Incidence of pulmonary thromboembolism (PTE) according to flight duration. Values shown above the bars are numbers of cases per 1 million passenger arrivals.
Pérez-Rodríguez E, Jiménez D, Díaz G, Pérez-Walton I, Luque M, Guillén C, Mañas E, Yusen RD. Incidence of Air Travel–Related Pulmonary Embolism at the Madrid-Barajas Airport. Arch Intern Med. 2003;163(22):2766-2770. doi:10.1001/archinte.163.22.2766
Prolonged air travel and the associated immobilization are risk factors for venous thromboembolism. The occurrence of pulmonary thromboembolism (PTE) under these circumstances is referred to as economy class syndrome. We assessed the incidence of symptomatic PTE in passengers on long-haul flights arriving at Madrid-Barajas Airport, Madrid, Spain, and the association with the number of flight hours.
We retrospectively reviewed cases of PTE among international travelers arriving at Madrid-Barajas Airport between January 1995 and December 2000. Patients presenting with symptoms of deep venous thrombosis but without symptoms of PTE were excluded. Pulmonary thromboembolism was identified using an algorithm of diagnostic tests. The incidence of PTE and the association with flight duration was assessed.
The average number of passengers per year who arrived at the airport on flights originating abroad in the period analyzed was 6 839 222. Sixteen cases of PTE were detected over the 6-year period. All patients with travel-associated PTE had flight durations of greater than 6 hours. The overall incidence of PTE was 0.39 per 1 million passengers (95% confidence interval [CI], 0.20-0.58). On flights that lasted between 6 and 8 hours, the incidence was 0.25 per 1 million passengers (95% CI, 0-0.75), while on flights longer than 8 hours, the incidence was 1.65 per 1 million passengers (95% CI, 0.81-2.49) (P<.001).
Air travel is a risk factor for PTE, and the incidence of PTE increases with the duration of the air travel. However, the low incidence of PTE among long-distance passengers, similar to that observed in other international airports, does not justify social alarm.
ALTHOUGH PREVIOUSLY described by several authors, the relationship between pulmonary thromboembolism (PTE) and air travel acquired notoriety as a result of the publication by Cruickshank et al1 describing the economy class syndrome (ECS). The reduced space and lack of mobility on long-distance flights could favor the development of venous thrombosis and ECS.
Interest in ECS has peaked as a result of 2 recent publications that include data referring to passengers arriving at Heathrow (London, England)2 and Charles de Gaulle (Paris, France)3 airports. These studies emphasized the incidence of deep venous thrombosis (DVT) and pulmonary embolism in economy class passengers of long-haul flights, creating a certain alarm in the media. However, in a case-control study, Kraaijenhagen et al4 found no association between travel and DVT, although most of the passengers had traveled less than 5 hours.
The Madrid-Barajas Airport, Madrid, Spain, receives a high influx of passengers each year. Ramón y Cajal Hospital is the referral center for this airport. This study assessed the incidence of PTE in international travelers arriving at Madrid-Barajas Airport and referred to Ramón y Cajal Hospital over a 6-year period.
Ramón y Cajal Hospital is the referral center for the Madrid-Barajas Airport, and all patients with signs and/or symptoms suggestive of PTE are sent to the emergency department by the airport medical services. We studied all referrals made from January 1, 1995, to December 31, 2000. Patients who had signs of DVT without signs PTE were excluded. In a retrospective cohort study, we assessed the incidence of PTE on international flights arriving at Madrid-Barajas airport.
Diagnosis of PTE was confirmed by integrating criteria for clinical probability,5,6 scintigraphic probability, and lower extremity venous compression ultrasonography results. All patients had lower extremity venous compression ultrasonography. Helical computed tomography (CT) was performed in patients with abnormal chest radiographs, or with underlying cardiopulmonary pathologic conditions. Pulmonary angiography was performed when the treating physician suspected massive embolism (hypotension, new severe respiratory failure, or right ventricle dysfunction on echocardiography) or when diagnostic uncertainty (intermediate- or high-probability symptoms with intermediate- or low-probability lung perfusion scintigraphy in the absence of DVT on compression ultrasonography7) was present.
