Rombaux P, Weitz H, Mouraux A, Nicolas G, Bertrand B, Duprez T, Hummel T. Olfactory Function Assessed With Orthonasal and Retronasal Testing, Olfactory Bulb Volume, and Chemosensory Event–Related Potentials. Arch Otolaryngol Head Neck Surg. 2006;132(12):1346-1351. doi:10.1001/archotol.132.12.1346
To investigate whether differences in olfactory function between healthy individuals and patients with olfactory loss could be detected by various diagnostic tests.
Psychophysical testing of orthonasal and retronasal olfactory functions, magnetic resonance imaging of olfactory bulb (OB) volume, and chemosensory event–related potential (ERP) measurement performed between January 1, 2005, and October 31, 2005.
Academic tertiary referral medical center.
Eleven healthy individuals with normal olfactory function (NL) and 11 patients with nasal polyposis (NP), 11 with posttraumatic olfactory loss (PT), and 11 with postinfectious olfactory loss (PI) were included in this study.
Main Outcome Measures
Orthonasal and retronasal olfactory test results, magnetic resonance imaging–based OB volume, and ERPs to both olfactory and intranasal trigeminal stimulation.
Orthonasal and retronasal testing revealed that NL individuals had higher scores than patients with NP, PT, or PI. Retronasal scores were higher in NP patients compared with PT and PI patients. The OB volumes were higher in NL individuals compared with NP, PT, and PI patients. The OB volumes in PT patients were significantly lower than those from NP and PI patients. Olfactory ERPs were different between NL individuals and NP, PT, and PI patients, and trigeminal ERPs were significantly different when comparing NL individuals with NP patients. For the entire cohort, a significant correlation was found between orthonasal testing and OB volume, between retronasal testing and OB volume, and between both orthonasal and retronasal testing and olfactory ERP amplitudes. Olfactory ERPs were recorded in the 11 NL individuals and in 3 NP, 3 PT, and 4 PI patients, defined as responders. Orthonasal and retronasal test scores, OB volume, and olfactory ERPs were significantly larger in responders compared with nonresponders.
Significant differences in various tests that evaluate olfactory function were detectable in a cohort of NL individuals and NP, PT, and PI patients. This finding suggests that these diagnostic tools provide information in terms of the clinical assessment of olfactory function. Future studies will investigate their combined use in terms of the prognosis of olfactory function in patients with olfactory loss.
This study compares the combined evaluation of olfactory function with psychophysical testing, magnetic resonance imaging (MRI), and chemosensory event-related potential (ERP) measurement in the clinical assessment of patients with olfactory loss. Psychophysical testing of olfactory function is typically used to determine orthonasal olfactory function.1- 3 Recently, interest has increased in retronasal olfaction because it has been demonstrated that, for example, patients with sinonasal disease have better retronasal than orthonasal olfactory function.4 Retronasal olfactory function is usually evaluated with odorized powders applied in the oral cavity. Patients are asked to identify an odor from a list of descriptors using a forced-choice paradigm. Structural MRI studies5 that focus on olfactory bulb (OB) volume have been shown to be valuable in the clinical diagnosis of patients with olfactory disorders. For instance, OB volume is decreased in patients with postinfectious olfactory loss as a function of both the duration of the disease and the presence of parosmia.6,7 Finally, electrophysiologic recordings (chemosensory ERPs) in response to olfactory and/or trigeminal stimuli are of great interest in the clinical evaluation of patients with olfactory disorders8,9 because the response is less biased than results of psychophysical testing.
To address the aim of the study, results of healthy individuals were compared with those of patients with decreased olfactory function due to nasal polyposis or sinonasal disease or postinfectious and posttraumatic olfactory loss. The battery of administered tests included psychophysical measurement using orthonasal and retronasal testing, MRI-based OB volume measurement, and ERP recording in response to both olfactory and intranasal trigeminal stimulation.
This prospective study was conducted at the Department of Otorhinolaryngology of the Cliniques Universitaires Saint Luc between January 1, 2005, and October 31, 2005. Informed consent was obtained after outlining the experimental paradigm. Rules of the Ethics Committee of the Université Catholique de Louvain (Brussels, Belgium) were followed, which are in accordance with the principles of the revised Declaration of Helsinki.
Patients with nasal polyposis or sinonasal disease (NP), postinfectious olfactory loss (PI), and posttraumatic olfactory loss (PT) were asked to participate in the following tests: psychophysical tests of olfactory function (orthonasal and retronasal); MRI of the brain, nose, and paranasal sinuses; and OB volume and ERP recording in response to both olfactory and intranasal trigeminal stimulation.
