Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Olfactory identification deficits at identification as ultra-high risk for psychosis are associated with poor functional outcome

Schizophrenia Research, 2-3, 161, pages 156 - 162

 “This month we focus on the need for more precise diagnostic tools in the early stage of illness – i.e. those at risk of transition to psychosis. With an emphasis on early recognition and effective intervention there is an urgent need to develop such tools in order to prevent accruing morbidity and disease progression. A lot of work is being done to identify clinical, cognitive and biological markers.”

Robin Emsley, Editor of the Schizophrenia Research Centre

Abstract

Background

We have previously reported that olfactory identification (OI) deficits are a promising premorbid marker of transition from ultra-high risk (UHR) to schizophrenia, but not to psychotic illness more generally. Whether this remains the case at longer follow-up, and whether there is decline in OI ability are unclear.

Method

The University of Pennsylvania Smell Identification Test (UPSIT) was administered to 81 participants at baseline (identification of risk for psychosis) and 254 individuals at follow-up. Forty-nine participants underwent UPSIT assessment at both time points. UPSIT scores were investigated at an average of 7.08 years after identification of risk in relation to transition to psychosis, a diagnosis of schizophrenia, and psychosocial/functional outcome.

Results

UPSIT scores at baseline and follow-up did not differ between participants who transitioned to psychosis and those who did not. Similarly, there were no significant differences on UPSIT scores at baseline or follow-up between individuals with a diagnosis of schizophrenia and transitioned individuals without schizophrenia. Those with a poor functional outcome showed significantly lower baseline UPSIT scores than participants with good outcome. There was no significant association between functional outcome and follow-up UPSIT scores. There were no significant changes in UPSIT over time for any group.

Conclusions

These results suggest that impaired OI is not a good marker of the onset of psychosis and schizophrenia, but may differentiate UHR individuals who experience a poor functional outcome, regardless of transition status.

Keywords: Olfactory identification, Olfaction, Smell, At-risk, Ultra-high risk, Psychosis, Schizophrenia, Orbitofrontal cortex, Longitudinal.

1. Introduction

Deficits in olfactory identification (OI) are a reliable finding in chronic schizophrenia and first-episode psychosis (Brewer et al, 1996a, Kopala et al, 1993, and Rupp, 2010), with a recent meta-analysis demonstrating a medium to large effect size for these deficits ( Moberg et al., 2013 ). They are also present in first-degree relatives (Kamath et al, 2014, Keshavan et al, 2009, Kopala et al, 1998, Moberg et al, 2013, and Roalf et al, 2006) and people with schizotypal features, although to a lesser extent ( Moberg et al., 2013 ), suggesting that they are a possible endophenotype for the disorder. This would imply that OI deficits are detectable in at-risk populations before the onset of frank psychotic illness. Indeed there is evidence that young people clinically at ultra-high risk (UHR) for psychosis also show impaired OI (Brewer et al, 2003, Kamath et al, 2014, Kamath et al, 2012, and Woodberry et al, 2010), with a pooled medium to large effect size (Moberg et al, 2013 and Turetsky et al, 2012). Moreover, OI deficits may be a marker of transition from the UHR state to schizophrenia specifically (rather than “psychosis” more generally) ( Brewer et al., 2003 ). It is worth noting, however, that Gill et al. (2014) failed to find a significant reduction in olfactory identification in the UHR sample, nor in relation to the onset of psychosis or schizophrenia.

One possible explanation for these findings is that OI deficits reflect neurodevelopmental compromise during adolescence ( Brewer et al., 2006 ), since the development of OI ability closely parallels orbitofrontal cortex maturation through to adulthood ( Doty et al., 1984 ). This suggests that OI deficits result from arrested prefrontal neural development, given that the lower-order pathways that mediate olfactory detection and sensitivity, despite some abnormalities (e.g.Kayser et al, 2013 and Turetsky et al, 2008), allow sensory information to reach orbitofrontal regions ( Brewer et al., 2006 ). This hypothesis is supported by the findings that OI deficits are associated with negative symptoms (Brewer et al, 1996a, Brewer et al, 2001, Corcoran et al, 2005, Good et al, 2006, and Ishizuka et al, 2010) and more specifically, the deficit syndrome of schizophrenia ( Strauss et al., 2010 ).

This model would suggest that baseline OI deficits in at-risk samples would be likely to predict poor functional outcome at follow-up, and that progressive OI impairments (or perhaps failure to show normal developmental gains) would be associated with a later diagnosis of schizophrenia. While it has been reported that baseline OI deficits predict functional outcome four years later ( Good et al., 2010 ) in people with a first episode of psychosis, it is unclear whether that is the case in at-risk samples. One reason this is not yet known is because follow-up times in at-risk research have generally been short and the degree of functional impairment has not been reported. The best available evidence (Barbato et al, 2012 and Woodberry et al, 2013) suggests that OI deficits are stable over time in individuals meeting UHR criteria. However, because few of the at-risk participants in these studies developed psychosis, it is unclear whether the development of psychosis or schizophrenia specifically has any impact, or whether those with poor outcome show more decline.

In the current study, we investigated OI in a group of individuals identified as UHR at the Personal Assessment and Crisis Evaluation (PACE) Clinic in Australia between two and 15 years previously, some of whom had also completed the OI task at baseline. Based on our earlier findings ( Brewer et al., 2003 ), we hypothesised that, at follow-up, OI deficits would be more apparent in those individuals with a diagnosis of schizophrenia relative to those who transitioned to psychosis more generally. Furthermore, we expected that baseline OI deficits would be associated with a diagnosis of schizophrenia and poorer functional outcome at follow-up. We expected that OI deficits would not show progressive decline regardless of outcome.

