Low-Dose Theophylline Reduces Symptoms of Acute Mountain Sickness


  • Thomas E.A.H. Küpper MD,

    1. Hermann Buhl Institute for Hypoxia Research and Sleep Medicine, Paracelsus Medical University Salzburg, Bad Aibling, Germany
    2. Institute of Occupational and Social Medicine, Faculty of Medicine, Aachen Technical University, Aachen, Germany
    3. Union Internationale des Associations d’ Alpinisme Medical Commission, Bern Switzerland
    Search for more papers by this author
  • Kingman P. Strohl MD, FCCP,

    1. Centre for Sleep Disorders Research, Louis Stokes VA Medical Center, Case Western Reserve University, Cleveland, OH, USA
    Search for more papers by this author
  • Markus Hoefer MD,

    1. Division of Pulmonary Medicine, University of Homburg, Saar, Germany
    Search for more papers by this author
  • Ulf Gieseler MD,

    1. Department of Internal Medicine, Diakonissenkrankenhaus, Speyer, Germany
    Search for more papers by this author
  • Cordula M. Netzer MD,

    1. Centre for Sleep Disorders Research, Louis Stokes VA Medical Center, Case Western Reserve University, Cleveland, OH, USA
    Search for more papers by this author
  • Nikolaus C. Netzer MD, PhD

    1. Hermann Buhl Institute for Hypoxia Research and Sleep Medicine, Paracelsus Medical University Salzburg, Bad Aibling, Germany
    2. Centre for Sleep Disorders Research, Louis Stokes VA Medical Center, Case Western Reserve University, Cleveland, OH, USA
    Search for more papers by this author

Thomas E.A.H. Küpper, MD, Institute of Occupational and Social Medicine, Faculty of Medicine, Aachen Technical University, Pauwelsstr 30, D-52074 Aachen, Germany. E-mail: tkuepper@ukaachen.de


Objective Headache, nausea, and sleeplessness at altitude [acute mountain sickness (AMS)] are major health problems for several million mountain recreationists who ascend to high altitudes each year. We aimed to test the efficacy of low-dose, slow-release theophylline for the prevention of AMS in a placebo-controlled, double-blind, randomized trial.

Methods Twenty healthy male volunteers (mean age 34.7 y) were randomized (random allocation) to receive either 300 mg theophylline daily or placebo 5 days prior, during ascent, and during a stay at 4,559 m altitude. AMS symptoms were collected using the Lake Louise Score on each day during ascent and at high altitude. A 12-channel sleep recorder recorded sleep and breathing parameters during the first night at 4,559 m. Theophylline serum levels were drawn prior to the sleep study.

Results Seventeen completed the entire study. Theophylline (n = 9) compared to placebo (n = 8) significantly reduced AMS symptoms at 4,559 m (Lake Louise Score: 1.5 ± 0.5 vs placebo 2.3 ± 2.37; p < 0.001), events of periodic breathing (34.3/h vs placebo 74.2/h; p < 0.05), and oxygen desaturations (62.3/h vs placebo 121.6/h; p < 0.01). No significant differences in sleep efficiency or sleep structure were present in the two groups. No adverse drug effects were reported.

Conclusions Low-dose, slow-release theophylline reduces symptoms of AMS in association with alleviation of events of periodic breathing and oxygen desaturations.

People in increasing numbers participate in recreational and professional mountaineering1 and all other types of altitude sojourn (leisure, occupation, etc.). Many suffer from acute mountain sickness (AMS),2,3 and the medical community has emphasized the importance of not only its recognition but also its prevention.4,5

AMS is characterized by symptoms of headache, nausea, and insomnia and can be an early sign of high-altitude cerebral edema.6 The consensus is that at altitudes above 4,000 m, the incidence of AMS reaches 70%.7 As a result, AMS is a major topic of interest for those in general medicine who counsel hikers.5,8

AMS is caused by hypoxic exposure.9 Hypoxia leads to increased vascular permeability,10–12 perhaps through effects of oxidative stress.13–16 Sleep-disordered breathing at altitude and Cheyne–Stokes respiration (CSR) in particular add to the hypoxic burden. While some studies showed a correlation between the AMS and the hypoxic burden during night (either directly17–20 or indirectly by reducing symptoms via oxygen enrichment during night21), there are controversial data.

