Effects of Presbyphagia on Oropharyngeal Swallowing Observed during Modified Barium Swallow Studies

Abstract

OBJECTIVES:

Understanding how aging impacts swallowing can help differentiate typical from atypical behaviors. This study aimed to quantify age-related swallowing alterations observed during a modified barium swallow study.

DESIGN:

Cross-sectional study.

SETTING:

Adult fluoroscopy suite in a metropolitan hospital at an academic center.

PARTICIPANTS:

195 healthy adults distributed across 3 age categories: 21–39; 40–59; 60+ years.

MEASUREMENTS:

17 physiologic components of swallowing across three functional domains (oral, pharyngeal, esophageal), including summed composite scores (Oral Total [OT] and Pharyngeal Total [PT]), from the validated and standardized Modified Barium Swallow Impairment Profile.

RESULTS:

Most components (65%) demonstrated no impairment (scores of “0”). The odds of a worse (higher) score increased significantly with age for: Tongue Control during Bolus Hold, Hyolaryngeal Movement, Laryngeal Closure, Pharyngeal Contraction, and Pharyngoesophageal Segment Opening. OT and PT scores for 40–59-year-olds were worse than the youngest group (p=.01 and p <.001, respectively). Adults 60+ years had significantly worse PT scores among all groups (p-values <.01).

CONCLUSION:

Oropharyngeal swallowing physiology evolves as healthy adults age and should be considered during clinical decision-making.

Introduction

Presbyphagia denotes age-related changes on swallowing performance (1). Previous studies employing modified barium swallow studies (MBSSs) to assess swallowing performance have demonstrated older adults had: increased mastication duration (2); initiated the pharyngeal swallow deeper in the pharynx (3, 4); increased oropharyngeal transit times (5); and reductions in lingual strength, tongue base retraction, and hyolaryngeal movement (6–8). Yet, the impact of these changes on swallowing safety (ingested material does not enter airway) and efficiency (ingested material is timely transported without material remaining) are equivocal.

For example, Daggett and colleagues (9) and Yoshikawa and colleagues (10) observed penetration events more often in older adults, while this was not corroborated by a more recent study by Garand et al. (11). Yoshikawa et al. (10) also observed greater oropharyngeal residue amounts in older adults, but such differences were not observed in a recent study by Mulheren et al. (12). Contradictory results are likely due to the methodological differences, such as differences in operational definitions of swallow-related measures, administration of a limited number of swallowing trials with a limited range of swallowing tasks (volumes, viscosities), age comparisons of only the extreme ends of adulthood, inclusion of only one gender, and sample size.

Alterations in swallowing performance resulting from aging have been postulated to result from various factors, such as loss of dentition, decline in oropharyngeal sensitivity and changes in cervical spine curvature (13). Mounting evidence supports that sarcopenia, in particular, may play a critical role. Sarcopenia is used to describe an aging syndrome characterized by progressive and generalized loss of skeletal muscle mass and strength (14). Sarcopenic-induced changes specifically in the bulbar musculature, therefore, may impact swallowing (15).

Additionally, there is increasing interest in the relationship between cognition and its role in swallowing, particularly in the aging population, since evidence suggests variations in neural activation and reorganization occur in older adults as a compensatory response to aging (16, 17). Yang et al. (18) observed increased odds of dysphagia in males with mild cognitive impairment, which suggests sex differences may also exist in age-related decline that may impact swallowing function. Yet, the relative contributions of the influences of age-related muscle loss in bulbar musculature or age-related alterations in brain structures on swallowing performance, particularly in the setting of disease, remain unknown. For example, Krishnamurthy and colleagues (17) failed to observe an age effect of a dual-task paradigm on swallowing in healthy adults, yet, residue significantly worsened in cognitive dual-task conditions in older patients with Parkinson’s disease (19).

Although age-related swallowing changes are not considered pathological, the clinical implications are still meaningful and warrant further investigation to enhance understanding of typical swallowing behaviors and the impact of aging on them. For example, alterations in swallowing function may lessen functional reserve and require the individual to perform at greater capacity than previously required (20). The difference between the ability to produce maximum output and task performance is referred to as homeostenosis (21). If homeostenosis impacts swallowing, this may impact health and well-being, particularly in the setting of disease/injury (22). Maeda and Akagi demonstrated sarcopenia as an independent risk factor for dysphagia in older adults, although causality remains unknown (23). This was highlighted in a study by Sporns and colleagues (13) in a retrospective cohort of 73 acute stroke patients, which demonstrated not only a decline in volume of various muscles involved in swallowing as age increased, but this was correlated with dysphagia severity. Thus, pre-existing muscle loss placed older individuals at increased risk for dysphagia after stroke onset (23).

