Clinical Features of Patients With Severe Altitude Illness in Nepal


  • Presented at the 10th Conference of the International Society of Travel Medicine, May 20 to 24, 2007, in Vancouver, Canada.

Eli Schwartz, MD, Center for Geographic Medicine, Sheba Medical Center, 52621 Tel Hashomer, Israel. E-mail:


Background Trekking in Nepal is a popular adventure travel activity involving more than 80,000 people of all ages annually. This study focuses on the demographic characteristics and clinical course of altitude illness patients evacuated to Kathmandu and estimates the rates of evacuation in different regions of Nepal.

Methods During the years 1999 to 2006, all patients who presented with altitude illness to the CIWEC clinic in Kathmandu were evaluated and included in the study if the final diagnosis was compatible with high-altitude cerebral edema (HACE), high-altitude pulmonary edema (HAPE), or acute mountain sickness (AMS). Altitude illness–related deaths were reported according to death certificates issued by selected embassies in Kathmandu.

Results A total of 406 patients were evaluated, among them 327 retrospectively and 79 prospectively. HACE was diagnosed in 21%, HAPE in 34%, combined HAPE and HACE in 27%, and AMS in 18%. Mean patient age was older than trekker controls (44 ± 13.5 vs 38.6 ± 13.9 y, p < 0.0001). Everest region trekkers were more likely to be evacuated for altitude illness than trekkers in other regions. The estimated incidence of altitude illness–related death was 7.7/100,000 trekkers. Most altitude illness symptoms resolved completely within 2 days of evacuation.

Conclusions Altitude illness that results in evacuation occurs more commonly among trekkers in the Everest region and among older trekkers. The outcome of all persons evacuated for altitude illness was uniformly good, and the rate of recovery was rapid. However, the incidence of altitude illness–related death continued to rise over past decade.

Trekkers in Nepal routinely hike to altitudes of 5,500 m or more, over a period of 1 to 3 weeks. An incidence of over 30% to 40% of altitude illness in trekkers above 4,000 m has been well described in previous studies.1–3

In the past, lack of communication, few helicopters, and a need to secure payment in advance made helicopter evacuation difficult to accomplish for any sick or injured person in the mountains. In recent years, the advent of satellite phones, private helicopter companies, and the acceptance of credit cards for payment have greatly lowered the threshold for requesting and obtaining a helicopter evacuation. Thus, altitude illness patients more often get evacuated directly from the mountains to a medical facility in Kathmandu at 1,310 m in altitude.

The CIWEC Clinic Travel Medicine Centre in Kathmandu is the most common medical facility to receive helicopter evacuations for altitude illness. This study describes the demographic and clinical features in patients with altitude illness who were evacuated to our clinic during a 7-year period.

Patients and methods

All altitude illness patient files from the years 1999 to 2005 were reviewed. The series included patients if the diagnosis of high-altitude cerebral edema (HACE) and acute mountain sickness (AMS) was compatible with the Lake Louise criteria and diagnosis of high-altitude pulmonary edema (HAPE) compatible with the 1991 International Hypoxia Symposium criteria.4 Patients were classified as having HACE, HAPE, combination of HAPE and HACE, or AMS. Exclusion criteria consisted of insufficient data in the patient file to diagnose altitude illness according to the criteria or a change in the final diagnosis to a cause of illness other than altitude illness. All patients were followed to complete resolution of symptoms.

The definition of a concomitant respiratory infection was based on the presence of a combination of clinical symptoms and signs suggestive of infection, such as productive cough with purulent sputum, sore throat, rhinorrhea, sinus or ear pain, high fever, leukocytosis, and failure of respiratory symptoms to resolve following evacuation to lower altitude. Infectious diarrhea diagnosis was confirmed by stool samples showing leukocytes and mucus.

From July 2005 to February 2006, a prospective study was conducted of altitude illness patients as they arrived at the clinic. These patients were asked to complete a questionnaire documenting demographic data, ascent profile, and symptoms.

The highest altitude usually attained with a normal trekking permit is 5,600 m (18,400 feet). Mountaineering in Nepal requires a special permit that allows climbers to attempt peaks from 5,600 to 8,848 m. Data on trekker numbers in the major trekking areas in Nepal were extracted from the Nepal Tourism Statistics Annual Statistical Report.5 Demographic data for the general trekker population were obtained from the trekking area permit registry at the Department of National Park and Wildlife Conservation for Sagarmatha National Park (Everest region) and from the Nature Conservation Trust for the Annapurna area. Detailed data on trekker profiles were available for trekkers in the Everest region during the period of September to October 2005 and for the Annapurna region trekkers during the period of January to December 2004. These data were used for trekker control group purposes.

