Significantly Larger Hypothalamic Volume in Narcolepsy Type 1 Following H1N1 Infection

Hilde T. Juvodden, MD

In a recently published study in Sleep, post-H1N1 infected patients with narcolepsy type 1 (NT1) demonstrated a significantly larger hypothalamic volume compared with the controls, in particular, bilaterally in the tubular-inferior hypothalamic subregions.¹ These findings suggest that neuroinflammation or increase in the number of histaminergic neurons in NT1 in the subregions affected could possibly be associated to gliosis.

Among 54 patients with NT1 (women, n = 39, mean age, 21.8 ± 11.0 years) and 114 controls (women, n = 77, mean age, 23.2 ± 9.0 years), findings showed a larger volume post-H1N1 compared with controls for the whole hypothalamus (Cohen's d = .71, P =.0028). Notably, a larger volume was also reported for the left (d = .70, P =.0037) and the right part of the hypothalamus (d = 0.65, P =.0075) as well as the left (d =.72, P =.0036) and the right tubular-inferior (d =.71, P =.0037) hypothalamic subregions.

“We believe our findings shed light on the disease processes occurring in the hypothalamus of NT1. Moreover, if structural brain changes are shown to be specific for hypocretin-deficient narcolepsy this could also in the future have potential to be used for noninvasive diagnostic purposes,” lead author Hilde T. Juvodden, MD, postdoctoral researcher at NevSom, Oslo University Hospital, in Ullevål, Norway, and colleagues wrote.¹ In this study, investigators compared the MRI-based volume of the hypothalamus among patients with NT1 and the controls in vivo.

Researchers utilized a segmentation tool based on deep learning to compute the volume of the whole hypothalamus, left/right part of the hypothalamus and 10 hypothalamic subregions. Differences among the groups were assessed with general linear models using permutation testing in Permutation Analysis of Linear Models. Investigators evaluated after 10,000 permutations, yielding 2-tailed p-values, and conducted a stepwise Bonferroni correction following the division of the into smaller regions.

In permutation testing, findings did not display any group differences for the left (d = 0.13, P =.60) and the right anterior–superior (d =.19, P =.43), left (d = 0.41, P =.095) as well as the right anterior–inferior (d =.43, P =.080), left (d =.41, P =.099) and right posterior (d =.15, P =.55) and left tubular-superior (d =.26, P =.28) subregions. Similarly, in a multiple linear regression analysis of disease duration and hypothalamic volumes, there were no significant results for the whole hypothalamus (d =.63), left (P =.55) or right part of the hypothalamus (P =.75) or the left (P =.49) or right tubular-inferior (P =.77).

Patients with comorbidities and a score of at least 5 on apnea-hypopnea index were included in the main analyses. In the subanalyses, these patients were excluded even though they produced similar results as the main analysis (36 vs. 54 patients). In a follow-up analysis after permutation testing, patients showed larger volume compared with the controls, for the whole hypothalamus (d =.57, P =.03), right part of the hypothalamus (d =.60, P =.027), left part of the hypothalamus (d =.47, P =.077), left tubular-inferior (d = 0.52, P =. 059) and right tubular-inferior (d =.61, P =.020) subregions.

“There was no significant correlation between the hypothalamic volumes (whole hypothalamus, right/left part of the hypothalamus, right/left tubular inferior) and disease duration. However, exploring correlations with disease duration in our sample has limitations as our sample is quite homogenous including for the mean disease duration 5.8 ± 1.4 years (range 9.15 years), this should therefore be further assessed in a multi-center study to obtain larger and more heterogenous samples,” Juvodden et al noted.¹

Authors noted that a limitation was not ideally being able to modify the hypothalamic segmentation into the most interested subregions, yet the automated segmentation was based on visible and reliably segmented anatomical landmarks.² Therefore, investigators noted that they will have to perform this automatic hypothalamic segmentation among sporadic NT1 patients to clarify and generalize the findings. In addition, the researchers recognize that their approach with MSLT could possibly lower the sleep pressure, but in their NT1 sample, the average SOREMs and average sleep latency on MSLTs were comparable with previous research.³

“There are several possible explanations for the increase in volume as previous studies have indicated neuroinflammation, gliosis, and increase in the number of histaminergic neurons in the hypothalamus of patients with NT1,” Juvodden and colleagues noted.1 “Another possible explanation could be that the current finding of larger volume is specific to H1N1-vaccinated narcolepsy patients as 88.9% of our patients were H1N1-vaccinated (although 3 of the H1N1-vaccinated patients had disease onset prior to vaccination).”

REFERENCES
1. Juvodden HT, Alnæs D, Lund MJ, et al. Larger hypothalamic volume in narcolepsy type 1. Sleep. 2023;zsad173. doi:10.1093/sleep/zsad173
2. Billot B, Bocchetta M, Todd E, Dalca AV, Rohrer JD, Iglesias JE. Automated segmentation of the hypothalamus and associated subunits in brain MRI. Neuroimage. 2020;223:117287. doi:10.1016/j.neuroimage.2020.117287
3. Arand DL, Bonnet MH. The multiple sleep latency test. Handb Clin Neurol. 2019;160:393-403. doi:10.1016/B978-0-444-64032-1.00026-6