2. Method

2.1. Participants and procedure

Participants were 286 individuals (45.8% male) identified as UHR for psychosis between 2.39 and 14.87 years previously (mean = 7.08; SD = 3.58; median = 6.43). They were part of a larger study (N = 416) aimed at following up all participants consecutively admitted to the PACE Clinic for baseline assessment between 1993 and 2006. Details of the outcome of this cohort are described in Nelson et al. (2013) . At follow-up assessment, 254 participants were assessed on the University of Pennsylvania Smell Identification Test (UPSIT; see Table 1 ). At baseline, 81 participants had been assessed on the UPSIT ( Table 2 ). Findings for this group after clinical follow-up at 18 months post-baseline have been reported previously ( Brewer et al., 2003 ), where we showed that OI deficits were evident in those that developed schizophrenia. An additional nine participants had transitioned to psychosis over the longer follow-up period. Forty nine participants had UPSIT assessment at baseline and follow-up ( Table 3 ).

Table 1 Characteristics of the subsample with UPSIT at follow-up.

Follow-up UHR-NP (n = 186) UHR-P (n = 68) UHR-P-Other (n = 45) UHR-P-SCZ (n = 23) Good outcome (n = 191) Poor outcome (n = 63) UHR-NP vs UHR-P UHR-P-Other vs UHR-P-SCZ Good vs poor outcome
Estimate df p-Value Estimate df p-Value Estimate df p-Value
UPSIT, M (SD) 32.32 (3.96) 31.57 (4.57) 32.24 (4.15) 30.26 (5.15) 32.41 (3.99) 31.24 (4.5) F = 1.13 1,251 0.3 F = 2.49 1,65 0.1 F = 3.56 1,251 0.06
Age (years), M (SD) 25.51 (4.88) 27.59 (5.34) 27.04 (5.10) 28.65 (5.75) 25.88 (5.09) 26.60 (5.05) t = − 2.94 252 0.004 t = − 1.18 66 0.2 t = − 0.97 252 0.3
Length of follow-up, M (SD) 6.84 (3.18) 8.85 (3.07) 8.82 (3.01) 8.89 (3.24) 7.18 (3.13) 7.95 (3.62) t = − 4.50 252 < 0.001 t = − 0.09 66 > 0.9 t = − 1.50 94.4 0.1
Female gender, N (%) 106 (57.0) 40 (58.8) 29 (47.8) 11 (64.4) 116 (60.7) 30 (47.6) χ2 = 0.01 1 > 0.9 χ2 = 1.12 1 0.3 χ2 = 2.82 1 0.07
Cigarette smoker, N (%) 82 (44.1) 34 (53.1) 26 (57.8) 8 (34.8) 87 (45.5) 29 (46.0) χ2 = 0.51 1 0.5 χ2 = 2.83 1 0.1 χ2 = 0.00 1 > 0.9
Ever a regular THC user, N (%) 56 (30.1) 22 (32.4) 15 (33.3) 7 (30.4) 56 (29.3) 22 (34.9) χ2 = 0.25 1 0.6 χ2 = 0.02 1 0.9 χ2 = 0.57 1 0.5
Neuroleptic medication in past 2 years, N (%) 5 (2.7) 33 (48.5) 17 (37.7) 16 (69.6) 17 (8.90) 21 (33.3) χ2 = 79.99 1 < 0.001 χ2 = 3.60 1 0.03 χ2 = 21.85 1 < 0.001
Current Full-Scale IQ, M (SD) 102.51 (14.31) 98.17 (15.31) 97.69 (17.01) 99.19 (11.15) 103.12 (14.21) 96.05 (14.90) t = 2.08 249 0.04 t = − 0.37 64 0.7 t = 3.36 249 0.001
BPRS psychotic subscale, M (SD) 5.70 (2.31) 8.81 (4.93) 7.09 (3.34) 12.17 (5.83) 5.60 (2.56) 9.37 (4.37) t = − 4.99 78.1 < 0.001 t = − 3.87 29.6 0.001 t = − 6.48 76.6 < 0.001
SANS total score, M (SD) 9.39 (11.58) 16.50 (16.60) 11.18 (11.66) 26.91 (19.89) 6.04 (6.77) 27.11 (16.01) t = − 3.25 92.2 0.002 t = − 3.50 30.0 0.001 t = − 10.15 69.5 < 0.001

Abbreviations: UHR-P, transitioned to psychosis; UHR-NP, not transitioned to psychosis; UHR-P-SCZ, transitioned and has diagnosis of schizophrenia; UHR-P-Other, transitioned, but with no schizophrenia diagnosis; THC, tetrahydrocannabinol; BPRS, Brief Psychiatric Rating Scale; SANS, Scale of Assessment for Negative Symptoms.

Table 2 Characteristics of the subsample with UPSIT at baseline.

Baseline UHR-NP (n = 50) UHR-P (n = 31) UHR-P-Other (n = 20) UHR-P-SCZ (n = 11) Good outcome (n = 37) Poor outcome (n = 16) UHR-NP vs UHR-P UHR-P-Other vs UHR-P-SCZ Good vs poor outcome
Estimate df p-Value Estimate df p-Value Estimate df p-Value
UPSIT score, M (SD) 32.14 (3.85) 31.32 (4.30) 31.90 (4.08) 30.27 (4.69) 32.92 (3.09) 29.81 (5.10) F = 1.06 1,78 0.3 F = 1.03 1,28 0.3 F = 6.36 1,50 0.02
Age (years), M (SD) 19.66 (3.40) 19.84 (3.62) 19.85 (3.57) 19.82 (3.89) 20.57 (3.52) 19.50 (2.97) t = − 0.22 79 0.8 t = 0.02 29 > 0.9 t = 1.06 51 0.3
Length of follow-up, M (SD) 9.19 (3.99) 7.82 (5.09) 8.40 (4.73) 6.77 (5.77) 11.22 (0.76) 11.32 (0.77) t = 1.28 52.6 0.2 t = 0.85 29 0.4 t = − 0.45 51 0.7
Female gender, N (%) 23 (46.0) 17 (54.8) 11 (55.0) 6 (54.5) 27 (73.0) 6 (37.5) χ2 = 0.30 1 0.4 χ2 = 0.00 1 > 0.9 χ2 = 4.57 1 0.03
Current Full-Scale IQ, M (SD) 102.14 (15.82) 93.77 (14.21) 93.45 (13.78) 94.36 (15.63) 101.03 (14.03) 89.00 (15.85) t = 2.40 78 0.02 t = − 0.17 29 0.9 t = 2.76 51 0.008
BPRS psychotic subscale, M (SD) 8.43 (2.75) 9.42 (2.47) 9.45 (2.72) 9.36 (2.06) 8.42 (2.71) 8.69 (2.65) t = − 1.63 78 0.1 t = 0.09 29 > 0.9 t = − 0.33 50 0.7
SANS total score, M (SD) 14.80 (11.10) 22.87 (15.90) 19.15 (14.48) 29.64 (16.80) 13.30 (11.16) 24.38 (15.71) t = − 2.48 48.2 0.02 t = − 1.82 29 0.08 t = − 2.56 21.8 0.02

Note: For participants who were not assessed at follow-up, length of follow-up was calculated based on the last time they were seen at the PACE clinic.

