Dear Readers

The British Thoracic Society made certain recommendations for patients with chest problems wishing to travel by air. I am a chest physician working in the UK and I hope this information will be helpful to the patients and health-care professionals in the North East. The article is very medical and lay person will find it difficult to understand. I, therefore, suggest that readers should consult a physician to understand the recommendation.

RESPIRATORY DISORDERS WITH POTENTIAL COMPLICATIONS FOR AIR TRAVELLERS
Asthma
Guidelines were identified relating to professional aircrew and potential recruits with asthma, but none were found relating to passengers. The flight environment experienced by commercial passengers should not pose a problem for most patients with asthma. In a review of all consecutive in-flight medical incidents reported for QANTAS airlines in 1993 there were 454 significant medical incidents, 9% of which were reported as respiratory tract infection or asthma.[38] A review of incidents on US commercial aircraft where an enhanced medical kit was used found that 10% of 362 episodes were due to asthma, lung disease or breathlessness.[39]

All airlines permit use of dry cell battery operated nebulisers, but there is usually a restriction during take off and landing because of the risk of electrical interference.[40] However, a Cochrane review has shown that spacers are as effective as nebulisers in treating acute asthma.[41] Co-morbidity may present a problem if the patient has severe airflow obstruction and hypoxia or if there is complicating cardiac disease. Low cabin humidity may present a theoretical risk of bronchospasm as a result of water loss from bronchial mucosa. A doctor's letter describing the patient's condition and listing medications is recommended.[42]

Cardiac disease
Cardiac disease is considered here briefly because it often co-exists with lung disease and may give rise to symptoms attributable to respiratory disease.

Co-morbidity may present more of a risk to the passenger than the respiratory disease alone, although no data exist to support or refute this view. One study measured SpO2 at simulated altitudes and on commercial flights in 12 patients with cyanotic congenital heart disease (CCHD) and acquired pulmonary hypertension and in 27 control subjects.[43] At the simulated altitude (equivalent to FiO2 15%) mean SpO2 fell from 86% (range 69–98%) to 78% (range 56–90%) in the patients and from 98% to 90% in the controls. During air travel the mean in-flight SpO2 was higher at 83% (range 78–94%). There were no changes in lactic acid concentrations, pH, or PaCO2 , and no clinical problems. The tolerance of patients with cardiorespiratory disease in a stable clinical condition to a moderate increase in hypoxaemia is unremarkable since they are effectively “acclimatised” to hypoxia. From the point of view of oxygen delivery to the tissues, a fall in SpO2 of 10% is easily overcome by a similar percentage increase in cardiac output. Hypoxaemia is a cardiac stimulant, and even patients in severe but stable heart failure can increase their cardiac output by 50% on mild exercise.

COPD
Data on patients with COPD are limited, and existing guidelines contain largely empirical advice based on relatively small studies. In addition to the risk of hypoxaemia, patients with severe COPD may be put at risk from high levels of carboxyhaemoglobin resulting from smoking. They may experience expansion of non-functioning emphysematous bullae and abdominal gases which could further compromise lung function.

Gong et al[27] studied 22 patients (13 men) with stable mild COPD (FEV1 < 80% predicted), 17 of whom reported variable discomfort (chest tightness or exertional dyspnoea) on previous flights. They inhaled sequential gas mixtures of 20.9% (sea level baseline), 17.1% (simulating 1524 m), 15.1% (simulating 2438 m), 13.9% (simulating 3048 m), and 20.9% oxygen (sea level recovery). With 15.1% inspired oxygen there was a mean fall in SpO2 of 11% from 94% to 83%. The lowest recordings were 87% on 21% inspired oxygen and 74% on 15.1% inspired oxygen. Progressive hypoxia induced mild hyperventilation resulting in small but significant falls in PaCO2 .

Supplemental oxygen was given during inhalation of 15.1% oxygen in five subjects and 13.9% oxygen in 16. PaO2 increased significantly with supplemental oxygen and PaCO2 returned to baseline or, in eight subjects, rose modestly above baseline. Heart rate rose and asymptomatic cardiac dysrhythmias occurred in 10 subjects; blood pressure was unchanged. Eleven subjects had no symptoms and 11 reported mild symptoms which did not correlate with hypoxia or hypoxaemia. Variable sleepiness noted by the investigators was partly reversed by supplemental oxygen.

