Meta-Analysis of Olfactory Dysfunction in 22q11.2 Deletion Syndrome

Abstract

A quantitative review of literature concerning olfactory function in 22q11.2 deletion syndrome (22q11DS) patients was performed detailing the scope/magnitude of deficits and probing possible moderators. We searched MEDLINE, EMBASE and PubMed to identify studies for inclusion. Effect sizes were based on differences in psychophysical olfactory tests between 22q11DS patients (n=194) and typically developing comparison subjects (n=466). 22q11DS patients exhibited marked olfactory dysfunction (d=−1.11, 95% CI=−1.29<δ<−0.92) that was homogeneous (p=0.86). Diffuse olfactory deficits were seen which were not moderated by age or sex. 22q11DS patients exhibit large/diffuse deficits in olfactory function that are of a similar magnitude to observed neuropsychological impairments.

1. Introduction

The 22q11.2 deletion syndrome (22q11DS) is a neurogenetic disorder resulting from a microdeletion on the long (q) arm of chromosome 22 that occurs de novo in approximately 90% of cases and is inherited in approximately 5–10% (Jonas et al., 2014). Olfaction has been extensively investigated in a number of neuropsychiatric (Turetsky et al., 2009; Larsson et al., 2017) and neurodegenerative disorders (Mesholam et al., 1998; Roalf et al., 2017) using olfaction as diagnostic or prognostic marker. Psychosis is found in approximately 23–30% of late adolescents and young adults with 22q11DS (Murphy et al., 1999; Bassett et al., 2005), and olfactory deficits are common in adults with schizophrenia and youth at clinical high risk for psychosis (Moberg et al.. 2014). Thus, olfactory deficits may allow for the identification of incipient psychosis in a subset of 22q11DS patients. In the first study of olfactory processing in 22q11DS, Sobin and colleagues (Sobin et al., 2006) reported that young children with 22q11DS demonstrated significant deficits in odor identification, relative to their healthy siblings, as measured by the University of Pennsylvania Smell Identification Test (UPSIT). The authors concluded that olfactory dysfunction is ubiquitous in 22q11DS across the age range, suggesting that olfaction has a central role in the phenotypic characterization of 22q11DS. In the current study, we conducted a comprehensive meta-analysis of existing studies examining psychophysical olfactory function in patients with 22q11DS. A meta-analytic approach allowed for the combination of results across studies to provide a more powerful estimate of true population differences.

2. Methods

This systematic review and meta-analysis were conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher et al., 2009).

2.1. Search Strategy

Two authors independently retrieved relevant studies (PJM, EAM) through computerized literature search using MEDLINE, PubMed, and EMBASE databases to find relevant studies with the search terms (“22q11 Deletion Syndrome,” OR “Velo-Cardio-Facial syndrome,” OR “DiGeorge Syndrome”) AND (“olfactory,” OR “olfaction,” OR “smell.”) Additionally, a thorough manual review of articles was performed utilizing cross-references from identified original articles and reviews.

2.2. Eligibility criteria

Inclusion of studies followed these criteria: 1) were original English language research articles that enrolled human subjects, 2) utilized standard or experimental psychophysical-based measures of olfactory function in patients with 22q11DS, 3) had an age-matched comparison group of healthy typically developing or sibling participants with no history of 22q11DS, and 4) provided data or statistical information that allowed for the calculation of effect size. Following completion of data extraction, the senior author (DRR) provided a final check of all the data.

2.3. Study selection and data extraction

Manuscript titles, abstracts and full-texts were independently screened by the first and third author. Disagreements were resolved by discussion and consensus. Items extracted included: 1) study details (author, publication year), 2) participant characteristics (sample size, age at time of testing, sex) and, 3) psychophysical test scores from measures of odor detection threshold sensitivity, odor discrimination, and odor identification. After two authors completed the data extraction, another two authors (BIT, CGK) checked all the data.