Diagnosis of NP was based on nasal endoscopy. The patients with NP did not receive treatment (nasal corticoids or oral corticoids) during the month before evaluation. Diagnostic criteria for posttraumatic olfactory loss included history of olfactory disorder after head injury, patency of the olfactory cleft at endoscopic examination, and exclusion of other causes of olfactory disorders (eg, sinonasal disease). Diagnosis of postinfectious olfactory loss was based on a history of olfactory loss after an acute infection of the upper respiratory tract, which was noted at least 6 weeks after onset of the infection; patency of the olfactory cleft or absence of signs for NP or chronic sinusitis; and exclusion of other causes of olfactory loss.
A total of 44 individuals were included in the study: 11 healthy individuals (NL; age range, 22-61 years; 5 men and 6 women), 11 NP patients (age range, 35-60 years; 4 men and 7 women), 11 PT patients (age range, 20-60 years; 5 men and 6 women), and 11 PI patients (age range, 27-63 years; 5 men and 6 women).
Psychophysical testing of olfactory function was performed with the “Sniffin’ Sticks” test.3 Normative data on a large cohort of healthy individuals and patients are available.3 In this evaluation, odorants are presented in felt-tip pens. For birhinal orthonasal testing, the pen's tip is placed approximately 2 cm in front of both nostrils. This test encompasses 3 different approaches to study olfactory acuity. First, odor threshold was assessed with N-butanol using stepwise dilutions in a row of 16 felt-tip pens. Second, participants were asked to discriminate 1 odor from 3 odors presented. For each discrimination task, 3 felt-tip pens were administered, 2 containing the same odorant and 1 containing the target odorant. This was repeated for 16 triplets of odorants. Third, participants were asked to identify a row of 16 odors using multiple choice lists of 4 items for each odor. Results from tests for odor threshold (T), odor discrimination (D), and odor identification (I) are summarized as a so-called TDI score. For healthy individuals, the TDI score at the 10th percentile is 30.25 for those 18 to 35 years old.3 Functional anosmia is diagnosed if the TDI score is 16 or less.
For retronasal olfactory testing, a standardized, validated test was used.10 This test is based on the identification of odorized powders or granules presented to the oral cavity. Twenty stimuli were selected, namely, coffee, vanilla, cinnamon, cocoa, raspberry, orange, garlic, strawberry, cloves, nutmeg, onion, cheese, curry, milk, banana, mushroom, coconut, lemon, paprika, and celery. Stimulants were applied to the midline of the tongue. Participants were asked to identify the odor from a list of 4 items. After administration of each powder, participants rinsed with water. For healthy individuals, retronasal testing yielded a median score of 19 for those 18 to 35 years old.
All participants were examined using a 1.5-T MRI system (Sigma Echospeed; GE Medical Systems, Milwaukee, Wis) with a standardized protocol for olfactory tract analysis. Identical parameters for imaging were used in all cases to ensure accuracy and comparability of the measurements. The protocol included (1) standard T2-weighted fast spin-echo imaging covering the whole brain (slices of 5-mm thickness) without interslice gap to rule out any organic brain disorder; (2) T2-weighted gradient-echo images using the echoplanar imaging technique (EPI-GRE-T2, slices of 5-mm thickness; GE Medical Systems) covering the whole brain to rule out the presence of any parenchymal or meningeal disorders; and (3) 2-mm-thick T1- and T2-weighted fast spin-echo images without interslice gap in the coronal plane covering the anterior and middle segments of the base of the skull. Volumes of the OB were calculated by planimetric manual contouring using standardized methods.7 The OB contours were delineated using an electronic cursor on all images where an OB was present. Areas for all slices were added and multiplied by 2 because of the 2-mm slice thickness to obtain a volume in mm3. Measurements were performed twice by the same observer (G.N.) on different days. When the 2 measurements of a particular OB differed by less than 10%, an average of the 2 measurements was used for further statistical analyses. In case the 2 measures differed by more than 10%, a second observer (T.D.) performed a third measurement. For further analyses, the average was used from 2 of the 3 volumes that were least different from each other.
The ERPs11 were recorded in response to olfactory and trigeminal stimulation using a computer-controlled stimulator based on air-dilution olfactometry (Olfactometer OM2S; Burghart Medical Technology, Wedel, Germany). This olfactometer allows delivery of the chemical stimuli without altering mechanical or thermal conditions in the nasal cavity. Stimuli reach the nasal cavity through Teflon tubing placed into a nostril with its opening beyond the nasal valve, pointing toward the olfactory cleft. The total flow rate was 8 L/min (36°C; 80% relative humidity; stimulus duration, 200 milliseconds; stimulus rise time, <20 milliseconds). To avoid auditory evoked responses due to possible switching clicks associated with the presentation of the chemical stimuli, study participants received white noise of 60 to 70 dB sound pressure level through headphones. Study participants were asked to keep their eyes open and to breathe through their open mouth during the session. Stimulation was presented monorhinally while patients were sitting in a well-ventilated room. Phenylethyl alcohol (50% vol/vol) was used for olfactory stimulation, and carbon dioxide (50% vol/vol) was used for trigeminal stimulation. The 2 stimuli were presented 20 times each in a randomized sequence with an interstimulus interval of 30 seconds.