Dillard et al [44] examined 100 patients (retired military personnel and dependents) with severe COPD over a period of 28 months. Forty four travelled on commercial flights, of whom eight reported transient symptoms during air travel but reached their destination apparently without complications. Those who did not travel by air had a lower mean FEV1 and greater use of domiciliary oxygen, suggesting that many patients with COPD choose not to fly. Christensen et al [45] studied 15 patients with COPD with FEV1 <50% predicted and sea level SpO2 >94%, PaO2 >9.3 kPa. Arterial blood gas tensions were measured at sea level, at 2438 m (8000 ft) and 3048 m (10 000 ft) in an altitude chamber at rest and during light exercise (20–30 watts). At 2438 m (8000 ft) PaO2 fell below 6.7 kPa in three patients at rest and in 13 during exercise. None developed symptoms, probably because of existing acclimatisation. Resting PaO2 >9.3 kPa or SpO2 >94% do not therefore exclude significant hypoxaemia at altitude in patients with severe COPD. Light exercise, equivalent to slow walking along the aisle of an aeroplane, may worsen hypoxaemia.

The risk of recurrent pneumothorax is discussed separately, but it should be noted here that COPD patients with large bullae are theoretically at increased risk of pneumothorax as a result of volume expansion at reduced cabin pressures. The volume of gas in a non-communicating bulla will increase by 30% on ascent from sea level to 2438 m (8000 ft). There is one case report of fatal air embolism in a patient with a giant intrapulmonary bronchogenic cyst.[46] However, there are no data to state with any confidence what the maximum volume of a bulla should be before it reaches an unacceptable level of risk of rupture leading to tension pneumothorax, pneumomediastinum, or air embolism. Recent UK guidelines on oxygen prescribing[47] quote evidence from two studies[24] [48] which suggest that the best predictor of PaO2 at altitude is pre-flight PaO2 on the ground. In one study the authors measured PaO2 and PaCO2 in 13 patients with COPD at 1650 m and 2250 m. No symptoms attributable to hypoxia were recorded although PaO2 fell from 9.1 kPa (68.2 mm Hg) at sea level to 6.6 kPa (51 mm Hg) at 1650 m and 6.0 kPa (44.7 mm Hg) at 2250 m. PaO2 on air at sea level measured some weeks before did not correlate with that measured at altitude, but PaO2 measured within 2 hours of flight time did. In the second study 18 retired servicemen with severe COPD were exposed to an altitude of 2438 m (8000 ft) in a hypobaric chamber.

Mean PaO2 fell from 9.6 kPa to 6.3 kPa after 45 minutes at steady state. The authors describe a predictive equation and recommend using the patient's pre-flight FEV1 to limit variation in the PaO2 result at altitude. In a review of acute responses of cardiopulmonary patients to altitude, Gong et al[49] recommend in-flight oxygen if the pre-flight PaO2 breathing 15% oxygen at sea level is <6.6 kPa. They conclude that equations do not accurately predict altitude PaO2 and favour the hypoxia altitude test.

A study of eight patients with mild to moderate COPD (FEV1 25–78% predicted) at sea level and after ascent to 1920 m (6298 ft) revealed no significant complications at altitude and 2,3-diphosphoglycerate levels remained unchanged.[50] This was despite levels of hypoxaemia similar to those observed in healthy mountaineers at altitudes of 4000–5000 m (13 000–16 000 ft).

The authors suggest that pre-existing hypoxaemia resulting from disease may facilitate adaptation of patients to hypoxia and prevent symptoms of acute mountain sickness. One study has examined the vasopressor responses to hypoxia in 18 men with severe COPD (mean (SD) FEV1 0.97 (0.32) l) at sea level, at 2438 m in a hypobaric chamber, and after oxygen supplementation at 2438 m.[51] Mean arterial pressure, systolic and diastolic blood pressure, and pulsus paradoxicus were unchanged at simulated altitude; oxygen reduced systolic blood pressure, pulsus paradoxicus, and pulse pressure. In one subject who developed increased cardiac ectopy, it was reduced by supplemental oxygen. The authors conclude that vasopressor responses to hypoxia do not increase the risk of flying in this group, but that in-flight oxygen may be beneficial. In summary, the clinical significance of temporary altitude induced hypoxaemia in patients with COPD is unclear.

