Table 2:
Descriptive characteristics of studies included in systematic review and meta-analysis
| Study | Patient Number | TBI Metric | Olfactory Tests | Recruitment location | Olfactory Scale | Conclusions |
|---|---|---|---|---|---|---|
| TBI Severity | ||||||
| Bratt (2018) | 28 TBI (25 OD, 3 normosmic) |
GCS, Imaging | Sniffin’ Sticks | Level 1 neurosurgical trauma referral center | TDI 1–48 Normosmia: >30.5 Hyposmia: >16.5 - ≤30.5 Anosmia: ≤16.5 |
Patients were initially evaluated based on self-reported olfactory status. Patients reporting no olfactory dysfunction were considered “normosmic” (154/182, 84.6%). Psychophysical testing performed in 28 patients with positive self-screening. OD identified in 25/28 (25/182, 13.7%), 15 of which had anosmia. Averaged TDI score was 17.4±9.3. Anosmia present in 8.2% in chronic phase after trauma (9–104 months post injury). |
| Mod/Sev | ||||||
| Green (2003) | 35 TBI OD 196 Control |
GCS, LOC, PTA, imaging | AST | Referred to private practice for psychological or neuropsychological assessment | AST 0–10 per nostril Normosmic: >2.4 OD: ≤2.4 |
Data provided for severity based on GCS and PTA (as well as when both were considered together), and stratified by abnormal imaging. Based on GCS: Both Mod TBI (2.92±2.9) and Sev TBI (3.41±27) patients had lower AST scores than control (5.8±2.1) (p<0.001). Greater OD with increased severity of head injury. |
| Mod/Sev | ||||||
| Han (2018b) | 19 TBI hyposmic 21 TBI anosmic 19 Control |
Imaging | Sniffin’ Sticks | Smell and Taste outpatient clinic | TDI 1–48 Normosmia: >30.5 Hyposmia: >16.5 - ≤30.5 Anosmia: ≤16.5 |
Imaging study evaluating gray matter density in setting of smell. Participant groups dictated by outcome of Sniffin’ Sticks scores. Olfactory function was impaired in all smell domains in patients compared to controls. TDI: control: 32.0±3.2, Hyposmia 22.3±4.1, Anosmia 11.0±2.7 (p<0.001) Threshold: control: 7.9±2.3, hyposmia 3.7±2.3, anosmia 1.1±0.1 (p<0.001) Discrimination: control 12.6±1.6, hyposmia 9.9±2.0, anosmia 6.2±2.2 (p<0.001) Identification: control 13.5±1.2, hyposmia 8.8±2.4, anosmia 3.7±1.6 (p<0.001) |
| Mixed | ||||||
| Osborne-Crowley (2016) | 23 TBI OD 15 Control |
PTA | B-SIT | Outpatient records from three brain surgery units | NR | OD identified in 8/23 (35%) TBI patients and 2/15 (13%) controls. Of the 8 OD patients, only 3 were aware of their dysfunction. |
| Severe | ||||||
| Sigurdardottir (2016) | 132 TBI (~30–80% OD) |
GCS, GOAT | UPSIT (65 patients) B-SIT (64 patients) |
Admitted to neurosurgical departments | UPSIT 0–40 normosmia: 34–40 (males), 35 to 40 (females) mild hyposmia: 26–33 (males), 26–34 (females) severe hyposmia: 19–25 (males, females) anosmia: 6–18 (males, females). possible malingering: 0–5 B-SIT 0–12 normosmia: 9 −12 hyposmia: 7–8 anosmia: 2–6 possible malingering: 0–1 |
UPSIT: OD overall 58/67 (89.2%) - hyposmia 35/67 (53.8%), anosmia 23/67(35.4%), malingering 2/67 B-SIT: OD overall 19/65 (29.7%) - hyposmia 6/65 (9.4%), anosmia 13/65 (20.3%), malingering 1/65 Significantly different results based on testing method (p<0.001), but concluded ~30% of severe TBI patients had anosmia. |
| Severe | ||||||
| Neumann (2012) | 106 TBI (59 OD, 47 normosmic) |
GCS, LOC, PTA | B-SIT | Outpatient brain injury rehabilitation centers, local brain injury support groups | B-SIT 0–12 Normative data used for classification. |
Dysosmia was present in 59/106 (56%) of patients. People with dysosmia had higher rates of self-reported OD than normosmics (p=0.016), but only 36% of dysosmics were aware of OD. B-SIT did not significantly correlate with GCS (p=0.128), LOC (p=0.058), or PTA duration (p=0.219). |
| Mod/Sev | ||||||
| Haxel (2008) | 82 TBI (20 OD, 62 normosmic) |
GCS, Imaging | B-SIT (82 patients) Sniffin’ Sticks (19 patients) |
Admitted to hospital for head injury / trauma | B-SIT 0–12 Normosmia: ≥9 OD: <9 TDI 1–48 Normosmia: >27 Hyposmia: >16 - ≤27 Anosmia: ≤15.5 |
