The Lake Louise Acute Mountain Sickness Score: Still a Headache

Acute exposure to altitude triggers, in many people, a variety of symptoms that are collectively known as acute mountain sickness (AMS). In 1991, experts from 13 countries met at the 7th International Hypoxia Symposium held at Lake Louise, Canada, to create an abbreviated scoring system for research purposes to diagnose AMS, known as the Lake Louise AMS Score (Hackett and Oelz, 1992).

The original score (Hackett and Oelz, 1992) is based on a self-administered questionnaire for the assessment of five variables (symptoms, here referred to as items) grouped within a common latent variable (AMS, here referred to as factor). The score items include headache, dizziness/lightheadedness, fatigue/weakness, gastrointestinal symptoms, and difficulty sleeping. Each item has a graded score from 0 through 3, depending on the severity of the symptoms. It was agreed that a score ≥3, assessed after the ascent to altitude, in the presence of a headache and at least one of the other symptoms, is diagnostic of AMS. More than 25 years later, experts from 21 countries reached a consensus and published an updated version (Roach et al., 2018). The 2018 version dropped the “difficulty sleeping” item. It was agreed that a score ≥3, assessed at least 6 hours after exposure to altitude or hypoxia, in the presence of a headache, is diagnostic of AMS. These modifications have revived a hot debate.

In this issue of High Altitude Medicine & Biology, two large and interesting studies conducted in China (Chen et al., 2021) and France (Richalet et al., 2021) re-examined the Lake Louise AMS Score using factor analysis, a statistical method used to evaluate the validity (accuracy) of a model and estimate the internal consistency (reliability) among items within the model.

Chen et al. (2021) conducted a retrospective study in 1,026 male soldiers aged 18–45 years old who ascended to 3,700 m. The original and the updated score versions were assessed 24 and 48 hours after the ascent (with ∼45% subjects lost to follow-up). Both versions had a good validity (as indicated by a relatively high factor loading for each item) and acceptable internal consistency (as indicated by a relatively high alpha coefficient for the factor) at 24 hours. At 48 hours, both score versions had an acceptable validity and internal consistency. In the 2018 version, headache and dizziness had the highest factor loadings regardless the timing of the assessment. The authors concluded that difficulty sleeping might contribute to AMS diagnosis.

Although the factor loading for difficulty sleeping item was acceptable in the study, the correlations of difficulty sleeping with headache and dizziness were the weakest, which is consistent with the findings from a previous study (Macinnis et al., 2013) and those from the article by Richalet et al. (2021). Collectively, the latter findings argue against the suitability of difficulty sleeping within the AMS score. Information on how much the alpha coefficient varies when a given item is deleted would be valuable to clarify this issue. Remarkably, in the article by Chen et al. (2021), 50% of the subjects were smokers and all were male, which may limit the generalizability of the findings.

Richalet et al. (2021) conducted a retrospective study in 484 male and female trekkers aged 13–87 years old who reached variable altitudes (2,735–8,200 m). Data were presented in terms of percentage of subjects lost in the diagnosis of AMS after the inclusion or exclusion of a given item. The authors recommended that headache should not be mandatory to define AMS and also concluded that difficulty sleeping contributes to the diagnosis of AMS.

The study by Richalet et al. (2021) does not provide information on the factor loading of each item for the entire study population. It is also unknown how much the alpha coefficient varied when a given item was deleted from the factor. However, the main limitation of the study is that the analysis included data obtained at different points in time. This approach makes it difficult to interpret the findings and introduces selection biases. Another limitation is that the range of altitude exposure was considerably wide. Naturally, symptoms will be more severe as altitude increases, and they will subside in the following days or may progress to cerebral or pulmonary edema.

From the data presented by Richalet et al. (2021), one can infer that the most common symptom related to altitude exposure is headache, as in the article by Chen et al. (2021). Thus, there appears to be no strong rationale, from a clinical perspective, to exclude the headache item as a requirement for the diagnosis of AMS. In contrast, the weak internal correlation of difficulty sleeping with headache could indicate that the former item does not accurately measure AMS, an issue that should be more carefully studied. So far, the existing scientific data do not appear to justify changes to the 2018 version of the Lake Louise AMS Score. Unless there is strong collective evidence to support that headache is not required to diagnose AMS, headache should remain mandatory for the diagnosis.

Since there is no gold-standard method to diagnose AMS, further attempts to modify the Lake Louise AMS Score will probably lead to more debates rather than a more consensual scoring system. Much is still pending, including a consensus on the score cutoffs to classify the severity of AMS and altitude cutoffs to define “mountain” sickness.

Future studies aiming to carefully evaluate the validity and internal consistency of the Lake Louise AMS score should design a prespecified timing of the assessment of the score and select a reasonable altitude range of exposure using a large population. It should be noted that due to the subjective nature of the items, sequential assessment of the score to the same subject could introduce learning bias. In my opinion, it would be advisable to collect information on pre-existing symptoms listed in the 2018 version. A pre-existing symptom immediately before the ascent to altitude should invalidate the score.

It is hoped that the two articles published in this issue (Chen et al., 2021; Richalet et al., 2021) will stimulate further research on this important, exciting, and controversial topic.