The available controlled studies involve relatively small numbers of patients with stable disease and no co-existing medical problems. Simulated altitude exposure did not generally exceed 1 hour. These studies also largely excluded additional stressors such as exercise, dehydration, sleep, and active smoking. The only report to study exercise suggested that FEV1 <50% predicted is a risk factor for desaturation. We therefore recommend that patients with severe COPD are assessed before flying. Although there are no data to support this view, we also recommend that patients who require in-flight oxygen should receive oxygen when visiting high altitude destinations. Major high altitude destinations are listed in Appendix 4.

Cystic fibrosis
There are few data on the risks of air travel to patients with cystic fibrosis. In 1994 a study of 22 children with cystic fibrosis aged 11–16 years examined the value of hypoxic challenge testing.[52] The children were assessed in the laboratory, in the Alps, and on commercial aircraft and all desaturated at altitude. Hypoxic challenge was found to be the best predictor of hypoxia. However, a more recent study[37] of 87 children with cystic fibrosis aged 7–19 years who travelled on flights lasting 8–13 hours suggested that spirometric tests were a better predictor of desaturation. Low cabin humidity may increase the risk of acute bronchospasm and retention of secretions with possible lobar or segmental collapse, but there are no data to quantify this risk. Diffuse parenchymal lung disease There are no published data; clearly this is an area needing future research.

Infections
There is concern about the potential for transmission of infectious disease to other passengers on board commercial aircraft. There is also concern about the effect of travel after recent respiratory tract infections. The most important consideration is that of transmission of pulmonary tuberculosis, especially that of multiple drug resistant (MDR) tuberculosis.

Seven cases of possible transmission of Mycobacterium tuberculosis on aircraft have been reported to the Center for Disease Control and Prevention (CDC), Atlanta, Georgia, USA. The first concerned a flight attendant with documented tuberculin skin test (TST) conversion who did not receive prophylaxis and who developed pulmonary tuberculosis 3 years later.[53] The CDC concluded that the index case transmitted M tuberculosis to other flight crew members, but evidence of transmission to passengers was inconclusive. The second case concerned a passenger with pulmonary tuberculosis on a transatlantic flight.[54] Following a TST in 79 crew and passengers, eight had a positive TST. All had received Bacille Calmette-Guërin (BCG) vaccine or had a history of past exposure to M tuberculosis. The CDC found no evidence of in-flight transmission of tuberculosis. The third report concerned a passenger with pulmonary tuberculosis who travelled from Mexico to San Francisco.[55] Ninety two passengers were on the flight. The TST was positive in 10 of the 22 who completed screening, nine of whom were born outside the US and the tenth was a 75 year old passenger who had lived overseas and was thought likely to have been exposed to tuberculosis previously. The San Francisco Department of Health found no conclusive evidence of M tuberculosis transmission during the flight.

In the fourth case a refugee from the former Soviet Union with pulmonary tuberculosis travelled on three separate flights from Germany to his final destination in the USA.[56] Of 219 passengers and flight crew, 142 completed screening. The TST was positive in 32, including five TST conversions. Twenty nine had received BCG vaccine or had lived in countries where tuberculosis is endemic. The five passengers with TST conversions were seated throughout the plane and none sat near the index case. None of the US born passengers had TST conversions. The investigation concluded that transmission could not be excluded but that the TST conversions probably represented previous exposure to tuberculosis. The fifth report was of an immunosuppressed US citizen with pulmonary tuberculosis domiciled in Asia. He flew from Taiwan to Tokyo, then to Seattle, and subsequently to two further US destinations.[55] Of the 345 US residents on these flights, 25% completed screening. Fourteen had a positive TST, of whom nine were born in Asia. Of the remaining five, one had a positive TST before the flight, two had lived in a country with a high prevalence of tuberculosis, and two were aged over 75.

The investigators concluded that transmission of tuberculosis could not be excluded but that the positive TST results may have resulted from prior M tuberculosis infection. In the sixth report a passenger with pulmonary tuberculosis flew from Honolulu to Chicago and then to Baltimore where she stayed 1 month.[57] She then returned to Hawaii by the same route. Of 925 passengers resident in the US, 802 completed screening. Six passengers on the longer flight had TST conversions, four of whom were born in the USA and sat in the same section of the plane as the index case. The investigation considered that transmission of M tuberculosis had probably occurred.

