xRead - Olfactory Disorders (September 2023)

20426984, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/alr.22929, Wiley Online Library on [04/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

532

INTERNATIONAL CONSENSUS ON OLFACTION

TABLE VIII.10 (Continued)

Age,mean (SDor range), year

No. of patients (male/female)

Correlation coefficients

Study groups Healthy patients

Correlated tests

P value

Study author Mahmut et al, 2012 1240

20 (NR)

39/40

SS-ID vs SS-D

0.28

0.001

SS-ID vs SS-T (butanol) 0.34 SS-D vs SS-T (butanol) 0.28

0.001 0.001

Weierstall and

23.5 (3.7)

52/52

Healthy patients

DODT vs UPSIT R

0.19

0.05

Pause, 2012 1172

DODTvsPEA threshold UPSIT R vs SS-D SS-ID vs SS-D

0.14

0.10

0.25 0.70

0.01

Soler et al, 2016 193

52.7 (16.1)

49/61

Rhinosinusitis

0.001

SS-ID vs SS-T (butanol) 0.69 SS-D vs SS-T (butanol) 0.62 UPSIT R vs SS-T (PEA) 0.65 UPSIT R vs SS-T (PEA) 0.63 SS (PEA) vs STT (PEA) 0.67

0.001 0.001 0.001 0.001 0.001 0.001

Dotyet al, 2019 1184

58.0 (16.10)

327/409

S&T clinic patients

Kasemsuk et al, 2020 1189

42.7 (15–84) 38/112

112/38

UPSIT R vsTOIT

0.64

Aniteli et al, 2020 1241

20–80

100

Healthy and S&T clinic patients

CCCRC ID vs B-SIT right nostril CCCRC ID vs B-SIT left nostril

0.90

0.001

0.90

0.001

Tian et al, 2021 245 50.1 (27–77) 14/24

Patients with PVOD SS-ID vs SS-D

0.80 0.55 0.48

0.001 0.001

SS-ID vs SS-T SS-D vs SS-T

0.001 AD = Alzheimer disease; CCCRC = Connecticut Chemosensory Clinical Research Center; DODT = Dusseldorf Odour Discrimination Test; ETOC = European Test of Olfactory Capabilities; ID = identification; NR = not reported; NS = not significant; ODT = odor detection threshold; ODMT = Odor Discrimination/Memory Test; PEA = phenylethyl alcohol; PEMEC = d,l-beta-phenylethylmethylethylcarbinol; PVOD = postviral olfactory dysfunction; S&T = smell and taste; SD = stan dard deviation; SOIT = Scandinavian Odor-Identification Test; SS = Sniffin’ Sticks; SS-D = Sniffin’Sticks discrimination only; SS-ID = Sniffin’ Sticks identification only; SS-T = Sniffin’ Sticks threshold only; STT = Smell Threshold Test; T&T = Toyoda and Takagi; TOIT = Thai Odor-Identification Test; UPSIT R = University of Pennsylvania Smell Identification Test.

reliability, their correlation with other types of tests, and practicality. A number of tests can be self-administered, minimizing physician involvement and personnel costs. In the era of COVID-19, throw-away identification tests may have the advantage of minimizing the likelihood of instrument contamination and viral spreading from breathing on test instruments. Although very brief screening tests (eg, four items) can be used to roughly screen for smell loss, longer tests are rec ommended to minimize the likelihood of obtaining false negative and false-positive responses. Shorter screening tests can only assess the presence or absence of dysfunction and do not make it possible, in individual cases, to detect probable malingering or to accurately establish clinically useful degrees of dysfunction. This is a major limitation as decreased smell function in the absence of anosmia can

be a significant liability and patients need to be counselled regarding their perceived smell problem and the degree of their deficit. Threshold tests are generally less reliable and are more time-consuming than identification tests, but, when done properly, correlate well with them. As with identifica tion tests, forced-choice responding should be employed. There is controversy whether threshold and other types of olfactory tests add anything to identification tests. Relia bility, and thus sensitivity, is increased when test results of nominally different test measures are combined. The most appropriate statistical approach for doing this is to first convert them to z scores or other appropriate met rics and implement well-established statistical methods that take into account scale differences and test reliabil ities, as described elsewhere. 1120 Interpretation of such