Acetazolamide and dexamethasone, respectively, are the most investigated and used for prevention of AMS. Both drugs, compared to placebo and to each other, have proven to be effective in the treatment and prevention of AMS.22–29 Another recent study compared acetazolamide with Ginkgo biloba and found superior efficacy with acetazolamide.30 However, the quite high dosages needed with acetazolamide can also produce adverse reactions.7 With recent studies, which have shown the efficacy of acetazolamide even at lower dosage and with less side effects,31–33 the debate about the best strategy is still open.

Theophylline is known to increase the ventilatory response in normal man and in patients with sleep-disordered breathing including CSR.34–36 Theophylline at a low dose is expected to increase ventilation and reduce hypoxic exposure. This central effect of theophylline on ventilation is reached with low dosages, ie, serum levels around 3 mg/L, lower than needed for bronchodilation.37,38 In a recent, open-label trial, theophylline showed a positive preventive effect on AMS in climbers at high altitude (partially simulated in an altitude chamber).39 But as discussed later in this paper, this study suffered from several limitations, some of which caused dropouts of subjects.

The primary objective in this investigation was to study the effects of low-dose theophylline on ventilation and oxygenation at night, on sleep, and on clinical symptoms of AMS in a randomized, placebo-controlled drug trial to high altitude. The secondary objective was to study the efficacy of a low-dose regimen with theophylline serum levels around 3 mg/L and starting medication before ascent to minimize side effects. This aimed to exclude any side effects and to avoid dropouts as much as possible.


This trial design was a double-blind, placebo-controlled study with healthy male volunteers and approved by the institutional review board and ethical committee of the Medical Faculty at the University of Ulm, Germany. The high-altitude study site was the European high-altitude research laboratory at the Margherita hut on Monte Rosa (Italy) at 4,559 m altitude.

Twenty healthy male experienced recreational mountaineers (sport students and members of the German Alpine Club, all lowlanders; aged 19–54 y, mean age 34.7 y) volunteered and gave written consent to participate in this trial. Portable seven-channel sleep recording (Polymesam; MAP, Martinsried, Germany) was performed at the volunteer’s home at sea level prior to beginning of the study to exclude an underlying sleep-disordered breathing. Additional exclusion criteria were a history of heart disease, high blood pressure, moderate- to high-altitude pulmonary or cerebral edema, drug or alcohol abuse, and systemic drug treatment. None of the subjects had a history of AMS.

Participants were randomized (random allocation; see Figure 1) to receive either 300 mg slow-release theophylline tablets (Unilair 300; 3M Pharmaceuticals Inc., Neuss, Germany) or an identical-appearing placebo. Randomization codes were broken on completion of the study and after collection and scoring of questionnaires, sleep parameters, and blood tests.

Figure 1.

Consort Statement flowchart of randomization and placebo-controlled drug trial.

Participants received a container with 20 tablets at the beginning of the study. Each participant was reminded to take one tablet (drug or a placebo) at 8 pm each day starting 5 days prior to the ascent and continued drug intake during ascent (2 d, 1 night at Mantova hut) and stay at the study site at 4,559 m altitude for another 5 days and 6 nights (altitude profile shown in Figure 2). At the last morning at high altitude, before descent, the container was recollected and the remaining tablets were counted.

Figure 2.

Altitude profiles and study places during study duration. “Wash in”: wash-in phase with study medication or placebo to reach therapeutic theophylline serum levels before ascent.

The AMS Lake Louise Score was obtained via Lake Louise questionnaire from each individual twice a day (8 am and 8 pm) during the entire study duration except descent. The Lake Louise questionnaire contains five questions for self-report regarding symptoms and signs of AMS, quality of well-being, and sleep; answers are given based on a four-grade scale.40 The maximum score to be reached is 15. In most studies, like in ours, an AMS score of 4 or more is defined as AMS.41–43 The AMS morning and evening scores of the first 3 days at 4,559 m were evaluated separately from the scores of all 5 days at 4,559 m to obtain information about the effects of acclimatization. Parallel to AMS reporting, the subjects were checked for theophylline-related side effects (questionnaire and heart rate).

Blood samples were drawn from each participant on the first and fifth night at high altitude (study days 7 and 12) at 10 pm, 2 hours after drug intake. Blood samples were centrifugated at 5,000 rpm for 10 minutes and stored in liquid nitrogen at −196° until analysis. The liquid nitrogen container with the serum samples was later transported via helicopter and car to Düsseldorf for analysis. Theophylline levels were measured in a batch analysis by fluorescence polarization immunoassay (Axsym; Abbott Laboratories, Chicago, IL, USA).

