This prospective observational cohort study involving 24-h ABP monitoring was conducted in Chengdu, China, in 2019. All subjects underwent a comprehensive medical examination before embarking on the expedition at LA (Chongzhou, 400 m above sea level) and underwent ABP monitoring and echocardiography both at the LA and HA (Litang, 4100 m above sea level). The exclusion criteria were as follows: (1) Any clinical conditions that may affect HA-adaptation, including known pulmonary diseases, cardiovascular diseases, hematological diseases, and so on. (2) Use of any medication. (3) Long term residence history and recent exposure history of altitude (last 6 months). Notably, subjects diagnosed with hypertension first according to the ABP data at LA (24-h BP > 130/80 mmHg and/or daytime BP > 135/85 mmHg and/or nighttime BP > 120/70 mmHg) (13) and 24-h ABP monitor recordings of <80% of the total data recorded (14) were excluded. Finally, 69 subjects were included, all of whom provided written informed consent. This study was performed in agreement with the Declaration of Helsinki and was approved by the Human Ethics Committee of the Xinqiao Hospital, Third Military Medical University (Identification code, 201907501) which registered at www.chictr.org.cn (ChiCTR-TRC-No.1900025728).

Two well-trained cardiovascular physicians performed ABP measurements using an ABM device (Spacelabs 90207, Redmond, WA, USA). The BP cuff was applied to the non-dominant arm on a weekday morning and was removed 24 h later. All participants were asked to remain still during the measurements. The subjects were instructed to avoid any unusual physical activities and follow a standard schedule at both LA and HA. Daytime and nighttime were defined as 6:00 to 22:00 and 22:00 to 6:00, respectively (11). The BP was recorded every 30 min during the daytime and every 60 min at night (15). The average real variability (ARV) of SBP and DBP was calculated as described in a previous study (16). BP load is the proportion of times that the BP exceeds normal values of the total number of recorded BP measurements during a certain period of time (day, night, and over 24 h) (17). To more accurately reflect the pressure load of the blood vessels, we calculated the cumulative BP load (cBPL), which is defined as the area between the fluctuating ABP curve and the time axis (11). Fitting the fluctuating BP curve by connecting adjacent data points with straight lines was used to determine the magnitude and durations of 24-h cumulative BP increases. By deconstructing the curve into many small trapezoids, we determine their areas, and summed the values. Due to missing data for some periods, we calculated the mean cBPL by dividing the total area by the number of periods measured. The interval between BP measurements at night (1 h) is twice as long as that during the day (30 min) (Figure 1); therefore, the number of periods measured doubled in the calculation of mean cBPL.

Echocardiographic examination was performed using an ultrasound machine (CX50, Philips Ultrasound System, Andover, MA, USA) to acquire left ventricular (LV) data. Images were saved digitally for subsequent offline analysis using QLAB software (QLAB 10.5, Philips Healthcare, Andover, MA, USA). We measured the LV dimensions and volumes by a computerized analysis software system. Ejection fraction (EF) was calculated by the LV volume data. Mitral inflow pattern from the tips level was analyzed for peak early diastolic velocity (E) as well as late diastolic velocity (A), and E/A. Reproducibility of main echocardiographic measurements was assessed in 20 randomly selected subjects. Interobserver variability was tested by two different physicians, and intraobserver variability was tested by the same physician at least 1 month apart. Both the interobserver and intraobserver variabilities were determined using the intraclass correlation coefficient (ICC). The ICC values were all over 0.85 and p-value < 0.001.

All subjects traveled by automobiles from LA to HA within 2 days (Figure 2). AMS was diagnosed using the latest Lake Louise questionnaire (2018) ~8 h after arriving at HA. This comprises a four-item self-administered questionnaire based on the most frequent AMS symptoms: headache, dizziness, light-headedness, gastrointestinal symptoms, and fatigue. Participants completed the questionnaire with the assistance of an experienced physician. Each item is scored from 0 to 3, according to the severity of the symptom (0: no symptoms, 1: mild symptoms, 2: moderate symptoms; and 3: severe symptoms). AMS was defined as a total score ≥ 3, with at least one point from the headache (18).