Electroencephalography (EEG) was performed from position Cz referenced to linked ear lobes (A1/A2). The EEG records contaminated by eye blinks and/or showing an activity higher than 50 μV were rejected before averaging. The EEG activity was averaged from 500 milliseconds in the prestimulus period until 1500 milliseconds in the poststimulus period. Chemosensory responses demonstrated a major negativity, N1 (latency, 290-490 milliseconds; amplitudes, <−2 μV), followed by a major positive complex, P2 (latency, 460-820 milliseconds; amplitudes, >+2 μV). All offline signal-processing procedures were performed using the LETSWAVE EEG toolbox (Université Catholique de Louvain).12
Statistical analysis was performed using the Medcalc pro-gram for medical statistics, version 8.1 (Medcalc, Mariakerke, Belgium).13 Variables are described as mean, standard deviation, standard error of the mean, median, and range. The normality of distribution of the studied variables was evaluated using the D’Agostino-Pearson test. Because some variables in the analysis turned out to be skewed, a natural logarithmic transformation was used to meet the normality assumption. Differences between subgroups of the study population were investigated using 1-way analyses of variance. The t test was used to investigate differences between groups. Regression analysis was performed for studying univariate associations. The receiver operating characteristic (ROC) curve was used to determine the performance of a diagnostic test in differentiating between healthy controls (negative group, n = 11) and patients with olfactory dysfunction (positive group, n = 33); pairwise comparisons of areas under the ROC curve were performed. Higher areas under the ROC curve are associated with a better performance of the test in differentiating between controls and patients with a possible maximum area under the ROC of 1. An a priori level of P<.05 was used to indicate statistical significance.
Descriptive statistics of all variables are presented in Table 1.
Characteristics of the 4 groups are summarized in Table 2. Age was not significantly different between groups (P = .68). Orthonasal testing revealed that NL individuals had higher scores than NP, PT, and PI patients (P<.001). The same was found for retronasal testing (P<.001). Interestingly, retronasal scores for NP patients were higher compared with PT and PI patients (P<.001). The OB volumes were higher in NL individuals compared with NP, PT, and PI patients (P<.001). The OB volumes from PT patients were significantly lower than those from NP and PI patients (P<.001). Olfactory ERP N1 and P2 amplitudes and N1P2 peak-to-peak amplitude were different between NL individuals and NP, PT, and PI patients (P<.001). Trigeminal ERP N1 and P2 amplitudes and N1P2 peak-to-peak amplitude were significantly different when comparing NL individuals with NP patients, respectively (P<.001), but not when comparing NL individuals with PT and PI patients.
When comparing univariate factors in the cohort study group, a significant correlation was found between orthonasal testing and OB volume (r = 0.47; P<.001), retronasal testing and OB volume (r = 0.46; P<.001), both orthonasal (r = 0.60; P<.001) and retronasal (r = 0.37; P<.001) testing and N1P2 peak-to-peak amplitude after olfactory stimulus, and orthonasal testing and N1P2 peak-to-peak amplitude after trigeminal stimulus (r = 0.21; P = .002) but not between retronasal testing and N1P2 peak-to-peak amplitude after trigeminal stimulation (r = 0.09; P = .05).
Study participants with identifiable olfactory ERPs were termed responders and those without identifiable ERPs were termed nonresponders. Significant differences between the 2 groups in terms of other measured variables are given in Table 3. Responders were found in the NL group (11 individuals), the NP group (3 patients), the PT group (3 patients), and the PI group (4 patients). Results from orthonasal testing, retronasal testing, OB volumetry, and N1P2 peak-to-peak amplitude in response to olfactory stimulation were significantly larger in responders compared with nonresponders (P<.001).
The ROC curve was used to determine the performance of a diagnostic test (orthonasal testing, retronasal testing, OB volume, N1P2 peak-to-peak amplitude of ERPs to olfactory stimuli, and N1P2 peak-to-peak amplitude of ERPs to trigeminal stimuli) between healthy individuals and patients (Table 4). The ROC curve areas were high for orthonasal testing (0.99), retronasal testing (0.98), N1P2 peak-to-peak amplitude to olfactory stimuli (0.91), and OB volume (0.90) but low for N1P2 peak-to-peak amplitude to trigeminal stimuli (0.78).
