngology patients. Use of thromboprophylaxis during hospitalization in 2021 in Aalborg Department of Otolaryngology was low (1.5%). The department represents the fourth largest otolaryngology department in Denmark. Although not necessarily similar throughout the country, these data underscores that thromboprophylaxis is not routinely prescribed for most head and neck surgery patients.

Some degree of misclassification is expected in the non-cancer group. Our supplementary analysis showed that 5% developed cancer during follow-up. The VTE risk was investigated separately for this group and found lower than the risk for patients with cancer known at baseline.

The ICD codes for VTE do not differentiate in proximal/distal deep vein thrombosis nor in pulmonary embolism location (central/sub-segmental), which may be a limitation of our study. Also, the procedure codes for surgery do not hold any information on the length of the surgery. Finally, the Danish population may not be representative of other populations with greater variability in race and ethnicity, affecting the external validity of our results. The nature of this study was descriptive and causal conclusions cannot be based on the observations.

We followed patients in national registries with prospectively collected data and virtually complete follow-up in a setting with free access to health services, thus largely eliminating selection bias [17].

The use of medical thromboprophylaxis in otolaryngology is an often-disputed clinical topic. Overall, the minority of patients suffer from VTE after head and neck surgery. However, the risk is higher in those patients with cancer. For patients with cancer, risk stratification could be further developed to help identify patients at highest VTE risk undergoing head and neck surgery in whom the incidence of postoperative VTE might warrant more routine thromboprohylaxis. Patients without cancer have a low VTE risk and may have a risk–benefit ratio that argues against routine prophylactic anticoagulation.