2.4. Statistical Analyses

Comprehensive Meta-Analysis (Biostat, NJ; Version 2.2.064) was used for analysis. For the overall analysis, mean differences in psychophysical measures of olfactory function between individuals with 22q11DS (n=194) and typically developing healthy comparison subjects (n=466) were standardized by calculating Cohen’s d, which is the difference between two raw means divided by the pooled standard deviation (SD). When means and SDs were not available, d was calculated from reported univariate F-tests, t-statistics, chi-square or p-values. In order to control for differences in sample size during effect size computation, studies were weighted according to their inverse variance estimates. Prior convention has classified effect sizes as: small (0.2≤d<0.5), medium (0.5≤d<0.8), large (d≥0.8) based on these methods (Cohen, 1988). Worse performance of 22q11DS subjects relative to controls on olfactory measures were defined by negative effect size values for ease of understanding. Confidence intervals (CI) and z-transformations of the effect size were utilized to determine whether mean effect sizes were statistically significant. Effect size homogeneity was assessed using the Cochran Q-statistic across studies for each olfactory domain. Random effect models were used to compute the significance levels of the mean effect sizes. In the case of overall effect-size heterogeneity, potential moderators were analyzed using the Q-statistic for categorical data and meta-regression techniques for continuous variables. Publication bias was evaluated graphically through examination of the funnel plot as well as calculation of an adjusted rank-correlation test, according to the methods of Begg and Mazumdar (Begg & Mazumdar, 1994), Egger and colleagues (Egger et al. 1997) and Duval and Tweedy (Duval & Tweedy, 2000).

3. Results

3.1. Study selection and study characteristics

A total of 15 articles were identified through the initial three-database search, and 11 articles that met criteria were included for systematic review and 5 articles were eligible for meta-analysis (Figure 1).

Figure 1.

Flow diagram for selection of studies (PRISMA flow diagram).

Included studies were published between 2006 and 2017 and were cross-sectional designs. Detailed characteristics of the included studies are shown in Table 1.

Table 1.

Studies included in meta-analysis.

Author	Sample	Age (years)	Sex (M/F)	Olfactory Measures
Buckley et al. 2017	50; 22q11DS 14; TD* controls	32 ± 11 39 ± 19	19 / 31 4 / 10	UPSIT‡; Identification
Butcher et al. 2017	13; 22q11DS 12; TD controls	41.5 ± 7.3 42.4 ± 8.7	8 / 5 8 / 4	UPSIT; Identification
Romanos et al. 2010	27; 22q11DS 27; TD controls	10.5 ± 2.7 11.0 ± 1.11	/ 8 / 7	Sniffin’ Sticks; 1) Identification, 2) Discrimination and, 3) Detection Threshold
Sobin et al. 2006	39; 22q11DS 23; sibling controls	9.7 ± 2.3 9.8 ± 2.9	19 / 20 7 / 16	UPSIT; Identification
Tang et al. 2016	31; 22q11DS 77; TD controls	16.3 ± 9.0 27.7 ± 11.7	14 / 17 38 / 39	Sniffin’ Sticks§; 1) Identification and 2) Discrimination. Detection Threshold to odorants: 1) Citralva (3,7dimethyl-2,6-octadienenitrile) and 2) Lyral (4-(4-hydroxy-4methyl-pentyl)-3cyclohexene-1carboxyaldehyde)

^*= Typically Developing Control;

^‡= University of Pennsylvania Smell Identification Test;

^§= Sniffin’ Sticks Olfactory Test.

3.2. Meta-analysis

Analysis of effect sizes across olfactory domains for the 22q11DS patients, relative to control subjects, revealed effect sizes in the large range of magnitude (k=10, (d=−1.11, 95% CI=−1.29<δ<−0.92) that were homogeneous (Q_B[9]=4.69, p=0.86). The latter indicates that diffuse deficits of a similar magnitude are seen across all three domains of olfactory functioning relative to typically developing controls. Meta-regression revealed no significant relationship of sex composition (k=6, Z=0.18, p=0.86) or age (k=6,Z=1.41, p=0.16) with effect size.

