Upper Airway Hematoma Secondary to Warfarin Therapy: A Systematic Review of Reported Cases.

Upper airway hematoma (UAH) is a rare but life-threatening complication of oral anticoagulants requiring early recognition. However, no consensus exists regarding the best approach to treatment. We therefore, sought to systematically review the published literature on UAH to elaborate its demographic and clinical characteristics, treatment, complications, and outcomes. A systematic electronic search of PubMed and EMBASE for case reports, case series, and related articles of UAH related to warfarin published from inception (November 1950) to March 2015 was carried out. Categorical variables were expressed as percentage and continuous variables as mean ± standard deviation (SD). Statistical analysis was done using Statistical Package for the Social Sciences (SPSS) version 20.0. All cases were reported to have UAH as a complication of anticoagulation therapy with warfarin. Demographic and clinical characteristics, treatment, complications and outcomes of UAH were studied. Thirty-eight cases of UAH were identified from 34 reports in the literature. No gender preponderance (male = 52.78%) was seen and the average age of presentation was 60.11 ± 12.50 years. Dysphagia, sore throat, and neck swelling were the most common symptoms and the mean international normalized ratio (INR)at presentation was 8.07 ± 4.04. Most cases had sublingual hematoma (66.57%) followed by retropharyngeal hematoma (27.03%). Of the cases, 48.65% were managed conservatively while the rest underwent either cricothyrotomy or intubation with the time to resolution being 7.69 ± 5.44 days. UAH is a rare butpotentially serious complication of warfarin therapy. It is more common in the elderly population with supratherapeutic INR; inciting events were present in many cases. Overall, it has a good prognosis with significant morbidity present only if concomitant respiratory compromise is present. Reversal of anticoagulation with low threshold for artificial airway placement in the event of airway compromise leads to a favorable outcome in most cases.

Introduction

Upper airway hematoma (UAH) secondary to warfarin therapy is rare but potentially life-threatening conditions. Hematomas/bleeding at various sites including sublingual, retropharyngeal, submaxillary and the epiglottis have been described. Although a sublingual hematoma can be confused with infectious processes such as Ludwig's angina, it is frequently obvious on examination. However, the other hematomas described can have more subtle signs until they lead to airway compromise.[1] UAH may occur due to different inciting events including cervical spinal injury, rheumatoid arthritis, neck surgery, injury to great vessels, and violent head movements. It may also occur spontaneously in patients on anticoagulation therapy or with a bleeding diathesis.[23] Although rare, UAH is a very serious event but only case reports exist with no higher level of evidence. Hence, with no consensus in place, the diagnosis and management of this condition remains a challenge. We therefore, sought to systematically review the published literature on UAH to elaborate its demographic and clinical characteristics, treatment, complications, and outcomes.

Materials and Methods

Ethical considerations

As the study did not involve human subjects or hospital chart review, institutional review board (IRB) approval/exemption was not required.

Search strategy and data collection

A systematic electronic search of Medline and EMBASE for case reports, case series, abstracts, and related articles of UAH secondary to warfarin therapy published from inception to March 2015 was performed independently by three authors (PK, RP, and PS) using two broad themes. For upper airway hematoma, the search terms used were “sublingual hematoma,” “sublingual hemorrhage,” “sublingual bleeding,” “epiglottic hematoma,” “supraglottic hematoma,” “neck hematoma,” “submaxillary hematoma,” and “retropharyngeal hematoma.” For the theme warfarin therapy, the search terms used were “warfarin” (mesh), “warfarine,” “warfarin sodium,” “warfant,” “warfarin,” “warfarin potassium,” and “Tedicumar.” The search was limited to human studies. Bibliographies of the reviewed articles were further scanned to identify additional reports. Care was taken to avoid duplication. Thirty-four articles with 38 cases in the English language were identified. The details of the selection process are outlined in Figure 1. The demographic variables, clinical presentations, diagnostic modalities, treatment, outcomes, and complications of UAH were studied. Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) Statistics for Windows, version 20.0 (IBM Corporation, Armonk, NY, USA). A P value of <0.05 was considered to be statistically significant. Categorical variables were expressed as percentages and continuous variables as mean ± standard deviation (SD). Binomial logistic regression analysis was used to identify independent predictors of airway compromise and independent t-test was used to compare the means between the subgroups.