Abbreviations: UHR-P, transitioned to psychosis; UHR-NP, not transitioned to psychosis; UHR-P-SCZ, transitioned and has diagnosis of schizophrenia; UHR-P-Other, transitioned, but with no schizophrenia diagnosis; BPRS, Brief Psychiatric Rating Scale; SANS, Scale of Assessment for Negative Symptoms.

Table 3 Characteristics of the subsample with UPSIT at both baseline and follow-up.

  UHR-NP (n = 31) UHR-P (n = 18) UHR-P-Other (n = 14) UHR-P-SCZ (n = 4) Good outcome (n = 34) Poor outcome (n = 15) UHR-NP vs UHR-P UHR-P-Other vs UHR-P-SCZ Good vs poor outcome
Estimate df p-Value Estimate df p-Value Estimate df p-Value
UPSIT baseline, M (SD) 32.45 (3.50) 31.11 (5.04) 31.71 (4.79) 29.00 (6.06) 32.94 (3.13) 29.73 (5.27)
UPSIT follow–up, M (SD) 32.42 (3.58) 29.50 (7.03) 30.64 (5.42) 25.50 (11.21) 32.35 (3.92) 29.07 (7.09)
Age (years) baseline, M (SD) 20.42 (3.12) 20.33 (3.91) 20.29 (4.03) 20.50 (4.04) 20.79 (3.49) 19.47 (3.07) t = 0.08 47 > 0.9 t = − 0.09 16 > 0.9 t = 1.27 47 0.2
Age (years) follow-up, M (SD) 31.45 (3.06) 31.83 (4.12) 31.57 (4.18) 32.75 (4.35) 32.00 (3.60) 30.67 (2.99) t = − 0.37 47 0.7 t = − 0.49 16 0.6 t = 1.25 47 0.2
Length of follow-up M (SD) 11.17 (0.68) 11.37 (0.79) 11.26 (0.78) 11.75 (0.84) 11.25 (0.75) 11.23 (0.69) t = − 0.95 47 0.3 t = − 1.11 16 0.3 t = 0.08 47 > 0.9
Female gender, N (%) 18 (58.1) 13 (72.2) 10 (71.4) 3 (75.0) 25 (72.5) 6 (40.0) χ2 = 0.47 1 0.5 χ2 = 0.00 1 > 0.9 χ2 = 3.70 1 0.06
Cigarette smoker follow-up, N (%) 11 (35.5) 13 (72.2) 12 (85.7) 1 (25.0) 17 (50.0) 7 (46.7) χ2 = 3.94 1 0.05 χ2 = 3.09 1 0.08 χ2 = 0.01 1 > 0.9
Ever a regular THC user at follow-up, N (%) 8 (25.8) 5 (27.8) 4 (28.6) 1 (25.0) 10 (29.4) 3 (21.4) χ2 = 0.00 1 > 0.9 χ2 = 0.00 1 > 0.9 χ2 = 0.05 1 0.8
Neuroleptic medication within 2 years of follow-up, N (%) 1 (3.2) 8 (44.4) 4 (28.6) 4 (100.0) 4 (11.8) 5 (33.3) χ2 = 9.51 1 < 0.001 χ2 = 3.00 1 0.08 χ2 = 3.15 1 0.08
Full-Scale IQ baseline, M (SD) 99.81 (14.57) 92.72 (15.50) 91.43 (15.69) 97.25 (16.09) 100.32 (14.02) 90.13 (15.72) t = 1.60 47 0.1 t = − 0.65 16 0.5 t = 2.26 47 0.03
Full-Scale IQ follow-up, M (SD) 99.35 (14.57) 93.06 (14.19) 92.14 (14.69) 96.25 (13.77) 100.38 (13.10) 89.47 (15.46) t = 1.47 47 0.1 t = − 0.50 16 0.6 t = 2.54 47 0.01
BPRS psychotic subscale baseline, M (SD) 8.23 (2.79) 8.72 (2.56) 8.86 (2.85) 8.25 (1.26) 8.38 (2.77) 8.47 (2.59) t = − 0.62 47 0.5 t = 0.61 12.2 0.6 t = − 0.10 47 > 0.9
BPRS psychotic follow-up subscale, M (SD) 6.45 (2.87) 7.61 (3.82) 6.64 (3.05) 11.00 (4.76) 5.82 (2.55) 9.27 (3.51) t = − 1.20 47 0.2 t = − 2.23 16 0.04 t = − 3.87 47 < 0.001
SANS total score baseline, M (SD) 14.03 (11.14) 20.39 (16.34) 17.64 (15.14) 30.00 (18.99) 13.15 (10.99) 23.67 (15.99) t = − 1.46 26.3 0.2 t = − 1.37 16 0.2 t = − 2.32 20.1 0.03
SANS total score follow-up, M (SD) 12.68 (16.93) 12.44 (17.45) 8.07 (8.37) 27.75 (31.95) 5.62 (5.66) 28.40 (22.86) t = 0.05 47 > 0.9 t = − 1.22 3.1 0.3 t = − 3.81 14.8 0.002

Note: No analyses are presented for the UPSIT because they are contained inTable 1 and Table 2.

Abbreviations: UHR-P, transitioned to psychosis; UHR-NP, not transitioned to psychosis; UHR-P-SCZ, transitioned and has diagnosis of schizophrenia; UHR-P-Other, transitioned, but with no schizophrenia diagnosis; THC, tetrahydrocannabinol; BPRS, Brief Psychiatric Rating Scale; SANS, Scale of Assessment for Negative Symptoms.