B-SIT: OD 14/82 (17.1%). Only 8/14 (57%) of OD patients self-reported OD. Sniffin’ Sticks: hyposmia 3/19 (15.8%), anosmia 7/19 (36.8%). Average TDI score of reported data was 23.0±9.7. GCS and LOC were not correlated with OD after head injury. |
| Mixed | ||||||
| Yousem (1996) | 25 TBI OD 8 Control |
Imaging | UPSIT 12-item memory test Single-staircase odor detection threshold |
Smell and Taste center | UPSIT: 0–40 (0–20 each nostril) Normosmic: >34 Mild hyposmic ≥27-≤34 Severe hyposmic: ≥18-≤25 Anosmic: <18 Odor memory: 0–12 per nostril PEA threshold: Anosmia: −2.0 |
UPSIT: OD 24/25 (96%); anosmia 12/25 (48%), severe OD 8/25 (32%), mild OD 4/25 (16%). Odor Memory: Control 16±2.8; 19/25 (76%) of TBI patients scored below 10 Odor Discrimination: Control 21.8±2.8; 22/25 (88%) of TBI patients scored below 20 Odor Detection Threshold: Control −6.3±1.6; 17/25 (68%) of TBI patients scored −2.0 MRI abnormalities: 22/25 (88%) OBs and tracts, 15/25 (60%) subfrontal regions, and 8/25 (32%) temporal lobes. |
| Mod/Sev | ||||||
| Meta-Analysis Inclusion Criteria Met | ||||||
| Yamaki (2020) | 31 TBI OD 10 Control |
GCS, Imaging |
OSIT-J | Admitted to hospital for severe TBI | OSIT-J 0–12 Normosmic: ≥8 OD: <8 |
OD identified in 28/31 (90.3%) of patients and 0/10 controls. Anosmia in 8/31 (25.8%), parosmia in 14/31 (45.2%), and both anosmia and parosmia in 6/31 (19.4%). OSIT-J scores were 2.9±2.8 in patients and 10.1±1.4 in controls (p<0.0001). |
| Severe | ||||||
| Green (2001) | 133 TBI OD 126 Control |
GCS, LOC, PTA, imaging | AST | Referred to private practice for psychological or neuropsychological assessment | AST 0–10 per nostril | AST score worse in “Definitive” TBI (7.3±5.4, p<0.0001) and severe TBI (7.0±5.9, p<0.0001) when compared to controls (11.4±4.1). After removing patients with low effort, there was modest correlations between smell test total scores and PTA duration (r = −0.23, p = 0.001), GCS (r = 0.30, p < 0.001), and abnormality on CT (r = −0.40, p < 0.001, n = 143). Patients with more severe TBI injuries were 10–12 times more likely to have OD than mild TBI. |
| Mod/Sev | ||||||
| Han (2018a) | 22 TBI hyposmic 24 TBI anosmic 22 Control |
Imaging | Sniffin’ Sticks | History of head injury | TDI 1–48 Normosmia: >30.5 Hyposmia: >16.5 - ≤30.5 Anosmia: ≤16.5 |
Imaging study evaluating gray matter density in setting of smell. Participant groups dictated by outcome of Sniffin’ Sticks scores. Olfactory function was impaired in all smell domains in patients compared to controls. TDI: control: 33.8±3.1, Hyposmia 22.0±4.0, Anosmia 11.3±2.7 (p<0.001) Odor Detection Threshold: control: 7.8±2.2, hyposmia 3.5±2.3, anosmia 1.1±0.3 (p<0.001) Odor Discrimination: control 12.4±1.7, hyposmia 10.1±1.9, anosmia 6.3±2.3 (p<0.001) Odor Identification: control 13.6±1.3, hyposmia 8.4±2.5, anosmia 3.9±1.6 (p<0.001) Compared to controls, hyposmic and anosmic patients had lower left (p-0.002, p<0.001)), right (p=0.01, p<0.001), and whole OB (p=0.002, p<0.001) volumes. |
| Mixed | ||||||
| Miao (2015) | 21 TBI OD 26 Control |
LOC, Imaging | Sniffin’ Sticks T&T | Department of Otolaryngology, patients complaining of OD | Normative data used for classification. | Study evaluating MRI and oERP in traumatic anosmic patients. TDI scores worse in cases (5.38±2.826) versus controls (32.05±2.89) (p=0.001). T&T scores worse in cases (5.92±0.13) versus controls (−0.99±0.97) (p=0.001). MRI: In patients with measurable olfactory bulbs (OB), OB volume was lower than controls on right (p=0.005) and left (p=0.012) sides. oERP: oERPs were detectable in 17 patients, but had longer latencies and lower amplitudes than in controls (p<0.05). Nine anosmic patients had no detectable oERPs. |
| Mixed (primarily Mod/Sev based off abnormal imaging) | ||||||
| Xydakis (2015) | 40 TBI OD 47 Control (normal neuroimaging) 8 Control (no neuroimaging) |
Imaging, Injury Severity Score | UPSIT | US service members with blast-related injuries requiring transfer to the US | UPSIT: 0–40 Normosmic: ≥33 Hyposmic: ≥25–33 Anosmic: <25 |