3.4. Publication Bias

Analysis for the presence of possible publication bias revealed an asymmetric funnel plot (Supplemental Materials) and significant Begg (p=0.007) and Egger (p=0.002) tests. Considering potential “file drawer” or publication bias in the 22q11DS literature, we calculated the potential missing studies using the Duval and Tweedie “trim and fill” method (Duvall & Tweedie, 2000). This procedure trimmed 1 study with an adjusted point estimate of −1.1, slightly larger than the original estimate of −1.11. In addition, calculation of a fail-safe N indicated that 321 “null” studies would have to be located and included in order for the combined 2-tailed p-value to exceed 0.05. As such, these analyses support the notion that the current meta-analytic data accurately represent the extant literature concerning olfactory function in patients with 22q11DS.

4. Discussion

This systematic meta-analytic review sought to further quantify both the magnitude and scope of olfactory deficit seen in individuals with 22q11.2 deletion syndrome and to identify any possible moderator variables. Results revealed very large and homogeneous olfactory deficits that were seen across a variety of psychophysical tasks. No significant mitigation of this effect by age or sex composition of the sample was observed. Consistent with the latter findings, Sobin and colleagues (Sobin et al. 2006) first reported that olfactory deficits appear to be common in 22q11DS as 68% of the 22q11DS patients were reported to exhibit scores ≥ two standard deviations below the UPSIT standardization sample mean as opposed to only 13% in their typically developing siblings. Notably, 22q11DS patients with velopharyngeal insufficiency (VPI) did not differ significantly in odor identification scores relative to those without, suggesting that olfactory deficit could not be explained solely by structural (e.g. cleft palate) or neuromuscular problems (see also Buckley et al., 2017). While the mechanism responsible for the olfactory deficit in 22q11DS has not been identified, abnormalities in prefrontal cortical development and dopaminergic dysregulation in olfactory eloquent brain regions such as prefrontal and orbitofrontal cortex (Kates et al., 2011; Jonas et al., 2014; Schmitt et al, 2015) and alterations in brain regions relevant to both olfaction and psychiatric symptomatology/psychosis risk have been reported (i.e., abnormalities in cortical thickness in inferior/medial frontal cortex, orbitofrontal cortex and temporal lobe) (Jalbrzikowski et al., 2013; Schmitt et al., 2015).

In one of the first studies to examine the relationship between psychosis-spectrum symptoms and olfaction in 22q11DS, Tang and colleagues (Tang et al., 2016) described a significant relationship between positive and negative symptoms and odor discrimination in 22q11DS, perhaps indicating a sensitivity of psychophysical measures to emerging psychiatric symptoms in this disorder. As olfactory measures are impaired in youth at clinical high risk for psychosis (Moberg et al., 2014), the use of these tests in 22q11DS may represent a simple and cost-effective way of screening for incipient psychosis. To this end, it is notable that the Danish 22q11 research initiative (Schmock et al., 2015) has included a brief test of odor identification in their neurocognitive battery to capture a domain of function not typically covered in a standard neuropsychological battery.

In conclusion, very large deficits in olfactory function are present in 22q11DS. Deficits are seen across the domains of detection threshold sensitivity, discrimination and identification, with only nominal differences seen among the three. Age or sex composition of the sample did not have significant moderating effects on the observed olfactory deficits. In light of these findings, simple-to-administer tests of olfactory function should be encouraged in the assessment of 22q11DS.

Highlights:

• Olfactory dysfunction is ubiquitous in 22q11DS across the age range, suggesting it has a central role in the phenotypic characterization of 22q11DS

• Quantitative review of effect sizes across the extant olfactory literature in 22q11DS revealed large and diffuse deficits in olfactory function

• No significant moderation of olfactory deficit by age or sex was seen

• Simple-to-administer tests of olfactory function should be encouraged in the assessment of 22q11DS