Nodular thyroid disease is very common in clinical practice and its diagnosis has increased with the performance of imaging tests, such as ultrasound [1]. Thyroid Fine-Needle Aspiration (FNA) is the most cost-effective and diagnostic method for thyroid nodules, with high sensitivity and specificity (76-98% and 71-100%, respectively) [1,2]. FNA can be performed by thyroid palpation or, more often, guided by ultrasound. Studies have shown that ultrasound-guided FNA (US-FNA) allows a lower rate of false positives and fewer non-diagnostic samples, comparing to FNA by palpation [1,3]. However, the effectiveness of US-FNA still depends on the experience of the operator [1]. Although rare, the most frequent complications of US-FNA reported are pain or cervical discomfort caused by the procedure itself, and bleeding, in some cases with formation of haematoma in the subsequent period [3,4]. Bleeding results from venous extravasation of blood into or out of the nodule. In most cases is easily managed through compression and local cryotherapy. This haemorrhagic complication is more frequent in highly vascularized thyroid glands or in patients with haemorrhagic risk factors, such as conditions that affect coagulation cascade (hypothyroidism, hepatic cirrhosis, chronic renal disease, among others) or under anticoagulation/antiplatelet therapies [4]. Thyroid dysfunction modifies the physiological balance between coagulation and fibrinolysis. Thyroid hormone deficiency induces a tendency towards hypo-coagulation by reducing coagulation factors such as von Willebrand factor and VIII, IX and XI factors. Thus, patients with uncontrolled hypothyroidism who perform thyroid FNA have a higher risk of haemorrhagic complications [5‐7]. In addition to previously mentioned factors, haemorrhagic complications could be inherent to the technique itself, such as use of larger diameter needles, an abrupt reduction of intra-nodular pressure during aspiration of nodules with a cystic component or even lack operator experience [4].

      In recent years, not only has been an increase in the number of US-FNA performed, but also, a greater complexity of the patients with respect to associated comorbidities and therapeutics. There is currently an increase in number of patients referred to thyroid US-FNA who are under anticoagulants/antiplatelet therapy. These drugs are mostly used to prevent thromboembolic events. The anticoagulants currently available on the portuguese market include: coumarins such as vitamin K antagonists and the direct oral anticoagulant (DOAC) which include direct inhibitors of thrombin and factor Xa. Antiplatelet medication includes Irreversible inhibitors of cyclooxygenase (COX1), Adenosine Di Phosphatase (ADP) receptor inhibitors and Glycoprotein IIb/IIIa (GPIIb/IIIa) inhibitors. Table 1 summarizes these different pharmacological classes available in Portugal. Since bleeding is a possible complication of thyroid US-FNA, performing this procedure in patients under anticoagulants/antiplatelet therapy is a challenge for both the prescribing professional and the practitioner. The danger of suspending treatment and thus increasing the risk of thrombotic events [8] and on the other hand, the risk of haemorrhagic complications due to the trauma caused by the procedure in hypo-coagulated patients, creates a dilemma for clinicians [9,10].

Table 1

Table 1: Anticoagulants and antiplatelet available in Portugal

Table 2

Table 2: Recommendations on manipulation of anticoagulant and antiplatelet drugs prior to minor bleeding risk procedures

Table 3

Table 3: Recommendations on manipulation of anticoagulant and antiplatelet drugs prior to minor bleeding risk procedures

