Materials and methods
This prospective single-center, randomized controlled pilot study was conducted in accordance to the Ethics Committee (Blinded for review) (ID 170728010).
In this study, we included two cohorts of subjects: healthy subjects, and patients with post-surgical olfactory dysfunction. Healthy patients were recruited voluntarily at the ENT clinic to assess fMRI testing paradigms and as a control group for the olfactory dysfunction group. Three healthy subjects with normal olfactory function participated in the study (two males and one female, with mean age 47 years, range 27 – 60). All 119 patients that underwent a transsphenoidal pituitary surgical approach were contacted. Forty-six patients that referred a smell impairment with onset after the surgical procedure were invited for further assessment. Twelve patients were enrolled (4 males with mean age 62 years, range 50 - 75; 8 females with mean age 45 years, range 36 – 64). Only patients with a definitive olfactory dysfunction diagnosis were finally included (determined by Sniffing Sticks Screening Test). Six patients were excluded from the study due to voluntary retirement.
Six patients and three healthy controls completed the study protocol. Healthy subjects completed only the first evaluation session, and all subjects in the patient group performed two evaluation sessions. In the first session, all participants underwent olfactory measurement test in order to diagnose olfactory dysfunction and assess severity. In addition, they were examined by an ENT specialist including a complete physical exam and nasofibroscopy. After clinical evaluation, fMRI measurements were performed using a sniffing paradigm and a trigeminal activation paradigm. Following the first session, the six subjects in the patient group were randomized in two groups of three subjects each using a 1:1 allocation ratio. Randomization sequence was generated by using computer-generated random numbers in the Statistical Package for the Social Sciences (SPSS) version 24.0, and each patient was allocated to an intervention group according to its assigned number. The first group was instructed to perform olfactory training over a 12-week period at home, and the second group received no treatment initially. On the second session, after 12 weeks, all patients were invited to repeat the olfactory testing and fMRI measurements, identical to the first session (S.1).
Olfactory dysfunction was assessed using the “Sniffing Stick Screening Test”(12). The test classifies the patients in three grades of severity depending on their performance, using a numerical 0-12 scale system: normosmia (8-12), hyposmia (5-7) and anosmia (< 5). Normative studies for the 12-odor test have been performed in several countries and is currently the only validated olfactory performance test in Chile(13–16). The Lawton et al. conversion table was applied to our results to obtain an approximate measure of the identification values of the 16 “Sniffing Stick Identification test”(17).
Olfactory training was performed using four essential oils: Lemon (Limonene 67.08% and b-pinen 12.52%), eucalyptus glovus (1-8-cineole 60.00%), clove (eugenol 75.49% and eugenol acetate 13.59%) and lavender (linalool 36.53% and linalyl acetate 32.80%). Patients were instructed to expose themselves twice a day to each odor, taking deep sniffs for 30 seconds, and resting 10 seconds between each oil. Additionally, they were instructed to evoke a memory or feeling during the odor exposure that was associated with the essential oil’s smell. Patients were contacted monthly to maintain compliance and motivation during the training period, and they confirmed that they performed smell training regularly twice per day. At the end of the 12-week training period, both groups were assessed using an olfactory performance test and a fMRI. After the second evaluation session, the group that did not undergo olfactory training received treatment.
For the scanning sessions, an odor infusion system was attached to a single air-line, with a nasal cannula, in order to perform the sniffing paradigm (S.2). This paradigm consisted of four sniffing blocks, and four normal breathing blocks. Each sniffing block consisted of 20 cycles of 40 seconds duration, and each normal breathing block consisted of 30 cycles of 60 seconds duration. The sniffing block was characterized by an odor intake, eucalyptus essential oil at 5Lt/hr., through the nasal cannula, to assure brain activation. On the other hand, the normal breathing blocks were characterized by non-odorized compressed air (5Lt/hr.) to guarantee odor cleansing of the nasal cannula during resting state fMRI. Also, a trigeminal activation paradigm was performed to assess the trigeminal system’s indemnity. This experiment consisted of five stimulation blocks and five resting blocks, each consisting of 10 cycles of 10-second duration. The stimulation block was characterized by the intake of odorless CO2 (50% at 2,5Lt/hr.) through the nasal cannula.
To compare olfactory performance scores between each group after each session, a Wilcoxon test and U-Mann-Whitney test and the alpha level for all statistical tests was set to α = 0.05.
All patients were scanned on a 1.5T MRI scanner with an 8-element head-coil. Anatomical reference images were acquired using a T1 weighted 3D Turbo Field Echo sequence. Three hundred slices with a voxel size of 0.9 x 0.9 x 0.6 mm3 with a Field of view (FOV) 230 x 253 mm. Repetition time (TR)/echo time (TE) of 7.4/3.4 ms with an inversion time delay of 876 ms. fMRI images were acquired using 2D single-shot, echo-planar imagining (EPI), gradient echo sequence. Twenty-five slices with AC-PC line orientation were acquired with a voxel size of 3.28 x 3.28 x 5.0 mm3 with a FOV of 210 x 210 mm. TR / TE of 2000/45 ms.
fMRI data was preprocessed through realignment, slice timing correction, co-registration to the T1 weighted images and normalization to the Montreal Neurological Institute (MNI) space and smoothing with (8x8x8 mm3), using SPM12, implemented in MATLAB 2018a. In addition, preprocessing steps were performed. STROBE reporting guideline was followed for this study.