Detailed criteria for the identification of the UHR group are described by Yung et al. (2003) and are summarized as follows: 1) attenuated positive symptoms, 2) brief limited intermittent psychotic symptoms, and/or 3) trait vulnerability for psychotic illness (schizotypal personality disorder or a history of psychosis in a first-degree relative) and a deterioration in functioning or chronic low functioning. In addition to these inclusion criteria, participants were aged 15 to 30 years and had not experienced a previous psychotic episode (treated or untreated). Exclusion from the current analyses was based on the following criteria: documented organic brain impairment; history of head injury with loss of consciousness; current viral or other severe medical condition, upper respiratory tract disease, cold, sinus problem or hay fever; a history of nasal trauma; documented poor eyesight or hearing; and inadequate command of English. All participants were neuroleptic naive at baseline assessment.

To locate and recontact participants in this cohort, an extensive tracking system was employed [see ( Nelson et al., 2013 )]. Transition and diagnostic information for participants with baseline UPSIT data who were deceased, not located or refused follow-up (N = 22) was obtained from the State of Victoria's public mental health service records, which documents all public mental health admissions in Victoria. Compared with the participants from whom clinical data was obtained at follow-up interview, those with transition and diagnostic information obtained from state records were more likely to be male (85%). All participants provided written informed consent in accordance with guidelines provided by the local mental health service research and ethics committee.

2.2. Measures

Olfactory identification was assessed using the UPSIT ( Doty et al., 1984 ), a standardized multiple-choice scratch-and-sniff test consisting of four booklets, each containing 10 items. UHR criteria at baseline and transition to psychosis were assessed using the Comprehensive Assessment of At-Risk Mental States (CAARMS; Yung et al., 2005 ). The Brief Psychiatric Rating Scale (BPRS, psychotic subscale; Overall and Gorham, 1962 ), and Scale for the Assessment of Negative Symptoms (SANS; Andreasen, 1982 ) were used to measure positive and negative psychotic symptoms respectively. Psychosocial functioning and quality of life at follow-up were assessed using the Social and Occupational Functioning Assessment Scale (SOFAS; Goldman et al., 1992 ) and Quality of Life Scale (QLS; Heinrichs et al., 1984 ).

Diagnosis of schizophrenia was made with the Structured Clinical Interview for DSM-IV (SCID; First et al., 1997 ). Here we report current/lifetime diagnosis as reported at follow-up assessment because of known diagnostic variability early in the illness course ( Schwartz et al., 2000 ). Participants were asked to report on experiencessincethey were last seen at the PACE Clinic, so those in full remission who had not experienced any psychotic symptoms since PACE would not rate for a specific diagnosis (even though they had transitioned to frank psychosis). Here we compared individuals with a diagnosis of schizophrenia at follow-up to transitioned individuals without schizophrenia.

Substance use data was inconsistently recorded at baseline and is not reported here. At follow-up, “ever a regular THC (tetrahydrocannabinol) user” was defined as any period over the participant's lifetime of weekly or greater THC use, reported on semi-structured interview. Current cigarette smoking was also recorded. Neuroleptic medication was recorded if the participants reported taking neuroleptic medication for some or all of the time during the two years prior to follow-up assessment.

2.3. Data analysis

Latent group analyses were conducted to determine poor and good functional outcome based on SOFAS and QLS scores at follow-up assessment using Mplus ( Muthén and Muthén, 1998–2007 ). We tested two, three and four group models using entropy and BIC as indicators of model fit. For good model fit, entropy should be greater than 0.8. Higher BIC indicates better model fit. An increase in BIC greater than 10 is strong evidence for one model over another.

Analyses were conducted in three sections: 1) the association between follow-up UPSIT scores and outcome (n = 254); 2) the association between baseline UPSIT scores and outcome (n = 81); and 3) the change in UPSIT scores over the follow-up period in relation to outcome (n = 49). Within each section, three group comparisons (outcomes) were made: a) participants who transitioned to psychosis (UHR-P) were compared to those who did not transition to psychosis (UHR-NP); b) participants who had a diagnosis of schizophrenia at follow-up (UHR-P-SCZ) were compared to those who transitioned but did not have a diagnosis of schizophrenia at follow-up (UHR-P-Other); and c) participants with poor and good functional outcome were compared, regardless of transition to psychosis.

Comparisons of clinical variables were analysed using chi-square (continuity corrected) for binary variables and independent groupst-tests for continuous ones. The associations between UPSIT scores and BPRS and SANS scores were assessed with Spearman's rho. Effect sizes were calculated using Cohen'sd.

The associations between UPSIT scores and group were analysed with ANCOVA. Raw scores were used and covaried for age. Change in UPSIT score over time was assessed using repeated measures ANCOVA with a two-factor design that included a between-subject (group) factor and a within-subject factor (UPSIT total score), covaried for age at baseline and the length of the follow-up period. Analyses were re-run covarying for gender. This did not change results and are not reported here. Because there were significant associations between the covariates (age and length of follow-up) and UPSIT scores, inclusion of these covariates could be attenuating real underlying group differences. To deal with this, analyses were repeated without covariates. These have not been reported here as there were no significant changes to associations.

3. Results

3.1. Latent group analyses of functional outcome

Latent groups analyses were performed on the entire cohort with available functional outcome data (QLS and SOFAS) at follow-up (n = 271). Analyses to determine functional outcome status showed that a two-group model had the best fit (entropy = 0.950; BIC = 4521.41). The two-group model was better than a three- (entropy = 0.851; BIC = 4428.59) or four- (entropy = 0.881; BIC = 4369.91) group model of functional outcome in this cohort. The groups were labelled “good outcome” and “poor outcome”.

All participants with follow-up UPSIT data also had functional outcome data (n = 254; see Table 1 ) and were classified according to the above model: 24.8% had poor functional outcome [mean QLS = 60.55 (SD = 17.68); mean SOFAS = 46.54 (SD = 8.65)] and 75.2% showed good outcome [mean QLS = 108.52 (SD = 11.64), mean SOFAS = 74.73 (SD = 11.25)]. 49.2% of the poor outcome group had transitioned to psychosis, compared to 19.4% of the good outcome group (χ2 = 21.51;p < 0.001).