To assess sleep hypoxemia and breathing pattern, one sleep study was performed via 12-channel polysomnography (Sidas GS; Stimotron/Respironics, Pittsburgh, PA, USA) during the first night at 4,559 m in each participant. Measurements were performed at constant room temperature of 18°C. SaO2 data were obtained at the right index finger.

For all subjects, the lights-out time was between 10 pm and 11 pm. The subject was allowed to awaken spontaneously at a time that varied between 6:30 am and 7:30 am. The following parameters and channels were measured: two electroencephalogram (EEG) channels, two electrooculogram (EOG) channels, two electromyogram (EMG) channels (chin), oxygen saturation via pulse oximetry, thoracic and abdominal respiratory effort via belts, electrocardiogram (ECG), nasal and oral flow via thermistor, and body position.

Sleep study data were recorded by notebook computer and stored for analysis. Sleep staging was performed by one technician using Rechtschaffen and Kales criteria32 after the end of the altitude measurements. Respiratory events were recorded using the following criteria. An apnea was defined as minimum 80% decrease of airflow for at least 10 seconds, and hypopnea was defined as 50% to 80% decrease of airflow for at least 10 seconds. Obstructive events were defined as a decrease or missing airflow without complete missing of abdominal or thoracic respiratory effort; central events were defined as missing airflow and abdominal or thoracic respiratory effort. CSR was defined by the typical respiratory patterns with short (duration less than 10 s) central apneas and central hypopneas. All respiratory events were added for the calculation of the respiratory disturbance index (RDI; respiratory events per hour of sleep). An oxygen desaturation was defined as a drop of 4 percentage points from baseline in oxygen saturation.

Polysomnographic parameters used for statistical analysis were RDI, oxygen desaturation index (ODI; oxygen desaturations per hour sleep), SaO2 mean (mean arterial oxygen saturation during sleep), SaO2 low (lowest arterial oxygen saturation during sleep), rapid eye movement (REM) sleep in % of total sleep time (TST), NREM 1 + 2 (non-REM sleep stages 1 + 2) in % TST, NREM 3 + 4 (delta sleep) in % TST, and sleep efficiency (TST/total time in bed in %).

The number of participants was chosen based primarily on space availability of the study site and secondarily on participant availability. Ten participants were in each group to compensate for dropouts or missing data.

Data were entered into data files using SPSS 8.00 software (SPSS, Chicago, IL, USA) for analysis. The two groups were compared with use of Mann–Whitney U-test for nonparametric variables. Lake Louise questionnaire data were analyzed by two-way analysis of variance with time as one factor and treatment group assignment as the other. The incidence of AMS of placebo versus theophylline group was tested by chi-square test. A p value of 0.05 or less was considered to indicate statistical significance.


Seventeen participants completed the study, 9 in the theophylline (mean age 36.2 y) and 8 in the placebo group (mean age 32.1 y). Three participants were unable to ascend to Margherita hut due to adverse weather conditions and personal time constraints.

No side effects were reported by either group; specifically absent were reports of nausea or palpitations. No participant developed high blood pressure (systolic pressure >160 and diastolic pressure >90) during the study. Mean theophylline serum concentrations were 0.31 mg/L (±0.18) in the placebo group and 3.41 mg/L (±0.87) in the theophylline group (p < 0.001; Figure 3).

Figure 3.

Mean theophylline serum concentrations in the placebo and theophylline group. Standard deviation marked by crossbars.

At sea level, there was no difference in the AMS score between the drug and the placebo group (0.5 ± 0.89 vs 0.7 ± 1.04). All subjects reported familiarity with the self-rating AMS scale before ascent.

Theophylline and placebo group showed a significant difference in the mean AMS at the intermediate stay at 3,440 m (theophylline 0.8 ± 0.75, placebo 1.6 ± 1.60; p < 0.05). In the placebo group, one individual had an AMS score of 7 after the night at this altitude indicating the presence of AMS. No individual of the theophylline group had an AMS score >4 at any time at this altitude. At highest altitude, the difference between placebo and theophylline expressed by the mean scores calculated for the first 3 days was significant (theophylline 1.5 ± 0.5, placebo 2.3 ± 2.37; p < 0.001; Figure 4). Calculated for all 5 days at 4,559 m, the difference was less pronounced but still significant (theophylline 1.5 ± 0.5, placebo 1.68 ± 1.87; p < 0.01).