Figure 1

Flowchart describing a systematic search and study selection process

Results

Demographics and clinical presentation

Among the 38 cases of UAH identified, there was no significant gender variation (male vs female, 52.78% vs 47.22%, P = 0.446). The mean age of presentation was 60.11 ± 12.50 years. A majority of the hematomas were sublingual (n = 25, 66.57%) and retropharyngeal (n = 10, 27.03%), with supraglottic, laryngeal, lingual, and submandibular ones being less common [Table 1]. The most common predisposing factors noted were: Violent coughing (n = 6), drug interaction with warfarin (n = 3), airway manipulation during intubation (n = 2), trauma (n = 2), and denture use (n = 2). There we no apparent predisposing factors in the rest of the cases (n = 16, 43.24%). The most common presenting symptom associated with UAH were sore throat (n = 29, 78.38%), dysphagia (n = 24, 64.86%), and neck swelling (n = 22, 59.46%). Respiratory compromise was seen in 18 cases (46.65%). Other less common presenting features were as follows: ecchymosis (n = 14, 37.84%) hoarseness (n = 11, 29.73%), dysphonia (n = 9, 24.32%), bleeding from another site (n = 9, 24.32%), drooling (n = 5, 13.51%), and trismus (n = 3, 8.11%). There was no difference in the mean age (63.35 vs 56.50, P = 0.108), warfarin dose (4.31 mg vs 6.31mg, P = 0.119), time to resolution (6.71days vs 8.60 days, P = 0.417), and international normalized ratio (INR) (8.95 vs 7.12, P = 0.357) between patients with and without respiratory compromise. On binomial logistic regression, there were no significant predictors of respiratory compromise (age, P = 0.112; INR, P = 0.396).

Table 1

Demographic and clinical features of the included cases

Warfarin dose and indications

The mean dose of warfarin taken by the patients was 5.34 ± 2.58 mg. The indications for warfarin were as follows: Mechanical valves (MVs) (n = 11, 29.73%),[24567891011] atrial fibrillation (AF) (n = 10, 27.03%),[3121314151617181920] venous thromboembolism (VTE) (n = 9, 24.32%),[12122232425262728] chronic thrombophlebitis (n = 2, 5.41%),[2930] and polycythemia vera (n = 1, 2.70%).[31] The indication of anticoagulation was unclear in four cases.[8223233] Three cases received concomitant antiplatelet therapy[163233] [Table 2].

Table 2

Laboratory parameters, management, and outcomes of the included cases

Laboratory parameters

The laboratory values at the time of presentation were: INR 8.07 ± 4.04, hemoglobin (Hb) 12.35 ± 2.04 g/dL, and platelets were within the normal range except for one case where thrombocytosis was noted.[31]

Treatment and prognosis

Conservative management with vitamin K, fresh frozen plasma (FFP), or prothrombin complex concentrate (PCC) was instituted in 18 (48.65%)[236891112131415171819202732] cases Table 2. Cricothryrotomy was performed in two (5.41%) cases,[2123] tracheotomy in 12 (32.43%) cases,[145112122232425263133] endotracheal intubation in five (13.51%) cases,[14102228] nasotracheal intubation in four (10.81%) cases,[571629] and surgical evacuation of hematoma in two (5.41%) cases.[47]

Most of the cases had a good outcome with the mean duration to resolution being 7.69 ± 5.44 days. The complications included respiratory compromise in 18 cases (48.65%), pulmonary edema,[21] aspiration pneumonia,[22] and mild pneumonitis[23] in one case each. One patient died due to anoxic brain injury.[11] There was no significant difference between the mean age (58.94 vs 61.17, P = 0.607), warfarin dose (6.03 mg vs 4.67 mg, P = 0.282), time to resolution (8.20 vs 6.53, P = 0.412), and INR (6.59 vs 8.86, P = 0.194) between the patients undergoing conservative and nonconservative management. Similarly, there was no significant correlation between time to resolution and age (Pearson's r = 0.044, P = 0.832) or INR (Pearson's r = −0.427, P = 0.113).

Most cases did not mention whether warfarin was restarted or not during the follow-up period. It was switched to aspirin in five cases[1718222426] and restarted immediately after resolution of the symptoms in three cases;[81127] however, the dose at which it was restarted was not clear.