Fifty-three participants with baseline UPSIT data also had functional outcome data (see Table 2 ). Of those, 30.2% showed poor outcome and 69.8% good outcome. 50.0% of participants in the poor outcome group had transitioned to psychosis compared to 29.9% in the good outcome group (χ2 = 2.00;p = 0.2). All participants with UPSIT data at baseline and follow-up (n = 49) also had functional outcome data. 30.6% had poor outcome. 46.7% of the poor outcome group had transitioned to psychosis, compared to 32.4% of the good outcome group (χ2 = 0.92;p = 0.3). The mean follow-up SOFAS and QLS scores in the baseline subsample, and baseline and follow-up subsample, were similar to those with UPSIT at follow-up, suggesting that outcome groups are comparable across the subsamples described.

3.2. Associations between UPSIT scores, clinical and demographic variables

Correlations between follow-up UPSIT and follow-up symptom scores were small. Spearman's correlation coefficient between UPSIT and BPRS-psychotic subscale was − .091,p = 0.2, and between UPSIT and SANS was − .084,p = 0.2. UPSIT scores at follow-up did not significantly differ between THC users and non-users [t209 = 0.94,p = 0.3], cigarette smokers and non-smokers [t237 = 0.28,p = 0.8], and participants who had used neuroleptics in the past two years compared to those who had not [t238 = − 1.17,p = 0.2].

There were very weak associations between baseline UPSIT and baseline symptom scores. Spearman's correlation coefficient between UPSIT and BPRS-psychotic subscale was − .04,p = 0.6, and between UPSIT and SANS was − .07,p = 0.4.

3.3. UPSIT at follow-up assessment

At follow-up, 254 participants were administered the UPSIT. Data from the follow-up sample is presented in Table 1 . UPSIT scores were compared between UHR-P (n = 68) and UHR-NP (n = 186) groups, and did not significantly differ (F1,251 = 1.13,p = 0.3,d = 0.18). Similarly, there were no significant differences when UHR individuals with a diagnosis of schizophrenia (n = 23) and transitioned individuals without schizophrenia (n = 45) were compared (F1,65 = 2.49,p = 0.1,d = 0.44). Comparison of UPSIT scores when the groups were defined by functional outcome (poor outcomen = 63; good outcomen = 191) also showed no significant differences (F1,251 = 3.56,p = 0.06,d = 0.28). Inspection of the effect sizes showed that the largest effect was between UHR-P-SCZ and UHR-P-Other.

3.4. UPSIT at baseline (at identification as UHR)

At baseline, 81 participants were administered the UPSIT. Data from this baseline sample are presented in Table 2 . There was no significant difference in baseline UPSIT scores between UHR-NP (n = 50) and UHR-P (n = 31;F1,78 = 1.06,p = 0.3,d = 0.20). Similarly, baseline UPSIT scores for individuals with a diagnosis of schizophrenia (n = 11) did not differ significantly from those transitioned individuals without schizophrenia (n = 20;F1,28 = 1.03,p = 0.3,d = 0.43). Finally, the poor outcome group (n = 16) showed lower scores than the good outcome group (n = 37). This difference was statistically significant and the effect size was large (F1,50 = 6.36,p = 0.015,d = 0.82).

Because of the strong association between transition to psychosis and poor outcome, we investigated whether the significant association between UPSIT at baseline and poor outcome was driven solely by the UHR-P cases with poor outcome. To test this, analysis of baseline UPSIT scores by outcome was repeated separately for UHR-P and UHR-NP. There was a statistically significant association between UPSIT at baseline and poor outcome in the UHR-NP group (F1,30 = 7.24,p = 0.01). The association was not significant for the UHR-P group (F1,15 = 0.44,p = 0.5).

3.5. Exploratory analyses of change in UPSIT scores over time

Forty-nine participants were assessed on the UPSIT at both baseline and follow-up assessments. Data for these participants is presented in Table 3 . Comparing UHR-P and UHR-NP groups, there was no significant main effect of time on UPSIT performance (F1,45 = 0.01,p > 0.9), and no significant Group by Time interaction (F1,45 = 1.46,p = 0.2). Comparison of UHR-P-SCZ and UHR-P-Other showed no significant main effect of time (F1,14 = 0.01,p > 0.9), and no significant Group by Time interaction (F1,14 = 0.66,p = 0.4). The statistical comparison should be interpreted with caution given the small number of participants with schizophrenia. The mean drop in UPSIT score in the group with schizophrenia was 3.5 points, but this was primarily accounted for by a drop of 13 points for one participant. Finally, analyses were repeated for the outcome groups. There was no significant main effect for time on UPSIT performance (F1,45 = 0.07,p = 0.8), and no significant Group by Time interaction (F1,45 = 0.02,p > 0.9).

4. Discussion

In this study we assessed OI in a large UHR cohort followed up over the medium to long-term. At follow-up, we were not able to show a significant difference in OI ability between the UHR groups when defined on symptomatic outcome (psychotic versus non-psychotic, schizophrenia versus transitioned without schizophrenia). Similarly, baseline OI ability did not differentiate between these groups, and there was no significant change in OI ability over time. However, we were able to demonstrate significant differences in OI ability at baseline when the groups were defined on the basis offunctionaloutcome.

These findings are at odds with those from our previous study ( Brewer et al., 2003 ), where we found that baseline OI deficits were present in the UHR patients who subsequently developed schizophrenia at 18 month follow-up, compared to those who developed a non-schizophrenia psychosis. Indeed, the effect size in that study was 0.74, indicating a sizeable discrepancy in performance. However, with a longer follow-up period, more participants have transitioned. In our original study, 22 of the 81 UHR participants had developed psychosis at the time of analysis, whereas in the current paper, 31 have now transitioned. Furthermore, diagnostic instability has resulted in changes in the numbers with a diagnosis of schizophrenia—12 in the original paper, but only 11 now, and only six people had a schizophrenia diagnosis at both time points. Despite this, although non-significant, it is worth noting that in the current study the effect size of the difference in UPSIT performance between those who did and did not have a diagnosis of schizophrenia at follow-up was moderate and stable (0.44 at follow-up and 0.43 at baseline), indicating that we may have been underpowered to detect OI deficits specific to schizophrenia. We would argue against discarding this idea without further study.