Statistical data related to numbers of passengers, origin of flights, and duration of flights for those years included in the study were supplied by the Department of Statistics of Madrid-Barajas Airport. Of the travelers in the study, data were not available regarding connecting flights and layover durations. Data were not available on travelers who did not complete connecting flights because of PTE, travelers who died during flight, and those who died at the airport.
In the patients diagnosed with PTE, data for the following variables were assessed: point of departure; class in which the patient traveled; duration of flight in minutes; age; sex; risk factors (obesity [>20% over ideal weight], varicose veins, neoplasias, orthopedic surgery and other major surgery in the previous 3 months, oral contraceptive use, and history of DVT or PTE8); clinical symptoms and signs of presentation; time of presentation; hemogram; arterial blood gases; and electrocardiogram, chest radiograph, and results of diagnostic tests (perfusion scintigraphy, chest CT scan, venous compression ultrasonography, and pulmonary angiography).
Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS version 10.0; SPSS Inc, Chicago, Ill) software. Dichotomous data with proportions and continuous data were described with mean (SD). The incidence of PTE cases per 1 million passengers was calculated. The incidence of PTE was also calculated within strata of number of flight hours (<6 hours, between 6-8 hours, and >8 hours). Incidences among strata were compared using the Fisher exact test.
From 1995 to 2000, 41 million passengers arrived at Madrid-Barajas Airport on international flights, producing an annual average of 6 839 222. The flights had originated in 77 countries or other geographical areas (Table 1). During the same mean time, 41 travelers were referred from Madrid-Barajas Airport to the emergency department at Ramón y Cajal Hospital. During the study period, 583 episodes of PTE were diagnosed, and 342 of these patients were admitted to the respiratory department.
Sixteen patients diagnosed with PTE came from flights that had arrived at Madrid-Barajas Airport. Of these, there were 12 women (75%) and 4 men (4%). The mean (SD) age was 70.1 (8.7) years, 71.2 years among the women and 66.7 among the men. Of the 16 cases, 9 (56%) had 1 PTE risk factor and 7 (44%) had at least 2 risk factors. The most common risk factors were obesity (n
= 7; 44%), varicose veins (n = 6; 38%), neoplasias (n = 2; 13%), history of DVT (n = 1; 6%), and thrombocytosis (n = 1; 6%). None of the patients had undergone recent surgery, taken oral contraceptives, or experienced previous PTE.
Regarding the initial presentation of symptoms in the 16 patients with PTE, 2 occurred in-flight and 11 occurred after the landing of the plane while still at the airport. In 2 cases, the onset of symptoms could not be precisely determined, and in 1 case, the first manifestations presented 8 weeks after the flight. The most common symptoms were dyspnea (n = 11; 69%), syncope (n
= 10; 63%); pleuritic pain (n = 5; 31%), and fever (n = 1; 6.3%). The chest radiograph most commonly revealed vascular alterations (absence of focal peripheral vascularization) (n = 8; 50%), cardiomegaly (n = 2; 13%), and diaphragmatic elevation (n = 2; 13%). Electrocardiograms were abnormal in most cases (n
= 11; 69%), and the most frequent alterations were S1Q3T3 pattern (n = 6; 38%) and alterations in T waves (n = 2; 13%). Other abnormalities included sinus tachycardia, extrasystole, and bundle branch block. On presentation, most patients (n = 13; 81%) had hypoxemia, and just under half (n = 7; 44%) had respiratory failure (defined as PaO2 <60 mm Hg). For PaO2 and PaCO2, the mean (SD) values were 61.5 (19.4) and 39.5 (5.3) mm Hg, respectively. Clinical suspicion of PTE was high in 11 cases (69%) and intermediate in 5 (31%). Clinical suspicion of PTE was never low.
In 15 of the 16 subjects, lung perfusion scintigraphy showed high-probability results (94%). In the remaining case, clinical probability was high, although lung perfusion scintigraphy showed a low-probability result, and compression ultrasonography of the lower extremities was positive for DVT. In this case, helical CT was performed, which did not reveal pulmonary arterial thrombosis. Six patients (38%) presented clinical signs of DVT in the lower extremities, which was confirmed by ultrasonography in 5 (2 popliteal, 2 femoral, and 1 bilateral iliofemoral in location).