Table 4

Table 4: Summary of results

      There is no consensus in national and international recommendations regarding the need to suspend anticoagulant/antiplatelet therapy prior to thyroid US-FNA (Table 2). There are few recommendations regarding thyroid cytology specifically, being more frequent protocols of action directed to invasive and surgical procedures. However, taking into account the characteristics of the thyroid FNA, we can include this procedure in the minor haemorrhagic risk procedure group. The purpose of these types of guidelines is to balance between the risk of thromboembolic events (in case of treatment interruption) and the haemorrhagic risk during the procedure (in case of keeping therapy). Regarding Cardiovascular and Interventional Radiological Society of Europe (CIRSE) guidelines, if INR ≤2 vitamin K inhibitors or Acetyl Salicylic Acid (ASA) monotherapy should be maintained [11]. However, clopidogrel monotherapy should be stopped before performing FNA. Nowadays the use of DOAC has been overlapping vitamin K antagonists, allowing a more simplified approach. Based on the protocol developed by Portuguese Anaesthesia Society, thyroid US-FNA could be included in the group of invasive procedures that do not require suspension of anticoagulant therapy [12]. It recommends INR value < 3 in the case of vitamin K inhibitors therapy and in the case of DOAC, to stop taking the day of the procedure and resume administration six to eight hours after the procedure with half of the usual dose. Regarding antiplatelet therapy this same consensus recommends suspension 3-5 days earlier if doing monotherapy in primary prevention, or maintenance the treatment if monotherapy in secondary prevention or in case of double anti-aggregation (and if both low haemorrhagic and thrombotic risk). Recently European Society of Cardiology in the latest DOAC management guidelines and double antiplatelet therapy update, reinforces the maintenance of these drugs when performing minor bleeding risk procedures, such as US-FNA [13,14]. Specifically to thyroid US-FNA, Korean Society of Thyroid Radiology present more preventive recommendations advising to discontinue vitamin K antagonists 5 days prior the procedure and performing US-FNA only if INR < 1.7 [3]. However, if thrombotic risk so requires, anticoagulation should be replaced by subcutaneous low molecular weight heparin within 2 days prior the procedure. In patients receiving antiplatelet therapy (ASA or clopidogrel), this therapy should be discontinued within 3-5 days prior, until 3 days after US-FNA, unless contraindication for discontinuation required by attending physician. There are few studies evaluating the safety of thyroid FNA in anticoagulated/antiplatelet patients in the literature [9,10,15]. Studies with FNA in breast and prostate tissue demonstrate that these drugs can be safely maintained [16‐18].

      It is important to highlight possible consequence of bleeding in the acquisition of cellular material suitable for pathological anatomy analysis. According to Bethesda classification for thyroid US-FNA, a non-diagnostic result is defined as a sample with insufficient number of thyroid cells associated with a cystic or haematic fluid component [1]. In these cases, it is recommended to repeat the cytological procedure. There are insufficient studies establishing relationship between FNA in hypo-coagulated patients and non-diagnostic samples, although published studies do not demonstrate causal link between the two variables [15].

      With this study, it is our purpose to verify if performing thyroid nodules US-FNA in patients under anticoagulant/antiplatelet medication, increases risk of haemorrhagic complications or on-diagnostic results.

      This is a retrospective study of patients undergoing thyroid US-FNA in Radiology department of Portuguese Armed Forces Hospital-Lisbon (HFAR-PL) between January 1st of 2017 and June 30th of 2018. All FNA were guided by ultrasound and performed systematically by the same radiologist, with the same device (ACUSO-S2000) and analyzed by the same pathologist. A total of 223 clinical processes were analyzed and telephone contact was made with all patients to confirm information from the clinical process. Variables assessed were age, sex, anticoagulant or antiplatelet therapy, presence of complication (pain, hematoma or other) and US-FNA result. Exclusion criteria were: FNA performed outside radiology department of HFAR-PL or by other professional and absence of available cytological results in clinical process. The final sample includes a total of 278 US-FNA. None of patients stopped anticoagulation/antiplatelet medication before the procedure. Patients were divided into two groups classified as: Hypo-coagulated group if under anticoagulant/antiplatelet therapy (including: warfarin, clopidogrel, Acetyl Salicylic Acid (ASA), ticlopidine, triflusal and Direct Oral Anti Coagulants (DOAC): Rivaroxaban, dabigatran and edoxaban); or Control group included patients without this kind of medication. In addition to consulting the clinical process, anticoagulant/antiplatelet medication as well as the report of complications after procedure were confirmed with each patient. Statistical analysis was performed using SPSS with significance index p<0,05. All values of p value were calculated by chi-square test.

      We analyzed 278 thyroid US-FNA from 223 patients. 55.6% of thyroid US-FNA were performed in female patients and 44.4% of males, aged between 22 and 89 years (mean age of 66.8 years). Patient’s characteristics are shown in Table 3. Control group comprised 73.4% (n=204) of total US-FNA performed and hypo-coagulated group included 26.6% (n=74). In the Hypocoagulated group: 67.5% of US-FNA were performed in patients receiving antiplatelet therapy (20% under ASA, 1.35% with dual ASA and clopidogrel aggregation, 41.9% with clopidogrel, 2.7% with triflusal and 1,35% with ticlopidine), 25.6% in patients taking DOAC (14.9% under rivaroxaban, 9.5% under dabigatran and 1.35% under edoxaban), and 6.7% in patients under vitamin K antagonist (warfarin). Complications (haemorrhagic or not) were reported in 2.5% (n=7) US-FNA performed: 1.96% for the control group and 4.1% for the hypo-coagulated group (p value = 0.32). Control group did not present any registry of haemorrhagic complications. In the hypo-coagulated group, haemorrhagic complications were reported in only one US-FNA procedure (1.35%). It was reported as a small hematoma formation without need for medical evaluation or treatment. This was a patient receiving rivaroxaban. Haemorrhage was not statistically significant between control and hypo-coagulated groups (p value = 0.11). Analyzing specifically the group of patients under DACO (n=19) in comparison to control group (n=204) there is a difference with statistical significance (p<0.001). However, this result appears to be biased in view of the size of the compared samples and should not therefore be extrapolated. Non-haemorrhagic complications, such as pain or cervical discomfort up to 48 hours after procedure, were found in 2.2% of US-FNAs: 1.96% in control group and 2.7% in hypo-coagulated group (patients on antiplatelet therapy with ASA and clopidogrel), with no statistical difference between the 2 groups (p value = 0.71). Non-diagnostic samples (Bethesda System) were recorded in 2.5% of all US-FNA, 1.47%(n=3) in control group and 5.4% (n=4) in hypo-coagulated group. In the latter group, all non-diagnostic samples were from patients under DOAC, specifically rivaroxaban (50%/n=2) and dabigatran (50%/n=2). The difference between the groups was not statistically significant (p value = 0.07). Results are resumed in table 4.