Continuous variables are presented as means ± standard deviations. Differences in measurements between men and women with normal distribution were tested using an independent-sample t-test, whereas the data that did not fit a normal distribution were analyzed using the Mann–Whitney U test. Categorical data were presented as percentages (%) and were compared using the chi-square test, continuity correction, or Fisher's exact test, as appropriate. Binary logistic regression was used to predict the risk factors of AMS. Spearman correlation coefficients were used to determine the correlation between the different BP index variations after acute HA exposure and AMS scores. Statistical significance was set at P < 0.05. Statistical analyses were performed using SPSS software Version 26 (IBM, Armonk, NY, USA). Statistical power calculations were performed using the PASS software, version 11 (NCSS, LLC, Kaysville, UT, USA). The results suggested that 69 subjects would provide more than 75% power to detect differences in targeted BP parameters between subgroups using a two-sided alpha of 0.05.

Twenty-three subjects developed AMS in the final data analysis. Age and BMI did not differ significantly between the two groups. The AMS group comprised a higher proportion of women (65.2 vs. 28.3%, P = 0.001) but a lower proportion of smokers (17.4 vs. 50.0%, P = 0.009). The AMS score and the percentage of AMS-related symptoms were also significantly higher in the AMS group (Table 1).

Patients with AMS had a higher nighttime heart rate (HR) (62.71 ± 8.29 vs. 57.58 ± 7.71 beat/min, P = 0.017) and a lower daytime BP (daytime SBP, 114.94 ± 10.77 vs. 122.60 ± 6.50 mmHg, P = 0.004; daytime DBP, 69.51 ± 5.83 vs. 72.58 ± 4.55 mmHg, P = 0.019) at LA. Both 24-h SBP load and daytime SBP load were lower in the AMS group than in the non-AMS group at LA (24-h SBP load: 15.75 ± 9.53 vs. 21.63 ± 10.25%, P = 0.025; daytime SBP load: 15.67 ± 11.35 vs. 23.09 ± 11.45%, P = 0.013). Moreover, 24-h mean cSBPL (92.81 ± 10.36 vs. 98.47 ± 7.86, P = 0.014), daytime mean cSBPL (114.98 ± 11.34 vs. 122.70 ± 57.30, P = 0.001), and daytime mean cDBPL (69.15 ± 6.02 vs. 72.75 ± 4.64, P = 0.008), which were calculated using the BP-time curve area, were higher in non-AMS participants at LA. Besides, there was no obvious difference in the ARV and cardiac function between the two groups at LA.

After arriving at HA, AMS subjects showed a lower SpO2 at HL, but no statistical difference was found in SpO2 variation between the two groups. Both daytime and nighttime BP increased after acute HA exposure in each group (Figure 3). 24-h BP increased significantly in the AMS group compared to the non-AMS group (24-h SBP variation: 10.52 ± 6.48 vs. 6.03 ± 9.27 mmHg, P = 0.041; 24-h DBP variation: 8.70 ± 4.57 vs. 5.03 ± 4.98, P = 0.004) which attributed to a higher elevation of daytime BP in subject with AMS (daytime SBP variation: 12.68 ± 8.91 vs. 6.46 ± 11.12, P = 0.023; daytime DBP variation: 9.15 ± 5.41 vs. 5.05 ± 6.35 mmHg, P = 0.010) (Table 2). 24-h DBP load variation (19.98 ± 11.42 vs. 11.21 ± 5.41%, P = 0.011) and daytime DBP load variation (19.71 ± 10.30 vs. 10.81 ± 6.66%, P = 0.008) were significantly higher in the AMS group compared to the non-AMS group. Furthermore, 24-h mean cBPL variation (24-h mean cSBPL 10.58 ± 10.99 vs. 4.02 ± 10.58, P = 0.016; 24-h mean cDBPL: 6.03 ± 5.87 vs. 2.89 ± 4.99, P = 0.034) and daytime mean cBPL variation (daytime mean cSBPL: 15.55 ± 13.99 vs. 6.77 ± 11.59, P = 0.007; daytime mean cDBPL: 9.00 ± 5.66 vs. 4.63 ± 6.97, P = 0.011) were also obviously higher in AMS subjects than in non-AMS subjects (Table 2). In addition, there was no obvious difference in the ARV and cardiac function between the two groups (Table 2).