OD reported in 14/40 (35.0%) of patients with moderate/severe TBI. Olfactory testing predicted abnormal neuroimaging better than chance alone (AUC 0.78, p<0.001). |
| Mod/Sev | ||||||
| Parma (2012) | 12 TBI OD 12 Control |
GCS | UPSIT | History of head injury | NR | Significantly worse UPSIT scores in cases (21.17±8.65) versus controls (32.27±3.77) (p<0.001). |
| Severe | ||||||
| Fujiwara (2008) | 46 TBI OD 25 Control |
GCS | Smell ID Test | 1 year post injury from consecutive admission lists | NR | Both Mod TBI (31.93±6.78, p<0.01) and Sev TBI (28.16±7.78, p<0.0001) demonstrated worse olfactory scores than controls (35.92±2.78). |
| Mod/Sev | ||||||
| Sandford (2006) | 7 TBI OD 36 Control |
GCS, LOC, Imaging | SDCOIT | Presenting to peds ED with blunt head trauma | Scored based on percentage of correctly identified odorants – 0–100% Hyposmia: ≤75% Anosmia: 0% |
Pediatric population, scores from mod/sev subset of TBI were used to generate study effect size Hyposmia was identified in 2/7 (28.6%) of patients. No patients were anosmic. Olfactory function was predicted by both GCS (p<0.05) and head CT abnormality (p<0.01). |
| Mod/Sev | ||||||
| Savage (2002) | 13 TBI OD 13 Control |
LOC, PTA, Imaging | CCCRC – threshold and olfactory-word ID Delayed Odor Recognition Memory Test |
2 medical rehab facilities, outpatient neuropsychology clinic | Threshold 0–10, higher score indicating more sensitive olfactory threshold ID: 0–8, higher score indicating better performance Recognition: 0–7, higher score indicating better performance |
Odor Identification: TBI patients had impaired odor-word ID scores on left nostril (p=0.001) and right nostril (p<0.00001) as compared to controls. Odor Detection Threshold: No difference between controls and TBI group. Odor Recognition: TBI patients had impaired recognition scores on left nostril (p=0.0007) and right nostril (p=0.0001) as compared to controls. |
| Mod/Sev | ||||||
| Yousem (1999) | 36 TBI OD 24 Control |
Imaging | UPSIT 12-item memory test Single-staircase odor detection threshold |
Smell and Taste center | UPSIT: 0–40 (0–20 each nostril) Odor memory: 0–12 per nostril PEA threshold: −1.9 - −10.0 |
UPSIT scores were significantly worse in patients (21.9±10.5) compared to controls (26.6±2.6) (p<0.0001). Anosmia was identified in 16/36 (44.4%) of patients. Patients had worse function on all olfactory domains after controlling for age (p<0.001). Left OB and tract volumes were correlated with left and total UPSIT scores. There was a significant difference in right and left OB and tract volumes between TBI and control patients, as well as between anosmic and hyposmic patients. |
| Mod/Sev | ||||||
| Levin (1985) | 45 TBI OD 19 Control |
GCS, LOC, PTA | Olfactory ID Test | History of head injury | ID test: 0–12 | Data provided for severity based on GCS, LOC, and PTA. GCS: In comparison with age-matched controls, olfactory naming and recognition were impaired in TBI patients. No difference noted in patients with mild TBI. Recognition: Mod TBI (median 8.3), Sev TBI (median 8.1), controls (median 9.9) (p<0.01) Naming: Mod TBI (median 2.0), Sev TBI (median 07), controls (median 4.1) (p<0.01) |
| Mod/Sev | ||||||
Note: AST: Alberta Smell Test; B-SIT: Brief Smell Identification Test; CCCRC: Connecticut Chemosensory Clinical Research Center; GCS: Glasgow Coma Scale; GOAT: Galveston Orientation and Amnesia Test; ID: Identification; LOC: loss of consciousness; Mod: Moderate; MRI: magnetic resonance imaging; NR: not reported; OD: olfactory dysfunction; oERP: olfactory event-related potentials; OSIT-J: odor stick identification test for the Japanese; PTA: post-traumatic amnesia; SDCOIT: San Diego Children’s Olfaction Identification Test; Sev: Severe; TBI: traumatic brain injury; TDI: Threshold, Discrimination, Identification; UPSIT: University of Pennsylvania Smell Identification Test