      Comparing the two groups there were no statistically difference in number of haemorrhagic complications. The incidence of haematoma in our study was approximately 1% similar result to other previously published studies [4,9,10]. In 2011, Abu-Yousef and his colleagues [9] published a study with 593 patients under anticoagulation/antiplatelet medication undergoing thyroid FNA, whose results showed absence of statistical difference between the groups, with a total hematoma formation rate of 1%. This review questioned the real need for suspension of anticoagulant medication for thyroid FNA and the need for protocols on the topic, a suggestion that created interest and controversy, leading to several reviews about the subject. With the introduction of DOAC was also a need to update recommendations of approach in FNA procedures. Thus, Lyle M. and colleagues elaborated a review about the approach of these new drugs in patients undergoing minor invasive procedures (including thyroid FNA in this category), stating as reasonably the conservation of treatment during this kind of procedure [19]. For this conclusion they have taken into account the half-life and safety profile of these drugs. In our study, in the hypo-coagulated group about 26% of US-FNA corresponded to patients under DOAC. In this universe only one patient (1.35%) had a small haematoma formation on the day after the procedure without need of medical evaluation or treatment compared to none (0%) in control group. There was no statistically significant difference between the groups (p = 0.11).

      Recently, H. Khadra and his collaborators [10] have published a series of 802 patients undergoing thyroid US-FNA divided in 2 groups (control vs. hypo-coagulated). They achieved a rate of haemorrhagic complications of 0.89% in the hypo-coagulated group, with no significant difference for the control group. These results are overlapping with those found in our series (1.35% vs 0.89%). In our study haemorrhagic complications were only found in patients under DOAC anticoagulation. Analyzing this subgroup comparison with control group, the difference was statistically significant (p<0.001), which could be explained by the difference in the size of both groups. Therefore, it seems improper to extrapolate these results. In our opinion this subject and in particular this therapy subgroup, need more robust studies in order to obtain more concrete results.

      In the design of this study it was also our aim to evaluate if there would be an increase in non-diagnostic cytological results (according to the Bethesda classification) in hypo-coagulated group versus control group. Literature on this subject includes mainly patients treated with warfarin, clopidogrel and ASA, with reference only to one study that included patients under DOAC. They present non-diagnostic results incidence between 3.9% [15] and 6% (10). Our results were similar, with the number of non-diagnostic samples being 1.47% in the control group and 5.4% in the hypo-coagulated group (all in patients under DOAC). None of the patients on conventional antiplatelet agents or anticoagulants presented this type of smears. This difference being no statistically significant (p = 0.07) confirms the non-interference of the hypo-coagulant state induced by these drugs and the quality of the acquired sample.

      As limitations of our work, we point out the fact that this is a retrospective study and in presence of a substantial variety of antiplatelet/anticoagulants, it is difficult to have a representative number of patients under each drug/class, which may influence the results obtained. We conclude that thyroid ultrasound guided fine-needle aspiration is widely used and considered as a method of choice for the evaluation and diagnostic approach of thyroid nodules. The most frequently reported complications are local pain and cervical discomfort and bleeding, although rare and usually without severity. In view of our results and the published literature, haemorrhagic risk is not significantly increased in patients receiving antiplatelet or anticoagulant agents. Despite we consider it essential to carry out more robust and targeted studies in this area, in our opinion, it seems reasonable the recommendation of non-cessation of anticoagulant/antiplatelet treatment prior to thyroid US-FNA.

      We would like to acknowledge to Luís Tatá (radiology department-HFAR director’s), Luísa Figueiredo (radiology departmentHFAR-PL) and Saudade André (pathology department- HFAR-PL) for their contribution in this study.