Further, these age-related changes may result in variability in clinical practice if clinicians are unable to delineate typical variations from true impairment with large-scale implications, including over-referrals and unnecessary (and even invasive) diagnostic and treatment procedures, contributing to increased healthcare costs (24). Therefore, distinguishing typical age-related swallowing changes from potential early manifestations of dysphagia is a critical first step when evaluating swallowing physiology. The purpose of our study was to quantify the impact of aging on swallowing physiology in a large sample of healthy adults using standardized metrics (the Modified Barium Swallow Impairment Profile [MBSImP^™]). Based on previous works (25–27), we anticipated differences in older individuals compared to their younger counterparts in physiologic components primarily involving lingual (Lingual Transport, Oral Residue) and pharyngeal musculature (Pharyngeal Stripping Wave, Pharyngeal Contraction, and Pharyngeal Residue).

Materials and methods

This cross-sectional study received Institutional Review Board approval. Participants provided informed consent prior to study participation.

Study Participants

Eligible participants were community-dwelling adults aged 21 years and older without neurological insult/disease (e.g., stroke, Parkinson’s disease, dementia), pulmonary disease (e.g., chronic obstructive pulmonary disease), head and neck cancer, large hiatal hernia (>2 cm), and history of anterior neck or spinal surgery. All participants self-reported a full oral diet without restrictions or use of compensations (corresponding to International Dysphagia Diet Standardisation Initiative [IDDSI] levels 0 [thin liquids] and 7 [regular solids]) (28). All participants denied a history of swallowing problems or aspiration pneumonia.

Modified Barium Swallow Study (MBSS) Procedure

Participants underwent a MBSS which employed continuous fluoroscopy with images digitally captured at a rate of 30 frames per second (Digital Swallowing Workstation Model 7100, Kay Elemetrics Corp, Lincoln Park, New Jersey, USA; TIMS Dicom System, TIMS Medical, Chelmsford, Massachusetts, USA). Each MBSS was completed with collaboration between one of five speech-language pathologists, each reliably trained in employing the MBSImP standardized protocol (29, 30), and one radiology assistant. The MBSImP protocol includes 12 swallows in 2 viewing planes (10 swallows in the lateral plane, and 2 swallows in the anteroposterior plane). Standardized consistencies of commercially prepared barium (Varibar^®, E-Z-EM, Inc.; 40% w/v ratio) were administered: thin liquid (<15 centipoise [cps], IDDSI Level 0 – thin), nectar liquid (<150–450 cps, IDDSI Level 2 – mildly thick), thin honey liquid (800–1,800 cps, IDDSI Level 3 – moderately thick), pudding/paste (4,500–7,000 cps, IDDSI Level 4 – extremely thick), and a half-portion of a shortbread cookie (Lorna Doone, Nabisco, IDDSI level 7 – regular) coated with 3 mL barium paste. Thin and nectar-thickened liquid tasks are administered in increasing volumes, including 5 mL via teaspoon, self-controlled cup sip, and sequential swallow tasks. The remaining consistencies were administered via teaspoon (5 mL), except the solid task. Average (SD) fluoroscopy time was 1.4 (0.3) minutes.

MBSImP Scoring

The first swallow task (5 mL thin) acclimates the participant and was not scored, leaving a total of 11 swallows eligible for scoring. Frame-by-frame playback capabilities of QuickTime Player allowed scoring of swallows using the published MBSImP scoring tool available online (29–31). Seventeen physiologic components were scored using an ordinal scale, with the smallest range of scores being 0 to 2, and the largest range from 0 to 4. An Overall Impression (OI) score is derived for each component reflecting the highest score observed across swallow tasks (as appropriate). Two experienced authors (KG and KD) completed consensus scoring after achieving >90% inter-rater reliability.