We looked for the risk factors of AMS which associated with the BP changes. After adjusting for age, sex, BMI, and smoking status, the results of the regression analysis showed that 24-h SBP variation (OR = 1.08, P = 0.036), 24-h DBP variation (OR = 1.21, P = 0.007), 24-h SBP load variation (OR = 1.05, P = 0.010), 24-h DBP load variation (OR = 1.06, P = 0.016), 24-h mean cSBPL variation (OR = 1.07, P = 0.024), 24-h mean cDBPL variation (OR = 1.14, P = 0.034) were independent risk factors of AMS (Table 3).

In order to more effectively evaluate the BP status of individuals and the relationship between BP changes and AMS during HA exposure, we analyzed the correlation between the changes in some BP indicators and the AMS. The variations in BP load (24-h SBP load variation, R = 0.369, P = 0.001; 24-h DBP load variation, R = 0.310, P = 0.005) as well as mean cBPL (24-h mean cSBPL variation, R = 0.504, P < 0.001; 24-h mean cDBPL variation, R = 0.290, P = 0.007) were also correlated with AMS score; which indicated that the BP load could also reflect BP status at HA and was associated with AMS (Figure 4; Supplementary Figure 1).

BP remained largely unchanged over the first few minutes or hours after HA exposure. However, the BP increased remarkably and persisted with prolonged altitude exposure over the next few days. The rise appeared to be continuous and proportional to altitude. It was particularly pronounced at night, which resulted in a reduction in the decline in nocturnal BP; the circadian rhythm of 24-h BP fluctuations disappeared. BP primarily returned to the baseline level when on the return to sea level (18, 19). While in hypertensive subjects, a higher hypoxia-driven upward shift and steepening of the BP response to exercise were observed (20, 21). Initially, the secretion of vasodilator factors, such as nitric oxide, could counteract the effect of sympathetic activation (22). Subsequently, pressor mechanisms begin to dominate. The enhanced sympathetic nerve activity caused by hypobaric hypoxia is related to the sympathetic reflex stimulated by chemoreceptors in the carotid sinus and aortic arch, which stimulates the release of catecholamines. Mazzeo and Reeves have also confirmed this point that while the activity of the renin-angiotensin-aldosterone system was inhibited, the increase in vasoconstrictor factor secretion results in the elevation of peripheral vascular resistance caused by contraction of the smooth muscle arterioles (23). However, difference in peripheral α-adrenergic sensitivity to norepinephrine between sexes may produce different hypoxia-induced BP changes (24). Factors including endothelin-1 levels and erythropoietin are also potentially involved (25). Although cardiac diastolic relaxation was decreased, an elevated systolic function may result in BP increase (26).

The BP load was introduced by pioneering studies published in the early 1990s (27). This may provide supplementary information for the diagnosis of hypertension in individuals when 24-h ABP would be misleading (27). At present, in addition to the mean BP, the degree of BP fluctuation has garnered increasing attention. Further, owing to the limited information provided by calculating the number of readings exceeding the normal BP threshold levels, several investigators have proposed a method for calculating the area under the curve (28). This reflects the extent and duration of the cumulative rise in BP, and may provide information that more closely reflects the actual BP status. In addition, other indicators that assess BP fluctuations, such as ARV, exhibited little independent prognostic significance (29). Previous studies suggested that the BP load was also closely associated with signs of target organ damage and adverse clinical cardiovascular events as the BP level (30, 31). Meanwhile, some studies have recommended to combining BP level and BP load to define hypertension and assess BP status (32). As no additional predictive value for the risk of target organ damage or cardiovascular complications after average the BP is considered (9, 33). However, the application of BP load to demonstrate BP status remains to be confirmed in extreme environments, and it may possess unique significance for the prediction and evaluation of HA cardiovascular disease.