In the final report a passenger with pulmonary and laryngeal tuberculosis flew from Canada to the US on three separate flights and returned 1 month later by the same route.[58] Five passengers had positive TST results but all had other possible explanations, and it was concluded that the likelihood of M tuberculosis transmission was low.

In all these reports the index patient was considered highly infectious and sputum specimens were heavily positive for acid fast bacilli. All were culture positive and had extensive pulmonary disease on chest radiography. Laryngeal tuberculosis is the most infectious form. In two instances the M tuberculosis strain isolated was resistant to at least isoniazid and rifampicin.[54] [57] Despite the highly infectious nature of all seven index cases, only two reports yielded evidence of TST conversion.[53] [57]

In the first case evidence of transmission was limited to crew members exposed to the index case for over 11 hours. In the second report transmission was demonstrated only in a few passengers seated in close proximity to the index case, and only on a flight lasting more than 8 hours. Although pulmonary tuberculosis does therefore appear to be transmissible during the course of air travel, none of the passengers with documented TST conversion have since developed active tuberculosis. The World Health Organisation (WHO) concludes that air travel does not carry a greater risk of infection with M tuberculosis than other situations in which contact with infectious individuals may occur, such as travelling by rail, bus, or attending conferences.[59]

There are other studies of potential transmission of airborne infectious diseases on aircraft. An influenza outbreak occurred in 1979 among passengers on a flight with a 3 hour ground delay before take off.[60] Seventy two percent of the 54 passengers developed symptoms; a similar virus was isolated from eight of 31 cultures, and 20 of 22 patients had serological evidence of infection with the same virus. The high attack rate was attributed to the ventilation system being switched off during the ground delay. Measles may be transmitted during international flights.[61] [62] In a study of patients with recent lower respiratory tract infections, Richards reported that 23 patients travelling by air after acute respiratory infection suffered no adverse effects.[63] There are no other data specifically relating to patients travelling after infection, and there is no evidence that recirculation of air facilitates transmission of infectious agents on commercial aircraft.

Neuromuscular disease and kyphoscoliosis
The data in this area are sparse, but there is one case report of cor pulmonale developing in a patient with congenital kyphoscoliosis after intercontinental air travel.[64] The patient was a 59 year old man with apparently stable cardiorespiratory function who developed a first episode of pulmonary hypertension and right heart failure after a long haul flight. The authors conclude that this resulted from prolonged exposure to the low FiO2 in the cabin. There are also anecdotal reports of oxygen dependent patients with scoliosis whose PaO2 has fallen precipitously during hypoxic challenge, despite a baseline oxygen saturation above 94% (A K Simmonds, personal communication).

Obstructive sleep apnoea
Few data exist regarding the effects of air travel on patients with obstructive sleep apnoea. Toff[65] reported a morbidly obese woman who developed respiratory and cardiac failure at the end of a 2 week tour involving two flights and a stay at altitude. It has been recognised since the 19th century that climbers to high altitude experience periodic breathing during sleep.[66] [67] [68] Apnoeic periods arise with reductions in arterial oxygen saturation and are nearly universal above 2800 m. Although generally thought harmless, periodic breathing can cause insomnia. It has also been speculated that the desaturations may contribute to altitude sickness. Three studies have examined this phenomenon in greater detail[69] [70] [71] but all the subjects were healthy volunteers. The apnoeas are thought to be central in origin. However, in the light of these observations it would seem prudent to recommend that patients using CPAP should take their CPAP machine with them when visiting high altitude destinations above 2438 m (8000 ft). Major high altitude destinations are listed in Appendix 4.

Previous pneumothorax
Thirty seven papers were reviewed. Airlines currently advise a 6 week interval between having a pneumothorax and travelling by air. The rationale for this recommendation is not explicit, but it is assumed that it reflects the time period during which a recurrence of a pneumothorax is most likely. In fact, the risk for a patient with a pneumothorax, if one were present, relates to ascent and descent, and a “new” pneumothorax occurring at altitude may be hazardous because of absence of medical care, but there should be no particular risk associated with pressure change. The “6 week rule” appears to have been arbitrarily applied with no account being taken of the type, if any, of underlying disease, or of any therapeutic intervention that has been undertaken, or of demographic factors.

The literature was reviewed to examine whether better evidence could be found for the timing of maximum risk of a recurrence of pneumothorax and to determine whether different advice should be offered to different subgroups of patients. Two papers were found relating to therapeutic interventions which included evidence about recurrence rates, and the following conclusions regarding timing of the recurrence or differences between subgroups were drawn.