The annual incidence of ECS in our experience was 0.39 cases per 1 million passengers (95% confidence interval [CI], 0.20-0.58). The incidence of symptomatic PTE in surviving patients referred to Ramón y Cajal Hospital in relation to flight hours (Figure 1), expressed in number of cases per 1 million passengers, is 0 if the duration of the flight is less than 6 hours, 0.25 (95% CI, 0-0.75) if the flight is between 6 and 8 hours, and 1.65 (95% CI, 0.81-2.49) if it is longer than 8 hours. Risk of PTE increases significantly if the duration of the flight is more than 8 hours (P<.001).
Air travel appears to be a risk factor for PTE. The incidence of PTE is significantly higher among those passengers flying for prolonged periods compared with those with shorter flights. The strength of the association may have been exaggerated by diagnostic suspicion bias applied to long-distance travelers with chest symptoms, but this seems unlikely to explain all of the observed relations. In this study, all the patients diagnosed with PTE had traveled more than 6 hours and most had traveled more than 8 hours. These data support those of Lapostolle et al,3 who reviewed the cases of PTE that required medical attention on arrival at Charles de Gaulle Airport in Paris during a 7-year period and found a direct relationship between distance traveled and the occurrence of PTE.
Despite these data, overall incidence of PTE in relation with long-distance flight is low. Some authors are of the opinion that the association between long-distance air travel and DVT-PTE could be merely circumstantial.9 Kraaijenhagen et al4 published a letter to the editor describing a comparison of 186 patients with DVT with 602 controls with similar symptoms who did not have DVT. In searching for a history of air travel in both cases and controls, no association was found between the trip and the formation of venous thrombosis, although it is true that most of the passengers had traveled less than 5 hours.
It is inevitable that some patients with PTE related to air travel were excluded owing to the study methodology. It is possible that we may have underestimated the incidence of ECS cases. Although our hospital is the mandatory referral center for all patients with signs and/or symptoms suggestive of PTE for the Medical Services of the Madrid-Barajas Airport, it is not possible to account for the number of passengers who died during the flight or at the airport and those who had asymptomatic or mildly symptomatic episodes that did not require medical attention on the plane or at the airport. Also, passengers on long flights with prior connections (especially Asian passengers in our case, who had prior connections in London or Paris) may not have completed the flight plan to Madrid-Barajas Airport because of PTE, death, or other reasons.
The incidence does not justify the social alarm created around air travel caused by the risk of developing PTE. The Madrid-Barajas Airport welcomes an average of 6 839 222 passengers per year originating on international flights, in spite of which only 16 cases of PTE were confirmed during a 6-year study period. This represents an overall incidence of 0.39 cases per 1 million passengers, a rate that coincides with those of other studies carried out in other European airports.3 It is difficult to compare our results with those of other published series, mainly owing to the different methods of selecting cases and the ways in which the results are presented.
Most patients in our study had at least 1 risk factor for PTE, including obesity, varicose veins, history of DVT, and neoplasia, a finding supported by previous studies. Likewise, almost all subjects in our study had comorbid illnesses such as diabetes, high blood pressure, and other cardiovascular diseases. Several authors have concluded that risk of flying is particularly high in these passengers compared with healthy passengers.10- 13 It is impossible, however, to translate these observations into practical advice for people who intend to travel without having conducted a formal examination of the risk factor profile in people without PTE who travel. The initial symptoms were manifest in most of our patients on arriving at the airport, shortly after having landed. Nevertheless, as was previously mentioned, the time interval until symptoms appear can be variable and even reach several weeks.1,3,14- 16
The most common form of clinical presentation was dyspnea or syncope, often associated with hypoxemia. This affirms that PTE associated with long flights is more severe than PTE unrelated to air travel.2,17 Patients who died were excluded from our series. Most of our patients had radiological alterations, mainly vascular alterations, and, in some cases, cardiomegaly or diaphragmatic elevation. Electrocardiographic alterations were evident in most cases. Clinical probability was high in most patients and low in none. Pulmonary thromboembolism was confirmed by means of high-probability lung perfusion scintigraphy in all cases except in one.