Eight embassies (of a total of 18 embassies) representing the nationalities of 67% of trekker arrivals to Nepal were contacted regarding tourist mortality data. Six of these eight embassies representing 49% of trekker arrivals to Nepal had data regarding timing, location, and cause of trekker deaths based on death certificates from 1999 to 2005. Thus, trekker deaths attributed to altitude illness were recorded from these six embassies: Japan, France, Germany, Holland, Australia, and Israel. The embassies of the UK and United States representing 17% of trekker arrivals to Nepal did not have data regarding the cause of death, and therefore, death cases were excluded from analyses.

Data were stored in Access (Microsoft Office 2003) electronic chart and transferred to Excel (Microsoft Office 2003) for statistical analyses. For odds ratio calculations, we used the Annapurna region as the control due to the low starting altitude and gradual exposure to altitude. We used Fisher’s exact test for categorical data and unpaired t-test for continuous data.


Of 581 patients initially evaluated, 175 were excluded. Insufficient data resulted in exclusion of 158 patients, while 17 patients were excluded due to a final diagnosis other than altitude illness. Thus, 406 patients were included; among them, 327 patients were evaluated retrospectively and 79 prospectively (Figure 1). HACE was diagnosed in 21% (N = 85) of all patients, HAPE in 34% (N = 136), combined HAPE and HACE in 27% (N = 111), and AMS was diagnosed in 18% (N = 74).

Figure 1.

Study flowchart. HAPE = high-altitude pulmonary edema; HACE = high-altitude cerebral edema; AMS = acute mountain sickness.

Sixty-seven percent of all patients were male compared with 53% of control trekkers (p < 0.0001). The ratio of male patients with HAPE (77%) was significantly higher compared to other altitude illness patients (Table 1). Mean patient age was significantly higher than trekker controls (44 ± 13 vs 38 ± 13 y, p < 0.0001; patient age range 15–73 y). HAPE patients were significantly older than other patient groups.

Table 1.  Demographics of altitude illness patients compared with trekker controls
  • HAPE = high-altitude pulmonary edema; HACE = high-altitude cerebral edema; AMS = acute mountain sickness; N/A = not available. Compared with trekker controls:

  • *

    p < 0.005.

  • HAPE, HACE, and combined HACE and HAPE patients compared with AMS patients: **p < 0.001.

  • Prospective group data.

Male gender (%)77*5968*5353.7
Mean age (y)48 ± 12.4* (22–73)44 ± 14.1* (15–70)43 ± 12.8* (19–70)41 ± 14.5 (17–71)38.6 ± 13.9
Body mass index25.1 ± 4.40** (10–35)23.9 ± 1.96** (21–29)23.3 ± 2.47** (19–27)20.4 ± 1.66 (17–22)N/A
Acetazolamide prophylaxis (%)28**29**3843N/A

In the prospective group, we recorded data on the use of acetazolamide prophylaxis, body mass index (BMI), whether trekkers participated in an organized group, and the average carried weight. Thirty-three percent of evacuated patients had used acetazolamide prophylaxis. HAPE and HACE patients reported significantly lower rate of acetazolamide prophylaxis compared with AMS patients (Table 1). The mean BMI was 23 ± 3.2 kg/m2 (17–35), and BMI was significantly higher in HAPE and HACE patients compared with AMS patients. We did not have information regarding the rate of acetazolamide prophylaxis and the average BMI of control trekkers. Eighty-five percent of evacuated patients had trekked with an organized group. The average carried weight carried by a trekker was 7.5 ± 3.7 kg (0–16).

Fifty-three percent of the patients were European citizens, 18% Asian citizens, 16% North American, and 11% from Oceania. A review of the medical history of the patients revealed 24 patients (5.9%) with a history of hypertension, 12 (2.9%) with a history of asthma, 6 (1.4%) who were diabetic, 4 (1%) with a history of a cerebrovascular event, and 4 (1%) with ischemic heart disease. Differences in the prevalence of medical conditions among the diagnostic groups were insignificant.

