The protocol was approved by the Ethics Committee of the General Hospital of the Chinese People's Liberation Army, and conformed to standards set by the Declaration of Helsinki. Written informed consent of all participants was acquired.
Apparently healthy male subjects who had relocated to Lhasa (3658 m) from plain districts for more than 2 years were recruited from communities by word-of-mouth. The high-altitude immigrant subjects (HA group) were screened for the following criteria: (1) permanent residence before relocating to HA at less than 1500 m; (2) no history of ischemic hypoxic encephalopathy as infants; (3) no serious brain injuries that resulted in loss of consciousness; (4) no tobacco or drug abuse or alcohol addiction; (5) no chronic or genetic diseases; (6) Qinghai score < 6 to exclude chronic mountain sickness; (7) a willingness to participate in the study and sign the informed consent form; (8) male; (9) right-handed; (10) 20 to 40 years; (11) immigrated to HA after reaching 18 years of age.
A group of sea-level-resident healthy male subjects (SL group) were included as a control group. SL subjects were recruited from communities near our hospital, and the screening criteria were the same as the HA group except for (6). The age of SL subjects was matched to HA subjects. Exclusion criteria consisted of contraindication for MRI.
Because headache can occur intermittently and its severity can change over time, the participants were asked to recall and rate the presence and extent of headache according to four levels (no headache, mild, moderate, and severe headache) during the past month. The headache score was defined as follows: 0, no headache; 1, mild headache; 2, moderate headache; 3, severe headache. According the headache score, the HA group was divided into two sub-groups: the subjects with a headache score of zero were classified as the non-headache group (HA[-]), and the subjects with a headache score greater than zero were classified as the headache group (HA[+]).
All participants were imaged on two identical 3.0T MR scanners (Discovery MR 750, GE Healthcare, Milwaukee, WI, USA) in the General Hospital of Tibet Military Region (Lhasa, 3658 m, HA group) and the First Medical Center of the General Hospital of the Chinese People's Liberation Army (Beijing, 50 m, SL group), respectively, by using two identical 8-channel head coils. Magnet hardware and software remained unchanged during the study. Given the time difference, the imaging of HA group was performed between 3:00 pm and 6:00 pm Beijing time, and the imaging of SL group was performed between 1:20 pm and 4:20 pm Beijing time, to minimize the influence of physiological rhythms. Participants abstained from alcohol, caffeine, tea (including Tibetan buttered tea), strenuous exercise, gluttony, and any medication for at least 24h before the MRI scan to minimize potential confounding eﬀects on brain volume and ICP. The three-dimensional fast spoiled gradient recalled echo (3D-fSPGR) sequence was used to acquire high resolution structural images. The parameters were as follows: repeat time=6.9ms, echo time=3.0ms, bandwidth=±31.25kHz, field of view=25.6cm, slice thickness=1mm, matrix =256 × 256, number of excitations=1, number of slices=192, acquisition time=4 min 47s.
Pituitary gland assessment
3D-fSPGR sagittal images of the pituitary gland were corrected for head tilt or rotation using a three-dimensional multiplanar reconstruction tool on the GE Advantage Workstation (AW) system. Pituitary gland height (PGH) was measured on mid-sagittal T1 images as the maximum orthogonal distance from the upper surface of the pituitary gland to the sella floor (Figure 1). PGD was assessed visually using the following grading scale modified from a previous report : normal appearance (convex and ﬂat dome of the pituitary gland, 0 points); mild PGD (loss of pituitary height h ≤ ⅓ of the sella height, 1 points); moderate to severe PGD (loss of pituitary height h >⅓ of the sella height to empty sella, 2 points) (Figure 2).
PGH measurements were performed by two radiologists (J. F. and X. Y., with 12 and 3 years of imaging experience, respectively). First, images were opened by X.Y. and were anonymized; then, PGD and PGH were assessed by J. F. Two weeks later, the two raters exchanged roles and PGH was assessed again. Image reformation and head position correction were independently performed by each radiologist to prevent any bias introduced by the other radiologist. Radiologists were blinded to clinical data and each other’s measurements during PGH assessments. The averaged value of PGH measured by the two radiologists was used for statistical calculations. Until the PGH measurements were completed, the extent of PGD was assigned points by consensus interpretation according to the PGD by the two radiologists. Disagreements were reviewed until consensus was achieved.
Brain to intracranial volume ratio calculation
Brain to intracranial volume ratio of HA group was calculated to assess the “tight fit” extent of intracranial system. All brain structure images were converted to NIFTI files using MRIcroN software (University of Nottingham School of Psychology, Nottingham, UK; www.mricro.com). The intensity of brain structure images was corrected using the N4 algorithm. SPM12 (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/) was employed to segment the brain images into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) using default parameters. Using the quantifications of gray matter volume (GMV), white matter volume (WMV), and total intracranial volume (TIV), BV/TIV was calculated as：BV/TIV = (GMV+WMV)/TIV.
Data analysis was performed using MedCalc Statistical Software version 19.0.4 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org; 2019). We report statistical significance and descriptive statistics. Continuous variables were assessed using mean ± standard deviation or median (IQR), and categorical variables were summarized with percentages. A two-tailed critical α was set to reject the null hypothesis at 0.05. The intraclass correlation coefficient (ICC) was used to assess the reliability of PGH measured by a single rater and different raters (J. F. and X. Y.). The ICC of the HA and SL groups were evaluated separately. The independent samples t-test or Mann-Whitney test was used to compare the differences between groups. The Cochran-Armitage test for trend was used to quantify whether there was a linear trend of the proportions of participants with PGD among SL, HA+ AND HA- groups. Fisher's exact test was used to perform pairwise comparisons of proportions of participants in each group. In addition, Kendall's tau rank correlation was used to investigate the relationship between PGD and the headache, and Spearman’s rho was used to calculate the correlation between PGD and the ratio of BV/TIV. Receiver operating characteristic (ROC) analysis was performed to evaluate whether PGD is a potential indicator of headache in HA immigrants.