Our findings do support our earlier work, and the work of Woodberry et al. (2010) , in that OI deficits do not distinguish between UHR patients on the basis of transition to psychosis. Where we extend previous work is our finding that OI ability at baseline is significantly worse in those with poor functional outcome many years later—indeed, this has the largest effect size of any comparison presented.

Overall, these results suggest that while risk for schizophrenia may be associated with deficits in OI ability, such deficits are more usefully associated with later poor psychosocial functioning. This fits with our previous work in this population showing that poor cognitive performance is a better predictor of later poor functioning than transition to psychosis (or schizophrenia) (Lin et al, 2011 and Lin et al, 2013). Furthermore, when we re-ran the analysis covarying for IQ, there was still a significant difference in OI ability between the good and poor outcome groups. This suggests that there is a unique contribution of olfactory identification to the prediction of poor functioning, and that it is not just a matter of global intellectual impairment. We suggest that this is the result of compromise in normal frontal lobe development (for review see Brewer et al., 2006 ), thereby interfering with the development of olfactory identification. This would also be consistent with the neurodevelopmental hypothesis of schizophrenia ( Pantelis et al., 2005 ).

We investigated whether the association between smell identification ability at baseline and outcome was driven solely by the transitioned cases with poor outcome (e.g. chronic schizophrenia) by reanalysing transitioned and non-transitioned cases separately. The result remained significant for the non-transitioned group only. The lack of a significant finding in the UHR-P group is likely due to a lack of statistical power to detect a significant result. Nevertheless, the findings suggest that the association between poor outcome and UPSIT at identification as UHR is not solely driven by transitioned cases with a very poor outcome.

Our results here suggest that OI performance is stable over time, even after the onset of psychosis. However, the baseline performance of this UHR sample is about three points higher than that of first-episode psychosis patients recruited from the same clinical centre ( Brewer et al., 2001 ). This suggests there may actually be a decline in performance during the transition from UHR to frank psychosis, which was not demonstrated in our exploratory analyses. Inspection of the data does provide some support for this—the difference between mean UPSIT score at baseline and follow-up for the UHR-NP group was − 0.03, but − 1.61 for the UHR-P group. This decline in the psychotic group was particularly large in those who developed schizophrenia (− 3.5), but the sample was too small (n = 4) to allow meaningful analysis. However, it may be that UHR and FEP samples are not directly comparable, and the discrepancy in OI ability is due to other illness-related factors (including medication). Clearly, longitudinal data from larger UHR cohorts, and from healthy controls, are required to explore them further.

In our previous studies of chronic schizophrenia ( Brewer et al., 1996b ), first-episode psychosis ( Brewer et al., 2001 ) and UHR ( Brewer et al., 2003 ) patients, a consistent relationship was reported between OI and negative symptoms. Similarly, Good et al. (2006) found that OI deficits appear to serve as one marker for a subtype of patients who are characterised by less remission, more negative symptoms and greater cognitive impairment. In that study, OI deficits appear to be most strongly related to overall social and occupational functioning, and specifically to maintaining employment and meeting basic needs. Furthermore, Strauss et al. (2010) reported reduced olfactory hedonic judgement in the deficit syndrome diagnostic subtype, along with a negative relationship between negative symptom severity and OI ability. However, in the current study, no relationship between OI and negative symptoms was found in any group. This may be because of the overall low levels of negative symptoms when compared to other clinical populations, but is interesting given that low OI performance was associated with poor functioning at outcome.

The results should be considered in light of some limitations. As already mentioned, the sample sizes were very small for some analyses and the conclusions that can be drawn from these are limited. Furthermore, it is also important to acknowledge the fact that clinical information for 22 participants with UPSIT data at baseline was obtained from state records. While this registry is likely to capture almost all people with psychotic illness, it is possible that some participants may have been misclassified. A final limitation is that only participants recruited into research in early cohorts at the PACE Clinic had baseline UPSIT data. We know that there are differences between participants recruited in early cohorts and those in later cohorts, including the rate of transition to psychosis (Nelson et al, 2013 and Yung et al, 2007). Thus results may not be generalizable across cohorts.

It is also possible that UPSIT deficits are epiphenomenal, reflecting more generalised endophenotypic neurocognitive impairment, rather than representing a new endophenotype in itself ( Compton et al., 2006 ). Further research should concentrate on the relative predictive power of OI deficits for poor functional outcome versus transition to schizophrenia, and try to more definitively determine whether there is a decline in OI ability over time.

Role of funding source

No funding source played any role in the collection, analysis, interpretation or publication of data.

Contributors

WJB, ARY, SJW and AL conceived the study; AL and BN collected the data; AL and SJW analysed the data; AL, WJB and SJW wrote the first draft and all authors contributed to revising the manuscript prior to submission.

Conflict of interest

No conflict of interest in relation to this manuscript exists.

Acknowledgements

This work was supported by NHMRC Program Grants (#350241 and 566529) and the Colonial Foundation, Australia. SJW and WJB were supported by the National Health and Medical Research Council (NHMRC) Career Development Awards. BN was supported by a Ronald Phillip Griffith Fellowship and a NARSAD Young Investigator Award. CP and ARY are the recipients of NHMRC Senior Principal and Senior Research Fellowships, respectively.