Comparing the ascent profiles of patients from the different trekking regions yielded significant differences in the starting point altitude. Patients in the Everest region had the highest starting point and the highest maximal altitude attained. The Langtang region had the highest rate of ascent (Table 2). Patients evacuated from the Everest area were significantly older than trekker controls from the Everest region (44.9 ± 13.9 vs 41.9 ± 13.2, p = 0.0011).

Table 2.  Demographic characteristics and ascent profile of patients with severe altitude illness per trekking area
  1. Trekking region compared with Annapurna: *p < 0.05.**p < 0.005.

Male gender (%)56696285
Average age ± SD (minimum–maximum)42.1 ± 12.9 (18–71)44.9 ± 13.9 (15–73)37.8 ± 10.3 (27–63)44.7 ± 12.5 (24–69)
Evacuated by air (%)90846965
Mean starting altitude ± SD (minimum– maximum) (m)1,011 ± 702 (400–3,350)2,760 ± 190** (1,900–3,800)1,981 ± 186** (1,600–2,200)2,650 ± 1,362** (400–5,600)
Mean maximal altitude ± SD (minimum– maximum) (m)4,388 ± 683 (2,700–5,455)4,637 ± 641* (3,440–6,200)4,219 ± 367 (3,500–4,700)5,948 ± 1,093** (3,200–8,845)
Mean ascent rate ± SD (minimum– maximum) (m/d)388 ± 110 (117–579)287 ± 79** (0–500)543 ± 155** (440–810)283 ± 90** (100–383)

During the study period, 21 altitude illness–related deaths were recorded in the six embassies. The mean age of trekkers who died was 50.9 ± 13.7 years (range 31–67). Sixty-eight percent were males. Only five patients had an autopsy that confirmed the diagnosis of altitude illness. When calculating the incidence of death attributed to altitude illness from 1999 to 2005 using death cases reported by the six embassies, the death ratio is 3.8/100,000 trekkers. When taking into account the fact that we had data from six embassies representative of 49% of the trekker population, the actual incidence may approach 7.7/100,000 trekkers.

The majority of trekkers in Nepal hike in the Annapurna region; however, most evacuations and death cases were from the Everest region (Table 3). During the study years, there was a rise in the yearly incidence of altitude illness–related evacuations from the Everest region from 56/100,000 trekkers in 1999 to 276/100,000 in 2005 (linear-by-linear association p < 0.001).

Table 3.  Incidence of altitude illness–related death or evacuation to CIWEC
  • *

    Mortality data from six embassies.

Total number of trekkers 1999–2005366,598147,86742,3496,441
Number of patients evacuated to CIWEC 1999–2005402441341
Evacuation to CIWEC: incidence per 100,000 trekkers10.9165.030.6636.5
Number of death cases 1999–2005*2827
Death of altitude illness incidence per 100,000 trekkers*0.545.414.72108.6

Trekkers in the Everest and Langtang regions were more likely to be evacuated for altitude illness than those who trekked in the Annapurna region [OR 15.1 (95% CI = 10.8–21.1) and OR 2.8 (95% CI 1.5–5.2), respectively]. Same holds true for altitude illness–related death in Everest and Langtang in which the odds ratios were 9.9 (95% CI 2.1–46.7) and 8.6 (95% CI 1.2–61.4), respectively (Figure 2).

Figure 2.

Odds ratio for altitude illness–related evacuation or death per trekking area (Annapurna region comparator). CI = confidence interval.

Many patients presented to CIWEC with an acute illness coexisting with or preceding the development of altitude illness during the trek (Table 4). The two major groups of diagnoses were respiratory infections and diarrhea. Sixty-five percent of HAPE patients were diagnosed with a concomitant respiratory infection, most commonly pneumonia (25%). HAPE patients had significantly more respiratory infections compared to the HACE patients (p < 0.001) and significantly less diarrhea cases (p < 0.05).

Table 4.  Concomitant clinical diagnoses in altitude illness patients, n (%)
  1. HAPE = high-altitude pulmonary edema; HACE = high-altitude cerebral edema; AMS = acute mountain sickness.