Discussion

Background

In spite of the increasing popularity of the newer anticoagulants, warfarin remains the most commonly prescribed oral anticoagulant in the United States with >25 million warfarin prescriptions in the United States in 2010.[34] Warfarin acts as a vitamin K antagonist by binding with the vitamin K 2, 3-epoxide reductase in the hepatic microsome and blocking the action of vitamin K-dependent factors II, VII, IX, X, protein C, and protein S. It is commonly used for chronic anticoagulation in patients with atrial fibrillation (AF), venous thromboembolism (VTE), and artificial heart valves. Warfarin levels are monitored with regular INR with a target of 2 to 3 in AF and VTE and 2.5 to 3.5 in patients with mechanical heart valves. Interaction with commonly used medications including broad spectrum antibiotics, quinidine, salicylate, and thyroxine as well as with alcohol and diet often make anticoagulation with warfarin challenging.[331] The concurrent usage of platelet inhibiting agents such as aspirin and nonsteroidal anti-inflammatories further increases the risk of bleeding.[3] The risk of bleeding, internal or external, is related to INR in a log linear fashion[3] and is known to be higher with INR levels >4.5,[4] which is consistent with the mean INR of 8 in our study. The incidence of bleeding in patients on warfarin is about 6.8%.[3] Upper airway hematoma is a rare complication. Sublingual, retropharyngeal, submaxillary, and epiglottic hematomas or bleeding have been described. It is important to recognize these early as they can lead to life-threatening complications such as airway compromise.

Clinical features and diagnosis

UAH may be preceded by predisposing factors such as violent coughing and trauma or it may be spontaneous (43.24% of cases in our study). Sore throat, dysphagia, and neck swelling are the most common presenting symptoms of UAH, which were consistent with our study. These are nonspecific findings that may be associated with many common clinical syndromes such as acute respiratory tract infections. Additionally, since our study did not find any difference in age, INR, or dose of warfarin between patients with and without respiratory compromise, a high level of suspicion is required for diagnosis at an earlier stage. This may be one of the reasons why more than half of the patients have respiratory compromise at presentation. Unlike what has been reported previously,[23] our study found that sublingual space (66.57%) was most commonly involved followed by retropharyngeal space (27.03%). Differentiation from acute infectious process such as Ludwig's angina or retropharyngeal abscess is crucial as they are managed quite differently.

Management

No definite consensus to treatment exists in the literature. While Hefer et al.[35] reported similar outcomes among all patients with retropharyngeal hematoma up to 1993 with observation versus aggressive early airway management, Cohen and Warman[22] support early tracheotomy in all patients, with observation limited to only mild cases. Similarly, Rosenbaum[11] recommends close intensive care unit (ICU) monitoring. The data of our review was in line with Hefer et al.[35] in that the outcomes (time to resolution) did not differ in the conservative and aggressive approach. Moreover, advanced age and higher INR failed to predict the likelihood of respiratory compromise. Hence, in mild cases with no airway compromise, our study favors medical therapy with reversal of the coagulopathy with vitamin K and FFP or PCC preferably in an ICU setting.[11318] The recommended dose of FFP and PCC is 4 units/kg with INR greater than 1.5 and 50 units/kg with INR greater than 6, respectively.[13] Although surgical drainage has been described,[36] it is not warranted in most cases as it carries the risk of increasing soft tissue edema and airway compromise.[37] Spontaneous resolution usually occurs with normalization of coagulation parameters.[1]

Patients with severe airway compromise should be considered a medical emergency and endotracheal intubation may be indicated as life-threatening hemorrhage can occur into the sublingual space rapidly.[1] Patients should be evaluated by an otolaryngologist (or any other physician capable managing and evaluating a critical airway including performing a flexible laryngoscopy) and those with impending airway obstruction should be managed by a team of experienced anesthesiologists and otolaryngologists. The preferred management should be fiber-optically-guided nasotracheal intubation; cricothyrotomy or awake tracheotomy should be done only in cases where intubation is not possible and orotracheal intubation contraindicated as mask ventilation may be impossible.[11]

Prophylactic antibiotics were used in six cases; however, they are not usually indicated as abscess formation does not occur.[1] Although steroids were used in 10 cases, there was no definitive evidence of benefit. The patient may be restarted on warfarin with regular monitoring of INR once the hematoma resolves, provided that he/she is able to maintain an optimum level of INR. None of the reported cases had INR in the desired range; hence, it is unlikely that UAH occurs with a normal INR.

Limitations of the study

Our study had several limitations. This is a retrospective series of reported cases in the current literature and has inherent biases related to such studies including selection and publication biases. Also, we included only articles available in the English language. Due to the small sample size, statistical analysis was limited. Asymptomatic cases of UAH are unlikely to be recognized; hence, the reported cases may not represent the overall patient population with UAH.

Conclusion

UAH is a rare but potentially serious complication of warfarin therapy, which should be differentiated from the more common infectious etiologies. It is more common in the elderly population with supratherapeutic INR and some inciting event present in many cases. Overall, it has a good prognosis with significant morbidity present only if concomitant respiratory compromise is present. Only mild cases should be observed, preferably in an ICU setting and conservative management is possible in these patients but no patient characteristics predict airway compromise or successful conservative management. Reversal of anticoagulation with low threshold for artificial airway placement in the event of airway compromise is the treatment of choice.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.