References

  • Andreasen, 1982 N. Andreasen. Scale for the Assessment of Negative Symptoms (SANS). (Iowa, 1982)
  • Barbato et al., 2012 M. Barbato, M.A. Colijn, R.S. Keefe, D.O. Perkins, S.W. Woods, K.A. Hawkins, B.K. Christensen, J. Addington. The course of cognitive functioning over six months in individuals at clinical high risk for psychosis. Psychiatry Res.. 2012;206(2):195-199
  • Brewer et al., 1996a W. Brewer, S. Francey, A. Yung, D. Velakoulis, V. Anderson, P. McGorry, D. Copolov, B. Singh, C. Pantelis. Olfactory and neuropsychological deficits in high-risk & first episode psychosis. Eur. Neuropsychopharmacol.. 1996;6:s136
  • Brewer et al., 1996b W.J. Brewer, J. Edwards, V. Anderson, T. Robinson, C. Pantelis. Neuropsychological, olfactory, and hygiene deficits in men with negative symptom schizophrenia. Biol. Psychiatry. 1996;40:1021-1031
  • Brewer et al., 2001 W.J. Brewer, C. Pantelis, V. Anderson, D. Velakoulis, B. Singh, D.L. Copolov, P.D. McGorry. Stability of olfactory identification deficits in neuroleptic-naive patients with first-episode psychosis. Am. J. Psychiatry. 2001;158:107
  • Brewer et al., 2003 W. Brewer, S. Wood, P. McGorry, S. Francey, L. Phillips, A. Yung, V. Anderson, D. Copolov, B. Singh, D. Velakoulis, C. Pantelis. Impairment of olfactory identification ability in individuals at ultra-high risk for psychosis who later develop schizophrenia. Am. J. Psychiatry. 2003;160:1790
  • Brewer et al., 2006 W. Brewer, C. Panelis, C. DeLuca, S. Wood. Olfactory processing and brain maturation. W. Brewer, D. Castle, C. Pantelis (Eds.) Olfaction and the Brain (Cambridge University Press, 2006)
  • Compton et al., 2006 M.T. Compton, L. McKenzie Mack, M.L. Esterberg, Z. Bercu, A.D. Kryda, L. Quintero, P.S. Weiss, E.F. Walker. Associations between olfactory identification and verbal memory in patients with schizophrenia, first-degree relatives, and non-psychiatric controls. Schizophr. Res.. 2006;86:154-166
  • Corcoran et al., 2005 C. Corcoran, A. Whitaker, E. Coleman, J. Fried, J. Feldman, N. Goudsmit, D. Malaspina. Olfactory deficits, cognition and negative symptoms in early onset psychosis. Schizophr. Res.. 2005;80:283-293
  • Doty et al., 1984 R.L. Doty, P. Shaman, M. Dann. The University of Pennsylvania Smell Identification Test: a standardized microencapsulated test of olfactory function. Physiol. Behav.. 1984;32:489-502
  • First et al., 1997 M.B. First, M. Gibbon, R.L. Spitzer, J.B.W. Williams. Structured Clinical Interview for DSM-IV Axis I Disorders. (American Psychiatric Publishing, Inc, Arlington, 1997)
  • Gill et al., 2014 K.E. Gill, E. Evans, J. Kayser, S. Ben-David, G.B. Messinger, G. Bruder, D. Malaspina, C.H. Corcoran. Smell identification in individuals at clinical high risk for schizophrenia. Psychiatry Res.. 2014;220:201-204
  • Goldman et al., 1992 H.H. Goldman, A.E. Skodol, T. Lave. Revising axis V for DSM-IV: a review of measures of social functioning. Am. J. Psychiatry. 1992;149:1148-1156
  • Good et al., 2006 K.P. Good, D. Whitehorn, Q. Rui, H. Milliken, L.C. Kopala. Olfactory identification deficits in first-episode psychosis may predict patients at risk for persistent negative and disorganized or cognitive symptoms. Am. J. Psychiatry. 2006;163:932-933
  • Good et al., 2010 K.P. Good, P. Tibbo, H. Milliken, D. Whitehorn, M. Alexiadis, N. Robertson, L.C. Kopala. An investigation of a possible relationship between olfactory identification deficits at first episode and four-year outcomes in patients with psychosis. Schizophr. Res.. 2010;124:60-65
  • Heinrichs et al., 1984 D.W. Heinrichs, T.E. Hanlon, W.T. Carpenter. The quality of life scale: an instrument for rating the schizophrenic deficit syndrome. Schizophr. Bull.. 1984;10:388-398
  • Ishizuka et al., 2010 K. Ishizuka, K. Tajinda, C. Colantuoni, M. Morita, J. Winicki, C. Le, S. Lin, D. Schretlen, A. Sawa, N.G. Cascella. Negative symptoms of schizophrenia correlate with impairment on the University of Pennsylvania Smell Identification Test. Neurosci. Res.. 2010;66:106-110
  • Kamath et al., 2014 V. Kamath, B.I. Turetsky, M.E. Calkins, C.G. Kohler, C.G. Conroy, K. Borgmann-Winter, D.E. Gatto, R.E. Gur, P.J. Moberg. Olfactory processing in schizophrenia, non-ill first-degree family members, and young people at-risk for psychosis. World J. Biol. Psychiatry. 2014;15(3):209-218
  • Kamath et al., 2012 V. Kamath, P.J. Moberg, M.E. Calkins, K. Borgmann-Winter, C.G. Conroy, R.E. Gur, C.G. Kohler, B.I. Turetsky. An odor-specific threshold deficit implicates abnormal cAMP signaling in youths at clinical risk for psychosis. Schizophr. Res.. 2012;182:280-284
  • Kayser et al., 2013 J. Kayser, C.E. Tenke, C.J. Kroppmann, D.M. Alschuler, S. Ben-David, S. Fekri, C.M. Corcoran. Olfaction in the psychosis prodrome: electrophysiological and behavioral measures of odor detection. Int. J. Psychophysiol.. 2013;90(2):190-206
  • Keshavan et al., 2009 M.S. Keshavan, A. Vora, D. Montrose, V.A. Diwadkar, J. Sweeney. Olfactory identification in young relatives at risk for schizophrenia. Acta Neuropsychiatr.. 2009;21:121-124