  2. HAPE compared with HACE: *p < 0.05.**p < 0.001.

Total136 (100)85 (100)111 (100)74 (100)
Respiratory infections (all combined)88 (65)**13 (15)47 (43)21 (28)
 Pneumonia34 (25)**3 (4)22 (20)1 (14)
 Bronchitis27 (20)**3 (4)10 (9)2 (27)
 URTI27 (20)**7 (8)15 (14)18 (24)
Diarrhea18 (13)*23 (27)23 (21)25 (34)

The time from descent to complete recovery of all symptoms and signs was documented in all patients. In HACE patients, ataxia was the last to resolve with mean resolution time of 41 ± 32 hours (0–144), followed by fatigue, diarrhea, dizziness, and headache (34 ± 27, 25 ± 19, 19 ± 35, and 18 ± 24 h, respectively). In HAPE patients, the last symptom to resolve was cough lasting for a mean of 46 ± 33 hours (0–192), followed by peripheral edema, fatigue, diarrhea, and shortness of breath at rest (37 ± 21, 36 ± 25, 24 ± 19, and 24 ± 24 h, respectively).

At CIWEC, 10% of all evacuated patients were admitted for overnight observation and treatment for dyspnea or severe ataxia (40 of 406 patients). The rest of the patients were discharged after examination confirmed their diagnosis and their ability to take care of themselves.

None of the patients evacuated to CIWEC died, and all symptoms completely resolved by the end of the follow-up period.


Trekking in Nepal attracted an average of 83,000 people per year during the study years. There are three popular trekking regions that attract 97% of all trekkers. The Everest region records 21,000 trekkers per year, the Annapurna region attracts 52,000 trekkers, and the Langtang region has 6,000 trekkers.5 The Everest region is the furthest from Kathmandu and has the highest average terrain. The trekking in this region generally takes place from 2,700 to 5,400 m, with many trekkers spending five to seven nights above 4,000 m. In contrast, the Annapurna region’s trekking begins at altitudes from 800 to 1,300 m, and trekkers may spend only two to three nights above 4,000 m during their trek, with a maximum sleeping height of 4,900 m. Trekking in the Langtang region begins at about 1,800 m and reaches a maximum of 4,300 m. These geographic features probably account for the increased odds ratio of altitude illness–related evacuation in the Everest region: the higher starting altitude, more difficult terrain, and remoteness from Kathmandu. Altitude illness–related death rates were also higher for the Everest region compared with Annapurna (Figure 2).

The current recommended ascent rate of 350 to 400 m altitude gain a day in the Everest region still allows for a 30% to 40% incidence of AMS at Pheriche (4,243 m) and higher incidence in trekkers who flew to Lukla (2,800 m).1 Our study suggests that there could be a benefit in designing even slower itineraries in this region. Trekking in the Langtang also carried higher incidence of death or evacuation compared with the Annapurna. The increased risk in the Langtang region may be due to a rapid ascent rate, two times faster than the rate in trekkers evacuated from the Annapurna region (Table 2). Mountaineering carries the highest risk in our study, as was previously shown.6,7

The number of helicopter rescues of patients in this study (311 patients) reflects an incidence of 55.8/100,000 trekkers. A previous study in Nepal during the years 1984 to 1987 reported 25.6 helicopter evacuations due to altitude illness per 100,000 trekkers.8 The incidence of helicopter evacuation from the Everest region (from where most of our patients were evacuated) rose significantly during our study (1999–2005) from 56 to 276/100,000 trekkers. The rise in the incidence of helicopter rescues may reflect the increased availability of helicopter rescue. Aerial evacuation carries a substantial cost and risk to both evacuated trekkers and rescue teams. Our study recorded approximately 45 helicopter rescues per year for altitude illness, a figure that likely underestimates the true amount because not all evacuated patients were examined by us. Even using this conservative number, the financial impact of helicopter rescue for altitude illness is large: at $5,000/rescue, about $225,000 per year is spent on helicopter transport to lower altitudes. Trekkers are also evacuated for trauma and nonaltitude-related illness.8

Two previous studies describing the incidence of altitude illness–related death in Nepal found a rate of 2.02 and 3.62/100,000 trekkers in 1984 to 1987 and 1987 to 1991, respectively,8,9 compared with our estimation of 7.7/100,000 trekkers in our study. Thus, altitude illness–related deaths have in fact significantly increased in 15 years (linear-by-linear association p = 0.05) despite the rise in helicopter rescues and the advances in communications.

The mean age of trekkers who died of altitude illness rose significantly from 35 ± 13 and 44 ± 16 years between 1984 to 1987 and 1987 to 1991, respectively, to 50.9 ± 13.7 years in our study (correlation test p = 0.048). This may reflect the fact that the average age of trekkers is increasing and the possibility that older trekkers are more likely to die from altitude illness.