If the pneumothorax was treated by a thoracotomy and surgical pleurodesis or by insufflation of talc (at thoracotomy), the recurrence rate should be so low that no subsequent restriction on travel is necessary.[72] Talc pleurodesis performed via a thoracoscopy may not be as successful in preventing recurrence of a pneumothorax—a 93% success rate was reported in one study[73] and a 92% success rate in another.[74] Similarly, other interventions via a thoracoscopy, even when using the same techniques as performed by a more major thoracotomy, may not always carry the same certainty of success,[72] although some good reports with no recurrence of pneumothorax have been published.[75] Further studies are required.

Non-talc chemical pleurodeses (for example, with tetracycline) are associated with a more significant and continued risk of recurrence—16% in one study with 50% of the recurrences arising in 30 days[76] and 13% in another.[74] The best figure found was a 9% rate of recurrence after chemical pleurodesis.[77] These recurrence rates suggest that, even after such an intervention, the patient should still be subject to travel advice applied to others after a spontaneous pneumothorax.

For patients who have not had a definitive surgical pleurodesis via a thoracotomy, a risk of recurrence should therefore be expected. While many studies have included details of the percentage of patients suffering a recurrence, very few have given much detail of the timing of these recurrences after the first episode, and few have characterised those most at risk. In one study a 54.2% recurrence rate was recorded with the majority occurring within 1 year of the first pneumothorax,[78] and in another study 72% of the recurrences occurred within 2 years of the first episode.[79]

Cumulative freedom from recurrence data have been published by Lippert et al[79] and stratified according to smoking history and underlying lung disease over a follow up period of up to 13 years. The shape of the curve (fig 2) does indeed imply that the biggest risk of recurrence is in the first year. One author has intimated that a further prospective trial he and colleagues are currently undertaking may provide a clearer month by month detail of recurrence rates.

At present the recommended 6 week cut off seems to be arbitrary, with a significant fall in risk only appearing to occur after 1 year has elapsed. Furthermore, current advice does not take into account those with a higher risk of recurrence such as smokers, those with pre-existing lung disease, taller men, and possibly women.[79] [80] Thoracoscopic examination of the pleura does not permit any greater prediction of those at greatest risk of recurrence.[79] [81]

In conclusion, a definitive surgical intervention makes the risk of recurrence of a pneumothorax negligible. Such patients may be able to fly 6 weeks after surgery and resolution of the pneumothorax in the absence of other contraindications. Careful medical assessment is required beforehand. For others the risk of a further pneumothorax is considerable for at least a year after the first episode. This risk is greatest for those with underlying lung disease and for continuing smokers.

While the likelihood of recurrence during flight is low and there is no evidence that air travel precipitates recurrence, the sequelae of recurrence at altitude may be significant. Recurrence of a pneumothorax while flying is likely to have more serious effects than a first episode, and recurrence in passengers with pre-existing lung disease is more likely to have serious consequences. Passengers may therefore choose to avoid this risk by delaying air travel for 1 year after a pneumothorax. This strategy should be given special consideration by those who smoke and/or have underlying lung disease.

Venous thromboembolic disease (VTE)
Fourteen papers were reviewed but the evidence is conflicting with many questions unanswered. BTS guidelines on suspected pulmonary thromboembolism list six major risk factors for VTE.[82] Air travel is classified as one of several lesser risks. The evidence quoted in favour of an increased risk of air travel relates to long haul flights.[83] [84] Such reports are supported by others dating back over 20 years,[85] [86] [87] [88] and by more recent surveys.[89] [90] [91] It is not possible from the published data to quantify the risk, and the underlying mechanisms have not been elucidated. Hypotheses include immobility, seated position, dehydration, and alcohol ingestion. Owing to delayed onset of symptoms and rapid dispersal of patients after a flight, many current reports are likely to underestimate the size of the problem. In small studies evidence suggests that co-morbidity may increase the risk of VTE associated with air travel.[89] [90] Some studies suggest that previous VTE increases the risk of air travel associated recurrence,[89] [90] [91] [92] [93] but the data are controversial.