  • Kopala et al., 1993 L.C. Kopala, C. Clark, T. Hurwitz. Olfactory deficits in neuroleptic naive patients with schizophrenia. Schizophr. Res.. 1993;8:245-250
  • Kopala et al., 1998 L.C. Kopala, K.P. Good, E.F. Torrey, W.G. Honer. Olfactory function in monozygotic twins discordant for schizophrenia. Am. J. Psychiatry. 1998;155:134-136
  • Lin et al., 2011 A. Lin, S.J. Wood, B. Nelson, W.J. Brewer, D. Spiliotacopoulos, A. Bruxner, C. Broussard, C. Pantelis, A.R. Yung. Neurocognitive predictors of functional outcome two to 13 years after identification as ultra-high risk for psychosis. Schizophr. Res.. 2011;132:1-7
  • Lin et al., 2013 A. Lin, A.R. Yung, B. Nelson, W.J. Brewer, R. Riley, M.P.C. Simmons, S.J. Wood. Neurocognitive predictors of transition to psychosis: medium to long-term findings from a sample at ultra-high risk for psychosis. Psychol. Med.. 2013;43(11):2349-2360
  • Moberg et al., 2013 P.J. Moberg, V. Kamath, D.M. Marchetto, M.E. Calkins, R.L. Doty, C. Hahn, K. Borgmann-Winter, C.G. Kohler, R.C. Gur, B.I. Turetsky. Meta-analysis of olfactory function in schizophrenia, first-degree family members, and youths at-risk for psychosis. Schizophr. Bull.. 2013;10.1093/schbul/sbt049
  • Muthén and Muthén, 1998–2007 L.K. Muthén, B.O. Muthén. Mplus User's Guide. (, 1998–2007)
  • Nelson et al., 2013 B. Nelson, H.P. Yuen, S.J. Wood, A. Lin, D. Spiliotacopoulos, A. Bruxner, C. Broussard, M. Simmons, D.L. Foley, W.J. Brewer, S.M. Francey, G.P. Amminger, A. Thompson, P.D. McGorry, A.R. Yung. Long-term follow-up of a group at ultra high risk (“prodromal”) for psychosis: the PACE400 Study. JAMA Psychiatry. 2013;70:793-802
  • Overall and Gorham, 1962 J.E. Overall, D.R. Gorham. The Brief Psychiatric Rating Scale. Psychol. Rep.. 1962;10:799-812
  • Pantelis et al., 2005 C. Pantelis, M. Yücel, S.J. Wood, D. Velakoulis, D. Sun, G. Berger, G.W. Stuart, A. Yung, L. Phillips, P.D. McGorry. Structural brain imaging evidence for multiple pathological processes at different stages of brain development in schizophrenia. Schizophr. Bull.. 2005;31:672-696
  • Roalf et al., 2006 D. Roalf, B. Turetsky, K. Owzar, C. Balderston, S. Johnson, C. Brensinger, R. Gur, S. Siegel, P. Moberg. Unirhinal olfactory function in schizophrenia patients and first-degree relatives. J. Neuropsychiatry Clin. Neurosci.. 2006;18:389-396
  • Rupp, 2010 C.I. Rupp. Olfactory function and schizophrenia: an update. Curr. Opin. Psychiatry. 2010;23:97-102
  • Schwartz et al., 2000 J.E. Schwartz, S. Fennig, M. Tanenberg-Karant, G. Carlson, T. Craig, N. Galambos, J. Lavelle, E.J. Bromet. Congruence of diagnoses 2 years after a first-admission diagnosis of psychosis. Arch. Gen. Psychiatry. 2000;57:593-600
  • Strauss et al., 2010 G.P. Strauss, D.N. Allen, S.A. Ross, L.A. Duke, J. Schwartz. Olfactory hedonic judgment in patients with deficit syndrome schizophrenia. Schizophr. Bull.. 2010;36:860-868
  • Turetsky et al., 2008 B.I. Turetsky, C.-G. Hahn, S.E. Arnold, P.J. Moberg. Olfactory receptor neuron dysfunction in schizophrenia. Neuropsychopharmacology. 2008;34(3):767-774
  • Turetsky et al., 2012 B.I. Turetsky, V. Kamath, M.E. Calkins, W.J. Brewer, S.J. Wood, C. Pantelis, L.J. Seidman, D. Malaspina, K.P. Good, L.C. Kopala. Olfaction and schizophrenia clinical risk status: just the facts. Schizophr. Res.. 2012;139:260-261
  • Woodberry et al., 2010 K.A. Woodberry, L.J. Seidman, A.J. Giuliano, M.B. Verdi, W.L. Cook, W.R. McFarlane. Neuropsychological profiles in individuals at clinical high risk for psychosis: relationship to psychosis and intelligence. Schizophr. Res.. 2010;123:188-198
  • Woodberry et al., 2013 K.A. Woodberry, W.R. McFarlane, A.J. Giuliano, M.B. Verdi, W.L. Cook, S.V. Faraone, L.J. Seidman. Change in neuropsychological functioning over one year in youth at clinical high risk for psychosis. Schizophr. Res.. 2013;146(1):87-94
  • Yung et al., 2003 A.R. Yung, L.J. Phillips, H.P. Yuen, S.M. Francey, C.A. McFarlane, M. Hallgren, P.D. McGorry. Psychosis prediction: 12-month follow up of a high-risk. Schizophr. Res.. 2003;60:21-32
  • Yung et al., 2005 A.R. Yung, H.P. Yuen, P.D. McGorry, L.J. Phillips, D. Kelly, M. Dell'Olio, S.M. Francey, E.M. Cosgrave, E. Killackey, C. Stanford. Mapping the onset of psychosis: the Comprehensive Assessment of At Risk Mental States. Aust. N. Z. J. Psychiatry. 2005;39:964-971
  • Yung et al., 2007 A.R. Yung, H.P. Yuen, G. Berger, S. Francey, T.C. Hung, B. Nelson, L. Phillips, P. McGorry. Declining transition rate in ultra high risk (prodromal) services: dilution or reduction of risk?. Schizophr. Bull.. 2007;33:673-681

Footnotes

a Telethon Kids Institute, The University of Western Australia, Australia

b Orygen Youth Health Research Centre, Centre for Youth Mental Health, University of Melbourne, Victoria, Australia

c Institute of Brain, Behaviour and Mental Health, University of Manchester, UK

d Melbourne Neuropsychiatry Centre, Department of Psychiatry, University of Melbourne and Melbourne Health, Victoria, Australia

lowast Corresponding author at: Telethon Kids Institute, 6008 Perth, Australia. Tel.: + 61 8 9489 7777.