Approximately one third of the patients with travel-associated PTE presented with signs of DVT. Of 6 patients with signs of lower extremity DVT, 5 had DVT detected by compression ultrasonography, mainly in proximal regions where there is a greater risk of embolizing compared with the distal regions.2,15,16 Other patients may have had undiagnosed DVT, since ultrasonography has a low sensitivity for detecting DVT in veins below the calves, especially in asymptomatic patients.2 In our study, we did not analyze the patients who only presented signs of DVT without PTE, which for most authors2,15 is much more frequent in relation to air travel than is PTE. Scurr et al,2 in a prospective randomized study of 231 travelers older than 50 years, without a history of thromboembolism, cardiopulmonary disease, or cancer, found that the presence of ultrasonography-documented asymptomatic calf DVT on long-distance flights (of at least an 8-hour duration) had an incidence of 10%.
Determining how important the conditions that develop in airplanes are in favoring the development of PTE is complex. Some authors1,9,13,14,16,18,19 have argued that any situation of prolonged sitting in a small space could create risk (ie, long car or train trips or even attending the theater for many hours). Circumstances present during air travel that favor the appearance of PTE. Venous flow speed is two thirds less when sitting than when in a decubitus position. Compression of the popliteal vessels on the edge of the seat promotes venous stasis.15 The decrease in atmospheric humidity, which increases imperceptible losses,20 insufficient intake of liquids, as well as the diuretic effect that is brought on by drinking alcohol, all bring about a certain degree of dehydration with the consequent hemoconcentration,1,11,15,16,19- 24 as proven by Simons and Krol.20 Plasma and urine osmolarity increases in healthy subjects after 8 hours of simulated flight. On the other hand, the system of pressurization on commercial airplanes can only maintain pressure similar to that of sea level at altitudes under 6800 meters. Most flights are flown at higher altitudes, along with the consequent partial reduction in oxygen pressure inside the plane.11,12,15,23 Hypoxia stimulates the production of oxygen-free radicals, and these in turn stimulate the production of relaxing factors derived from the vascular endothelium with the consequent venous vasodilatation, which favors stasis.23 At the same time, this diminishes the fibrinolytic capacity of plasma and, together with the decrease in atmospheric pressure, activates coagulation.16,25 Moreover, tobacco smoking would have an effect at 2 levels: increasing blood viscosity as well as the subject's degree of hypoxia.26
Another aspect that has created controversy is determining what type of PTE prevention measures airlines should communicate to their passengers. Different proposals have been made, ranging from simple measures including exercises for the lower extremities, changes in posture, intake of abundant liquids during the flight, and abstinence from tobacco and alcohol,1,10,11,13,15,18,23,24,27 to others such as the use of compression stockings28 (although these may favor the appearance of superficial thrombosis)2 or the use of prophylactic anticoagulation on an individual basis, according to the characteristics of each patient.3,11- 13,15- 17,19,21,26,27,29 The efficacy of these measures, however, has still not been established.3,17- 19 Our data indicate that special precautions are probably not required for flights less than 6 hours, but may become appropriate when passengers with risk factors for PTE embark on prolonged flights of 6 hours or longer.
In conclusion, air travel is a risk factor for PTE, and the incidence of PTE increases with the duration of the air travel. However, the low incidence of PTE among long-distance passengers, similar to that observed in other international airports, does not justify social alarm. Travelers on long-duration flights who have risk factors for PTE may benefit from simple preventive measures.
Corresponding author and reprints: David Jiménez, MD, Emergency Department, Ramón y Cajal Hospital, 28034 Madrid, Spain (e-mail: email@example.com).
Accepted for publication April 2, 2003.
From the Pneumology Department (Drs Pérez-Rodríguez, Díaz, Luque, Guillén, and Mañas) and the Emergency Department (Dr Jiménez), Ramón y Cajal Hospital, Madrid, Spain; Respiratory Department, Southampton Hospital, Southampton, England (Dr Pérez-Walton); and the Divisions of Pulmonary and Critical Care Medicine and General Medical Sciences, Washington University School of Medicine, St Louis, Mo (Dr Yusen).
We express many thanks to Luis Madrigal, Department of Statistics at the Madrid Barajas Airport and Alfonso Muriel, Department of Statistics at the Ramón y Cajal Hospital of Madrid.