An important result of this study was the correlation between olfactory function and different diagnostic tools, such as orthonasal and retronasal testing, OB volume, and electrophysiologic responses obtained in response to olfactory stimulation. In addition, the combined use of different measures allows the development of criteria that can be applied to the diagnostics of the cause of the olfactory loss (Table 5), at least in the sense of generating hypotheses that would have to be confirmed in future studies.
Orthonasal olfaction is defined as the perception of odors that occur during sniffing and inspiration. Retronasal olfaction is defined as the perception of odors that reach the olfactory epithelium through the pharynx during eating and drinking or during expiration. In this cohort study, retronasal testing revealed that NL individuals had higher scores than PT and PI patients. In fact, this finding confirms previous work that indicates that orthonasal function is lower than retronasal olfactory function in NP patients.4,14,15 It could be hypothesized that NP patients have a gradual olfactory decline due to polyp growth patterns and, as a compensatory mechanism for orthonasal olfactory loss, develop higher capacities in retronasal olfaction to ensure olfactory acuity. On the other hand, patients with sudden onset of olfactory loss, such as PT and PI patients, would not have the time to develop higher retronasal function and, thus, typically describe loss of flavor sensations.
The OB volume varies with regard to olfactory function. This relation has been demonstrated for PT and PI patients.6,7 In PT patients, it has been hypothesized that the olfactory loss is secondary to the severing of the fila olfactoria on their passage through the cribriform plate, leading to a decreased sensory input to the OB and therefore to a decreased volume. Similarly, in PI patients it is also believed that the periphery of the olfactory system is damaged, which would lead to OB volume decrease. Interestingly, the OB volume was significantly lower in the PT group compared with the NP and PI groups. This finding may indicate that olfactory loss is typically less complete and that regeneration of olfactory function is more likely in NP and PI patients compared with PT patients.1
Results of analyses of amplitudes of olfactory ERPs exhibited an excellent correlation with results of psychophysical testing. This finding confirmed that olfactory ERPs are a useful means in the diagnosis of an olfactory deficit that is minimally biased by the patient's or investigator's expectations. Notably, trigeminal ERPs were not significantly different between groups, with the exception of ERP amplitudes, which were found to be smaller in NP patients. A possible explanation for this observation may relate to polyp formation. Specifically, polyps produce a mechanical block of the nasal cavity and thus reduce the area of respiratory epithelium with its trigeminal receptors that could be activated.
Some patients with olfactory loss, as determined through psychophysical tests, exhibited olfactory ERPs. Because olfactory ERPs signify the presence of olfactory function,8 this may be viewed as a positive prognostic factor in terms of the recovery for PT and PI patients and possibly as a positive factor of the response to corticosteroid treatment in NP patients. Future studies are needed to investigate this hypothesis.
Although activation of the trigeminal system does not seem to be appropriate in the investigation of olfactory function, it is the interaction between the trigeminal and the olfactory system that is of interest in this clinical situation.16 Patients with olfactory loss exhibit smaller responses to trigeminal stimulation. In turn, normal responses to trigeminal stimuli may reflect recovery of olfactory function.
When considering the ROC analyses, the diagnostic tools used in this study were demonstrated to discriminate healthy individuals with normal olfactory acuity from patients with olfactory disorders. However, the rate of discrimination between the 2 groups was higher than 90%, with the only exception being response amplitudes to trigeminal stimulation. Accordingly, responses to trigeminal stimulation are of only limited value in the actual diagnosis of olfactory loss. Nevertheless, valuable information may still be contained in this response with regard to patients' chances of recovery from olfactory loss.
In conclusion, the present data highlight the significant differences of various tests to evaluate olfactory function in healthy individuals and NP, PT, and PI patients. It is hypothesized that these diagnostic tools provide information in terms of the prognosis of olfactory loss and/or the response to treatment with anti-inflammatory drugs in patients with sinonasal disease.
Correspondence: Philippe Rombaux, MD, Department of Otorhinolaryngology, Université Catholique de Louvain, Cliniques Universitaires Saint Luc, Hippocrate Avenue 10, 1200 Brussels, Belgium (firstname.lastname@example.org).
Submitted for Publication: February 10, 2006; accepted May 3, 2006.
Author Contributions: Drs Weitz and Hummel had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Weitz and Hummel. Acquisition of data: Rombaux, Weitz, Mouraux, Nicolas, Bertrand, and Duprez. Analysis and interpretation of data: Rombaux, Weitz, Mouraux, Nicolas, Bertrand, Duprez, and Hummel. Drafting of the manuscript: Weitz and Hummel. Critical revision of the manuscript for important intellectual content: Rombaux, Mouraux, Nicolas, Bertrand, and Duprez. Statistical analysis: Hummel. Administrative, technical, and material support: Weitz. Study supervision: Mouraux, Nicolas, Bertrand, Duprez, and Hummel.
Financial Disclosure: None reported.