Composite scores representing Oral Total (OT) and Pharyngeal Total (PT) were also derived. Consistent with clinical practice, OT was calculated as the sum of OI scores for Components 1–6, with a score of “1” for Components 1 (Lip Closure) and 5 (Oral Residue) treated as “0”. Similarly, PT was calculated as the sum of OI scores for Components 7–16, with a score of “1” for Components 15 (Tongue Base Retraction) and 16 (Pharyngeal Residue) treated as “0”. Scores of 1 for Components 1, 5, 15 and 16, indicating “trace”, are not included in the calculation of composite scores as these are included to enhance visual discrimination between structural coating and residue collection (29).

Statistical analysis

Power Analysis

Age categories were a priori defined based on previous investigations (32–34). A sample size calculation and power analysis revealed that 70 participants per age category (21–39 years; 40–59 years; 60–79 years; and 80 years and older) for a total of 280 participants would yield sufficient power (≥80%) to detect small, but clinically significant, age-related trends in dichotomized MBSImP Component OI scores (lower vs higher scores) based on a continuity corrected Cochran-Armitage trend test with α = 0.05. This sample size would have also provided power exceeding 99% to detect clinically meaningful age-related trends in Composite scores (OT/PT). Despite extensive recruiting efforts, we were unable to achieve the target sample size in the oldest age category (80 years and older) primarily due to interested participants not being able to meet our stringent eligibility requirements. Therefore, all analyses were completed using three age categories; i.e., 21–39 years; 40–59 years; and 60 years and older.

Descriptive statistics were performed for all variables of interest. Analyses were performed using R version 4.0.3. The level of statistical significance was set at 0.05.

Component Scores

The Cochrane-Armitage trend test was performed at the swallow task level to determine if there was an association between age categories and a higher component score. Data were dichotomized; scores greater than 0 were considered “high” except for Components 1, 5, 15, and 16 in which scores greater than 1 were considered “high” (24, 25). A regression model was constructed for each component to determine if there was a significant relationship between age and the odds of having a high score for a task within the given component. Further, sex was considered in the modeling process to check for confounding. To account for the expected correlated nature of the task level data within a component, a generalized estimating equations (GEE) framework was used with an exchangeable correlation structure. The model parameters were interpreted using odds ratios, obtained by exponentiating the parameter estimates from the model, and corresponding 95% confidence level. For age, the parameter estimates were interpreted by 10-year changes in age. For Component 3 (Bolus Preparation/Mastication), a logistic regression model was fit to the data instead because there is only one task administered to assign an OI score. A piecewise linear model was fit to the data for Component 14 (Pharyngoesophageal Segment Opening [PESO]) due to a non-linear relationship between age and the probability of having a high score.

Composite Scores

The range of potential scores for OT and PT are 0–22 and 0–29, respectively. Pairwise t tests were used, and a Bonferroni Type-1 error correction was employed to determine significant differences in OT and PT scores across age categories. A linear regression model with sex and age included as covariates was used to assess for confounding in the modeling process.

Results

Participants

Out of 211 participants who underwent a MBSS, 16 were excluded: 10 removed due to error in fluoroscopy rate setting; 1 removed when imaging revealed structural anomaly (enlarged tonsils) which altered bolus flow; and 5 removed when imaging revealed an exclusion criterion of previous anterior neck surgery not reported during eligibility interview. Therefore, a total of 195 participants (109 females, 86 males) with a mean age (SD) of 47 (17.4) years were eligible for analyses. Demographic characteristics across age categories are provided in Table 1.

Table 1.

Participant demographics across age categories. Data are given as frequency (percent), unless otherwise noted

Variable	21–39 Years (n = 70)	40–59 Years (n = 70)	60+ Years (n = 55)	Total (N = 195)
Age (Mean ± SD)	28.2 ± 4.6	48.9 ± 6.2	68.7 ± 8.0	47 ± 17.4
Sex
Females	36 (51)	35 (50)	38 (69)	109 (56)
Males	34 (49)	35 (50)	17 (31)	86 (44)
Race
Caucasian	50 (71)	44 (63)	50 (91)	144 (74)
African American	16 (23)	23 (33)	5 (9)	44 (23)
Other	4 (6)	3 (4)	0 (0)	7 (4)
Ethnicity
Hispanic/Latino	2 (3)	1 (1)	1 (2)	4 (2)
Non-Hispanic/Latino	68 (97)	69 (99)	54 (98)	191 (98)

Physiologic Component Scores

The majority of components (11; 64.7%) were observed to have a score of “0”.