Further research is needed to determine whether delay in travel for those at risk is beneficial, and whether avoidance of alcohol and dehydration and upgrading reduce risk. Research is also required to examine the potential role of prophylactic low molecular weight heparin, full formal anticoagulation, and mechanical prophylactic methods including graded elastic compression hosiery and full leg pneumatic compression devices. The latter may be impractical on board an aeroplane and have not been studied in this context. However, they have been shown to have an additive effect in other at risk situations.[94] A recent study suggests that symptomless deep vein thrombosis may occur in up to 10% of airline passengers, and that wearing elastic compression stockings during long haul flights is associated with a reduced incidence.[95]

The role of aspirin in this setting also requires investigation. A study of 13 356 patients undergoing surgery for hip fracture and 4088 patients undergoing elective arthroplasty showed that aspirin reduces the risk of pulmonary embolism and deep vein thrombosis by at least one third throughout a period of increased risk.[96] The authors conclude that there is now good evidence for considering aspirin routinely in a wide range of groups at high risk of thromboembolism.

Thoracic surgery
There are few data available, but it is clear that the volume of gas in air spaces will increase by 30% at a cabin altitude of 2438 m (8000 ft). Postoperative complications such as sepsis or volume depletion should have resolved before patients undergo air travel. Severe headache precipitated by airline travel has been recorded 7 days after a spinal anaesthetic, presumed to be due to cabin pressure changes inducing a dural leak.[97] North American guidelines[13] highlight the fact that postoperative patients are in a state of increased oxygen consumption due to surgical trauma, possible sepsis, and increased adrenergic drive. Oxygen delivery may be reduced or fixed in patients who are elderly, volume depleted, anaemic, or who have cardiopulmonary disease. Reduced use of transfusions means that postoperative patients are now often more anaemic than previously.

Logistics of travel with oxygen
Berg et al[98] have investigated the effects of oxygen supplementation in a group of 18 patients with severe COPD (mean FEV1 31% predicted). Baseline PaO2 at sea level was 9.47 kPa, which fell to 6.18 kPa when exposed to an altitude of 2438 m in a hypobaric chamber. The subjects were then given supplemental oxygen; 24% oxygen by Venturi mask increased PaO2 to 8.02 kPa, 28% oxygen by Venturi mask increased PaO2 to 8.55 kPa, and 4 l/min via nasal prongs increased PaO2 to 10.79 kPa. This suggests that, in patients with COPD, 24% and 28% oxygen via Venturi masks (and probably 2 l/min via nasal prongs) will improve hypoxaemia at 2438 m but will not fully correct it to sea level values. However, oxygen given at 4 l/min via nasal prongs will overcorrect hypoxaemia to produce values above sea level baseline values.

In practical terms, aircraft oxygen delivery systems are usually limited to 2 or 4 l/min. This is probably best delivered by nasal prongs as the simple oxygen masks provided by many airlines may allow some re-breathing and worsen carbon dioxide retention in susceptible subjects. Using 100% oxygen at a rate of 4 l/min via nasal prongs from a cylinder will produce a PaO2 at 2438 m (8000 ft) cabin altitude slightly higher than sea level PaO2 on air. Using 2 l/min via nasal prongs should correct the fall in oxygenation. Patients who require LTOT are not excluded from air travel, but no randomised controlled trials exist on which to base recommendations on the optimal flow rate.

The method of oxygen delivery depends upon the specific aircraft, but the supply is usually from cylinders. In some aircraft oxygen can also be tapped from the “ring main” of oxygen.[99] Patients are not allowed to use their own oxygen equipment on the aircraft but can take an empty oxygen cylinder or oxygen concentrator as baggage. Charges may be made for both services, as well as a charge for supplemental oxygen. Regulations vary with each airline, which can decline the patient's request to travel.[100] A comparative study of arranging in-flight oxygen on commercial air carriers was performed by members of the respiratory therapy department at the Cleveland Clinic Foundation in Cleveland, Ohio;[101] 76% of the 33 carriers contacted offered in-flight oxygen. There was significant variation in oxygen device and litre flow availability. Flow options varied from only two flow rates (36% of carriers) to a range of 1–15 l/min (one carrier). All carriers provided nasal cannulae, which was the only device available on 21 carriers. Charges varied considerably. Six carriers supplied oxygen free of charge while 18 carriers charged a fee ranging from $64 to $1500. Charges for an accompanying empty cylinder ranged from none to $250. Most carriers required 48–72 hours advance notice; one required one month's notice.