The relationship between changes in BP after altitude exposure and AMS remains controversial in previous studies. Although study has shown that BP remains relatively safe and stable at altitudes without to be symptomatic (34), it was also previously demonstrated that the BP elevation in AMS is relatively obvious after the acute plateau (35), and that higher BP at HA is related to the onset of AMS (7). Our previous study also revealed that the increase in BP during a certain period is related to sympathetic hyperexcitation, and excessive morning BP surge is associated with the onset of AMS (36). However, in some large size studies, increased BP after acute HA exposure has not been confirmed to be associated with AMS (37), and even a medical history of hypertension may be associated with a lower risk of AMS (37, 38). These conflicting results may be mainly due to individual differences among the populations included in the studies. For example, increased arterial thickness, decreased production of the vasodilator nitric oxide cerebral and lower blood flow might enable hypertensive patients less susceptible to AMS (15, 18). Thus, we excluded individuals with potential hypertension in this cohort study, which may have affected our observed results at HA in order to accurately reflect the relationship between BP changes during HA and the incidence of AMS. Because single BP measurement is not a good reflection of systemic circulation status, we also performed ABP measurement. Previous studies have confirmed some AMS risk factors. And certain factor, such as sex difference, may have an effect on BP changes during hypoxia (39, 40). We performed multivariate logistic regression and found BP load variation was independent risk factor of AMS after adjusting. Besides, we also demonstrated that there was a linear relationship between BP level variation and AMS although the degree of correlation was weak. Few previous reports have analyzed the relationship between high BP load and the risk of cardiovascular disease at HA. The excessive BP load is primarily due to the sympathetic excitation of the body (21). Moreover, AMS was associated with autonomic nervous dysfunction in previous studies. Patients with AMS development may accompanied with a disturbed BP control and an insufficient vasoconstriction function during altitude exposure (41, 42). And hyperactivation of the short-sympathetic after the acute plateau may be the main mechanism of pathogenesis. AMS may be induced by the elevation of BP after HA exposure, because an excessive rise in BP might result in symptoms, such as dizziness and headaches (36, 43). Therefore, we believe that BP load might be a good indicator of systemic circulation status assessment potentially during HA exposure. Our results also confirmed that changes in the BP load calculated using different methods are associated with AMS. Especially 24-h mean cSBPL showed a better correlation with AMS score compared with other ABP parameters. However, we also found that the variation in BP variability, which also reflects the degree of sympathetic excitation, was poorly associated with AMS. BP load may be a better indicator that reflect systemic circulation status in individuals with AMS at HA. A larger population study is required to confirm this point in the future. Notably, the explicit pathophysiology of AMS still remains unclear now. Previous study has proven that hypoxia could induce the blood-brain barrier disruption and lead to the development of cerebral edema subsequently. An increased brain volume with hypobaric hypoxia elevated intracranial pressure and impaired intracranial buffering capacity, which contributes to the development of the symptoms that define AMS (44). Sympathetic excitation causes the increase of systemic circulation BP load, cerebrovascular may also appear similar hemodynamic changes related to cerebral edema. But unfortunately, the relevant data were not tested. But at least, clinical evidences of the autonomic nervous system effects on AMS development could stimulate future research.

There are several limitations to the present study. Most importantly, due to the limitations of the HA conditions in this study, the BP data were not as complete as those of some studies conducted at LA, although the ABP recordings were still more than 80% of the total ABP data recorded. Therefore, we adjusted the calculation method of the cBPL by calculating the mean cumulative BP load. Notably, there might be a superior method of calculating the area under the fitting curve that is above the hypertension threshold (30). However, we did not use this method, because we were constrained by the discontinuity of the data. Secondly, the diagnosis of AMS was based on self-report, which could have led to classification bias. Moreover, we did not assess sleep status. Sleep may be disturbed during night ABP measurement. In addition, we did not examine some important indexes which could reflect the potential mechanism of AMS, such as cerebral hemodynamic monitoring. Lastly, different demography, such as race, smoking and drinking history may affect the established results. This conclusion is still need to be examined.