Influence of Age

When controlling for swallow task, as age increased, there was an increase in the odds of having a high score on one oral component (Tongue Control During Bolus Hold) and five pharyngeal components (Anterior Hyoid Excursion, Pharyngeal Stripping Wave, Pharyngeal Contraction, PESO, and Pharyngeal Residue) (all p-values <0.05). Overall, a 10-year increase in age was associated with a significant increase in the odds of having a higher component score. When both swallow task and gender were controlled, three additional pharyngeal components (Laryngeal Elevation, Laryngeal Vestibular Closure, and Tongue Base Retraction) reached significance. Table 2 provides parameters estimates as odds ratios for a 10-year increase in age across all components controlling for swallow task and gender.

Table 2.

Odds ratios (OR) and 95% Confidence Intervals (CI) for a 10-year increase in age across MBSImP Components. Model accounts for gender and swallow task

MBSImP Component	OR (95% CI)	p - value
C1 – Lip closure	0.95 (0.72, 1.25)	0.70
C2 – Tongue control during bolus hold	1.22 (1.06, 1.41)	0.01
C3 – Bolus preparation/mastication	1.10 (0.88, 1.38)	0.42
C4 – Bolus transport/lingual motion	1.07 (0.90, 1.26)	0.45
C5 – Oral residue	1.02 (0.91, 1.16)	0.70
C6 – Initiation of pharyngeal swallow	1.02 (0.91, 1.14)	0.78
C7 – Soft palate elevation	1.24 (0.70, 2.21)	0.46
C8 – Laryngeal elevation	1.20 (1.01, 1.42)	0.04
C9 – Anterior hyoid excursion	1.28 (1.09, 1.50)	<.001
C10 – Epiglottic movement	1.26 (0.82, 1.94)	0.28
C11 – Laryngeal vestibular closure	1.41 (1.05, 1.88)	0.02
C12 – Pharyngeal stripping wave	1.66 (1.20, 2.28)	<.001
C13 – Pharyngeal contraction	1.38 (1.09, 1.77)	0.01
C14 – Pharyngoesophageal segment openinga
<45 years	0.85 (0.60, 1.21)	0.37
>45 years	2.05 (1.55, 2.70)	<.001
C15 – Tongue base retraction	1.19 (1.01, 1.40)	0.04
C16 – Pharyngeal residue	1.34 (1.15, 1.56)	<.001
C17 – Esophageal clearance in the upright position	0.88 (0.76, 1.02)	0.10

^a.Piecewise linear model was fit

Composite Scores

The overall mean (SD) for OT and PT were 5.5 (2.3) and 4.7 (2.6), respectively.

Influence of Age

Significant differences were observed in OT across age categories, including the middle age category having higher OT scores than the youngest age category (p = 0.01). Significant differences were observed in PT across age categories, including the oldest age category having higher PT scores than the middle and youngest age categories (p<0.001 and p = 0.001, respectively). Results remained unchanged after controlling for sex (p-values <.01).

Boxplots of the OT and PT Composite scores across age groups are provided in Figure 1. When examining quartile distributions of Component OI scores within the Oral Domain across age categories, a higher median score was observed in the middle and oldest age categories (Mdn = 3) relative to the youngest age category (Mdn = 2) for Component 6 (Initiation of Pharyngeal Swallow) (Supplementary Material, Table 1). Within the Pharyngeal Domain, Components 15 (Tongue Base Retraction) and 16 (Pharyngeal Residue) had higher median OI scores in the older age categories relative to the youngest (i.e., median scores of 2 in the oldest age categories compared to a median score of 0 in the youngest age category for both components) (Supplemental Material, Table 1).

Figure 1.

Distribution of Oral Total (A) and Pharyngeal Total (B) scores across age categories

* p = .01; ** p = .001; *** p < .001.

Discussion

We quantified age-related changes in swallowing physiology using a prevalent approach for measuring MBSS observations (MBSImP) in 195 healthy adults. These unique results fill a critical gap in the literature comprised of studies with small sample sizes, limited swallow tasks, and measures not commonly employed in clinical practice, all of which made it difficult for clinicians to translate research findings into practice and for comparison to patients.