Figure 4.

Mean acute mountain sickness (AMS) (Lake Louise) scores of placebo and theophylline group for each day at 4,559 m. Standard deviation marked by crossbars. Level of significance (p) calculated by Mann–Whitney U-test. Variance analysis by two-way analysis of variance shows global significance for the main effect (p < 0.0001), which is caused by group (p < 0.0001) as well as by time (p < 0.0001). There is no interactive effect between the two factors (p = 0.920).

At highest altitude, the maximal score was 10 in the placebo group and 9 in the theophylline group. AMS by definition occurred in seven of eight cases   (87.5%) of the placebo group and in five of nine   (55.6%) cases of the theophylline group. The difference in AMS incidence between the two groups did not reach statistical significance (p = 0.29). Therefore, theophylline did reduce the severity but not the incidence of AMS in this study. Maximum scores in both groups were reached after the first night at high altitude.

The analysis of the different AMS symptoms using the subgroups of questions in the Lake Louise questionnaire showed no difference between both groups in gastrointestinal problems and dizziness/lightheadedness; both were rare or absent during the study. There was significantly less headache in the theophylline group at high altitude ( p < 0.05). Fatigue/weakness was less in the theophylline group at high altitude ( p < 0.04), while at moderate altitude, there was no difference between these parameters. The difference in self-reported sleep quality was highly significant between placebo and theophylline group at moderate altitude ( p < 0.005) as well as at high altitude ( p < 0.001).

The comparison between the morning versus evening ratings of the main AMS symptom “headache,” showed higher values, respectively more severe headache, in the morning. This was significant at high altitude in both groups (theophylline p < 0.001; placebo p < 0.02). The symptom scores assessed with questions regarding the mental status, ataxia, peripheral edema, and functional assessment showed no high-altitude influence at all in both groups.

Respiratory parameters in the theophylline and placebo group showed significant differences regarding the RDI (mean RDI in theophylline group 34.3, in placebo group 74.2; p < 0.05; individual values are shown in Table 1) and the ODI (mean ODI in theophylline group 62.3, in placebo group 121.6; p < 0.01; Table 1). All recorded respiratory events were classified as CSR; obstructive apneas or obstructive hypopneas were not recorded in either group. Mean oxygen saturation during the night was 73% in the theophylline group and 77% in the placebo group, a difference that did not reach statistical significance (Table 1). No difference was seen in regard to lowest oxygen saturation between placebo and theophylline (Table 1).

Table 1.  Respiratory parameters of the individual subjects during the study night at 4,559 m on study day 7/8 (first night at high altitude)
IDTheophylline/placeboAge (y)RDIODISaO2 mean (%)SaO2 low (%)
  1. RDI = respiratory disturbance index (respiratory events/Cheyne–Stokes respiration/h); ODI = oxygen desaturation index (oxygen desaturations >4% SaO2/h); SaO2 mean (mean arterial oxygen saturation during the night); and SaO2 low (lowest arterial oxygen saturation measured during the night).


Sleep data could not be evaluated in one participant of the placebo group due to electrode problems. Results are based on the other 16 individuals. Sleep efficiency showed a tendency to be better in the placebo group (85% in placebo vs 75% in theophylline), but this did not reach statistical significance  (p = 0.14).

No statistically significant difference could be found between placebo and theophylline group regarding the percentages and distribution of sleep stages during the study night.

Average TST—percentage of REM—was 14.1% in theophylline and 14.1% in placebo, NREM 1 + 2 was 51.0% in theophylline and 38.0% in placebo, and NREM 3 + 4 was 35.0% in theophylline and 46.0% in placebo.


The present randomized, double-blind trial establishes the efficacy of low-dose theophylline in the reduction of AMS. We found that compared to placebo, theophylline reduced symptoms of AMS significantly at moderate altitude during ascent and during a 5-day stay at high altitude. Participants on theophylline subjectively felt a better sleep quality and exhibited improved indices for sleep-disordered breathing than the placebo group, without affecting overall sleep efficiency or sleep structure that was present.