Several previous series of trekkers in Nepal compared altitude illness patients to control trekkers: Hackett and Rennie1 showed that older age was protective against AMS (35.2 ± 13.3 vs 31.4 ± 9.7, p = 0.029). The mean age of trekkers in their study (Everest region) was 33 years. Murdoch3 surveyed 283 trekkers in the Everest region with a mean age of 31.5 years and found a negative correlation between AMS score and age. However, in our study, the mean age of trekker controls in the Everest region was higher (41.9 ± 13.2 y), and HACE and HAPE patients evacuated to CIWEC were significantly older than trekker controls. This may be due to the fact that older people are at higher risk of severe altitude illness, but it is also possible that older trekkers are more willing to evacuate themselves by helicopter rescue and therefore seen more often in our clinic. A study conducted in the Alps showed that older age and male gender increased the risk of altitude trekking–related death.10

HAPE patients were significantly more males compared with trekker controls. Several studies have previously shown higher frequency of HAPE in men.11–14 Other studies showed that the rates of women developing nonsevere altitude illness were higher than the rates of men.2,15

The BMI of HAPE and HACE patients in our series was higher than AMS patients. Obesity was previously described by several authors as a risk factor for altitude illness.2,16–18 It might also place patients at risk for more severe forms of altitude illness.

It is notable that only 33% of trekkers evacuated due to altitude illness used acetazolamide prophylaxis. Due to lack of data regarding the rate of prophylaxis use in trekkers, we were unable to determine whether the low rate in altitude illness patients was a risk factor. A recent study done in Nepal and northern India has shown that only 40% carried acetazolamide with them and less than one third of travelers trekking at altitude listed the drug as a preventative option for AMS.19 Possibly, low rates of acetazolamide prophylaxis contribute to the relatively high incidence of altitude illness evacuations.

Trekking with an organized group is common in Nepal. In the Everest area, 94% of evacuated patients were trekking with a group compared with 46% of the trekkers in the Annapurna region. It might be that more trekkers in the Everest area participate in organized tours, but we were not able to obtain these data. Group trekking is characterized by a rigid ascent schedule and peer pressure to continue the trek despite fatigue or illness. This might be another factor increasing the risk seen in the Everest region.

HAPE patients in our study suffered from a high rate of respiratory infections. In a study of respiratory infections among hikers in the Everest region, Murdoch3 found a higher incidence of AMS among those with symptoms of infection. Durmowicz and colleagues20 studied the incidence of various upper and lower respiratory tract infections in children with HAPE in Colorado. They found symptoms of infection preceded HAPE in 79% of the study subjects, and of them, 54% suffered from upper respiratory tract infections. Although it might be difficult to differentiate respiratory symptoms caused by infection from those caused simply by HAPE, we believe that the presence of an infection might predispose trekkers to altitude illness.

In our group, all symptoms completely resolved within 1 to 2 days of evacuation, and none of the patients suffered from permanent sequelae. Of 44 HACE patients reported by Dickinson21 in 1983, 10 recovered in 2 days or less after evacuation, 32 were in the hospital for a time ranging from 3 to 14 days, and 2 patients remained unconscious for over 3 weeks. Hackett described nine HACE patients with a varied hospital course with an average of 5.6 days of hospitalization. Time to normal neurologic examination averaged 2.4 weeks (range 1–6).22 The shorter recovery rate in our study population may be due to the fact that our patients were rapidly treated and descended (shown by the high rate of air evacuation in our series) as opposed to previous series where patients may have spent longer time in altitude or arrived to treatment later in the course of disease. Shorter recovery time may represent the natural course of severe altitude illness treated with rapid evacuation soon after diagnosis. A limitation of our study is that we did not have information about patients admitted to other hospitals, in whom course of recovery might have been longer.

Severe altitude illness is preventable in trekkers, and early recognition of symptoms can allow the trekker to either rest for a day at the same altitude or descend a short distance to recover. The difference between a successful trek and having to be evacuated for severe altitude illness is often just one day’s difference in the acclimatization schedule. It would appear that efforts to educate trekkers are still necessary to avoid mortality and potentially dangerous rescue situations.


We thank Dr Gad Segal for his valuable assistance in preparing the manuscript.

Declaration of interests

The authors state that they have no conflicts of interest.