Physiologic Component Scores

The majority of participants were observed to have a score of 0 for most physiologic components, as was anticipated, since we employed strict eligibility criteria to reduce confounders (e.g., conditions known to have dysphagia as a prevalent consequence) that may otherwise explain study findings. The low scores observed across components support the interdependence and interrelatedness across aerodigestive structures to facilitate a safe and efficient swallow (35). For example, it is well recognized the hyoid and larynx move as a functional unit and influences PESO (at least partially) because of the mechanical traction on the cricoid cartilage when the hyolaryngeal complex moves in a superior-anterior direction during swallowing (36, 37). As a collective result, this biomechanical action influences swallow safety (laryngeal closure is achieved, and airway is displaced out of the bolus path) and efficiency (complete bolus entry into the esophagus) (38).

When examining the odds ratios for a 10-year increase in age across individual physiologic components after controlling for swallow task and gender, one Oral Domain component (Tongue Control During Bolus Hold) and eight Pharyngeal Domain components (Laryngeal Elevation, Anterior Hyoid Excursion, Laryngeal Elevation, Pharyngeal Stripping Wave, Pharyngeal Contraction, PESO, and Tongue Base Retraction) revealed significantly increased odds for having a high score as age increased. Alterations in lingual and pharyngeal musculature attributed to sarcopenic changes may explain differences in these particular component scores (26, 39).

Oral alterations attributed to aging have been previously reported, with a predominance of studies focusing on the tongue because it plays a key role in providing sensory input to trigger a pharyngeal swallow and in generating oral pressures for bolus transport (40, 41). Current study findings revealed an age-related increase for higher scores for only Component 2 (Tongue Control During Bolus Hold). Previous reports are equivocal regarding alterations in oral somatic sensations in older individuals (42–44), likely due to the method employed to assess oral sensitivity (e.g., viscosity, 2-point discrimination). Age-related alterations in lingual musculature have also been reported (e.g., reduction in muscle fiber diameter (46)) with subsequent reduction in maximum lingual pressures (39). However, age-related reductions in lingual swallowing pressures have not been supported (47).

Further, reduction in pharyngeal muscle bulk has been demonstrated in older adults with subsequent increased pharyngeal lumen size (48, 49) and associated with larger amounts of pharyngeal residue (26). Interestingly, the Pharyngeal Residue measure (Component 16) did not demonstrate an age-related trend. Again, this may be due to differences in methodology for determining the presence and amount of residue accumulation (e.g., pixel-based measures in anatomical space versus visuospatial judgement considering entire pharyngeal cavity).

Similarly, Mulheren and colleagues (12) observed higher MBSImP scores in the Pharyngeal Domain in 32 community-dwelling older adults (>62 years) compared with a historical younger cohort of 33 adults (18–29 years old). In the current study, higher scores in four pharyngeal components were similarly observed; i.e., Anterior Hyoid Excursion, Pharyngeal Stripping Wave, PESO, and Tongue Base Retraction. The current findings not only provide support for this previous investigation (12) but provide additional information from a larger sample (55 participants were 60 years and older in the current sample) and examined swallowing physiology across a wider variety of tasks commonly employed in clinical practice. Further, as reported in the Mulheren study (12), the current study did not observe higher rates of airway invasion in older adults (11). Therefore, these age-related alterations in pharyngeal swallowing function did not contribute to impaired swallow safety.

Composite Scores

OT and PT Composite scores averaged 6 and 5, respectively, with majority of participants earning composite scores of 0. Specific components appeared to influence OT and PT Composite scores within the entire study sample. For example, Components 5 and 6 (Oral Residue and Initiation of Pharyngeal Swallow, respectively) had the largest proportion of scores >1 compared to the other Oral Domain components. Similarly, Components 15 and 16 (Tongue Base Retraction and Pharyngeal Residue, respectively) predominantly contributed to the highest scores observed across the Pharyngeal Domain components.

When examining differences across age categories, interestingly, only the participants in the 40 to 59 year old category demonstrated significantly higher OT scores. Although individuals aged 60 years and older had a higher average OT score than the youngest age category (5.5 versus 4.8, respectively), differences failed to reach statistical significance (p = .31).