Theophylline did not increase mean oxygen saturation, but it did reduce CSR and oxygen desaturations. The impaired ventilation at night at high altitude with CSR and oxygen desaturations is considered the main factor for the pathogenesis of subjectively reduced sleep quality and morning symptoms of AMS.18–20,44 Influenced by the ventilatory response to hypoxia, CSR occurs more often during sleep at high altitude.45,46 Our observations are consistent with the beneficial effect of low-dose, slow-release theophylline on sleep-disordered breathing at sea level.36

The mechanism we found for the beneficial effect was most likely related to a reduction in the number of oxygen desaturations during sleep by a stimulation of respiratory drive.35 The observation is consistent with a study showing that those treated with theophylline for CSR at sea level also experience daytime benefit.36 Additional effects by theophylline on AMS symptoms might include a decrease in adenosine-mediated cerebral blood flow37 or a reduction in inflammatory responses and vascular permeability as a result of its phosphodiesterase inhibitor activity.47 The very low pharmacodynamically ineffective serum theophylline level measured in our placebo group likely resulted from coffee or tea consumption, which was not restricted in either group.

No adverse side effects were observed. The main systemic sympathomimetic effects of theophylline like increased heart rate, palpitations, seizures, sleeplessness, and diuresis are usually not found at serum levels below 25 mg/L. The serum levels needed to improve sleep-disordered breathing are achieved by once-daily treatment. Obviously, our drug regimen, which differs from those of Fischer’s trial39 by administering the drug well before ascent, is the main factor that eliminated any side effects completely. Although the altitude where our study was performed was significantly higher than Fischer’s, no tachycardia was observed, which is in contrast to Fischer’s data.39 As an open trial with a relative high number of dropouts, Fischer’s study is limited. Most of these limitations were solved by our randomized, double-blind setting and the drug regimen, although the number of subjects in our study was relatively small.

In contrast to our drug regimen, the recommended dosages of acetazolamide (500–750 mg/d48) to prevent AMS are associated with paresthesia and polyuria7 and may contribute to volume depletion, although this problem could be partially solved by lower dosage (2× 125 mg/d), which has been tested for efficacy.31–33

A recent publication is in accordance with our investigation regarding the normalization of ventilation at night at high altitude (3,454 m) and reduction of oxygen desaturations with theophylline but describes improved basal oxyhemoglobin saturation only with acetazolamide.49 The conclusion of this recent publication is that acetazolamide is more effective in the prevention of AMS. However, summarizing effects and possible side effects, we would consider theophylline a good alternative especially for those who cannot tolerate acetazolamide or who are allergic to it. More direct comparisons of the different drug regimens will be needed to establish both relative efficacy and side effect profiles.

As a study of clinical efficacy, the limitation of this study is the low number of observations. With the data obtained, the sample size to investigate the reduction of AMS incidence can be calculated: assumed a power of 80% and p < 0.05 as significance, 31 subjects would be needed (calculated with nQuery software). As a result, we were not able to demonstrate a significant decrease in the incidence of AMS through administration of theophylline, despite the fact that we could show a significant reduction of AMS symptoms. The low number of participants is a problem with high-altitude mechanistically oriented drug trials because of the environmental constraints and limited space at high altitude or in hypoxic research chambers. Our participation rate is comparable to other placebo-controlled drug trials at high altitude.7,8 It added to the limitation of the low number of subjects that for safety reason, our subjects had no history of AMS. Having subjects with AMS risk might have increased significance of theophylline effects.

In conclusion, our results confirm the previously observed positive effect of theophylline in the reduction of AMS symptoms found in the open-label study. This effect of low-dose, slow-release theophylline administered once is mediated by improved ventilation at night at high altitude. This regimen with onset of drug administration well before ascent is well tolerated and causes significant less side effects than others without a “wash-in phase” like Fischer’s.39 The clinical utility of low-dose theophylline, however, will be determined by comparison trials to acetazolamide while using our drug regimen.


The authors wish to thank Prof Peter Bärtsch, MD, Heidelberg, Germany, for the idea for this investigation, his continuous advice, and constructive criticism and Mrs Claire Küpper, PhD, Düsseldorf, for her assistance with randomization and preparation of drug and placebo samples. This investigation was supported by an unrestricted grant of 3M Pharmaceuticals Inc., Neuss, Germany. 3M Pharmaceuticals Inc. also provided the study medication and placebo. Respironics Inc., Pittsburgh, PA, USA, provided logistic support (sleep recorders and laptops during study duration and helicopter flights for transport of this material). The Margherita hut research lab is supported by several European universities, the Italian Alpine Club, and structural and research funds of the European Union.

Declaration of interests

The authors state that they have no conflicts of interest.