Older individuals had higher average PT scores than their younger aged counterparts, supported by the number of MBSImP components in the Pharyngeal Domain that demonstrated significant age-related trends (described above). Muscle loss, which can lead to sarcopenia, can begin as early as middle adulthood (4th decade of life) (50), with increasing declines in muscle mass rates as an individual ages (51, 52). Since sarcopenia can impact pharyngeal swallowing musculature, this may help to explain these alterations observed between the two older age categories relative to the youngest age category. However, although differences were observed across age categories for OT and PT, the average Composite scores for the entire study sample were lower than the upper limit thresholds reported in a previous dysphagia study with various patient populations (53). Specifically, 98% of the current study participants were below the upper OT limit of 10, and 100% of participants were below the PT upper limit of 13. These findings support the clinical use for classifying severity categories as current data falls within the “none-mild” (lowest class) severity category (53).

Study Limitations

One limitation includes the sample size and sex imbalance in the oldest age category. For example, previous works have implicated sex differences in sarcopenic changes in bulbar musculature (54) and increased likelihood of swallowing impairment in those with cognitive decline (18). Although, participants underwent a standardized and validated cognitive screening procedure, there is no guarantee that individuals with cognitive decline were not included (false negatives). The cross-sectional design also warrants caution, as some of these individuals may have presented with subclinical neurologic changes from age-related disease onset not yet recognized or identified.

Further, motor performance variability has been implicated to result from muscle loss (55). The MBSImP protocol administers each swallow task only once, because it has been shown that the limited tasks with each consistency is sufficient for capturing swallowing impairment in dysphagic adults (56). As such, the potential variability that may occur across trials of the same task was not tested current study. Additionally, dental status/oral health, nutritional status (e.g., body mass index), or physiologic markers of sarcopenia and frailty (e.g., hand grip strength) were not collected. Bahat et al. (57) observed a significantly higher number of swallowing symptoms on a validated patient-reported questionnaire (Eating Assessment Tool (58)) in older adults identified as frail, with frailty identified as an independent factor for swallowing symptoms. However, the same questionnaire (EAT-10) was collected in the same current study sample with an average score of 0.6 (59), which was well below the cut-off score of >15 used in the Bahet et al. study (57).

Clinical Implications

Despite these limitations, findings from this observational study have several advantages to assist clinicians and researchers with an interest in swallowing. First, this large sample size helps to reduce variability in measures that may have been previously reported in studies where sample sizes were small. Further, results may be immediately translated to clinical research by providing a platform for comparison of patient performance to that of healthy controls. This helps to differentiate disease impact on oropharyngeal swallowing function from potential confounders, particularly for the thousands of clinicians and clinical investigators that employ the MBSImP protocol and interpretation method.

Next, greater variability in swallowing function of older adults has previously been reported (60). It is worthwhile, then, to attend to which components are contributing to higher Composite (OT/PT) scores. Our findings demonstrated certain components had greater influence on Composite scores. As designed, any MBSImP physiologic component score must be interpreted within the context of other physiologic components for its overall impact on swallow safety and efficiency (61, 62). For example, a score of 3 on Initiation of Pharyngeal Swallow (Component 6) may be seen in healthy non-dysphagic adults, yet this score in conjunction with a score of 2 on Laryngeal Vestibular Closure (Component 11) may have dire consequences for airway protection. Further, MBSImP scores are intended for consideration within the context of the underlying disease process, patient complaint, and impact on health and quality-of-life. Even with the recent rise of publications employing machine learning to measure critical swallowing-related events efficiently and reliably (63–65), the clinician is still needed to provide meaningful interpretation of the output. Lastly, alterations in swallowing function may place an individual at risk for dysphagia should a stressor occur, such as onset of illness or disease, and age, therefore, should be recognized as potential risk factor to ensure appropriate care for our aging population (20, 22). Otherwise, adaptations in the swallowing mechanism may not only allow for successful swallowing of various foods and liquids under various circumstances in healthy individuals (66) but may also allow for successful aging (67).

Conclusion

This work supports and quantifies age-related alterations in both oral and pharyngeal swallowing physiology. Variability observed across several physiologic components of swallowing highlights the need for consideration of each component within the context of the swallow mechanism as a whole, as well as the clinical circumstances to determine impact on health and need for targeted intervention. Further investigation into swallow physiology in the oldest population (80 years and older) is substantially lacking and necessary to identify and detail potentially significant changes in swallowing physiology that may impact health and well-being of the growing US population.