Longevity and Volume Expansion of Hyaluronic Acid Dermal Fillers: A Cross-sectional 3-dimensional Magnetic Resonance Imaging Study

Abstract

Background:

This investigation aims to analyze the volumetric expansion of injectable facial fillers over time using volumetric magnetic resonance imaging (MRI) analysis.

Methods:

A retrospective cross-sectional study of 14 patients with a recorded history of filler injections underwent quantitative MRI evaluations. The volume measured was then compared with the documented volume of filler injected. This study focused on detailed volume measurements in anatomically critical zones, including the tear trough, malar region, midcheek, zygoma, and nasolabial region.

Results:

Volumetric analysis revealed that MRI-measured volumes substantially exceeded the injected volume of filler, with a ratio of 1:2.8. Specifically, total MRI-measured volumes varied from 0.7 to 38 cm³, contrasting with initial injected volumes ranging from 0.55 to 11 cm³.

Conclusions:

This is the first study to compare the initial volume of injected hyaluronic acid (HA) filler with the volume detected using 3-dimensional MRI. Certain HA fillers may undergo significant volumetric expansion after administration rather than gradual diminution. These findings contribute to our understanding of the behavior of HA fillers and may explain postinjection edema. This phenomenon should be carefully considered when planning such treatments. These results also demonstrate the benefits of using advanced imaging modalities such as MRI to aid accurate topographical analysis, assessment of residual filler, and prediction of long-term volumetric outcomes in the relevant aesthetic patient.

Takeaways

Question: How do hyaluronic acid facial fillers behave over time in terms of volume retention and expansion?

Findings: Three-dimensional magnetic resonance imaging analysis revealed that the volume of dermal filler present was, on average, 2.8 times greater than the amount initially injected, with detectable filler up to 15 years postinjection.

Meaning: Hyaluronic acid fillers not only persist far longer than expected but also expand in volume, highlighting the need for conservative dosing and long-term planning in aesthetic treatments.

INTRODUCTION

In recent years, the use of hyaluronic acid (HA) dermal fillers has continued to expand, reflecting growing global demand for minimally invasive facial rejuvenation treatments. HA fillers remain one of the most frequently performed nonsurgical aesthetic procedures worldwide. According to the International Society of Aesthetic Plastic Surgery International Survey on Aesthetic/Cosmetic Procedures performed in 2023, a total of 5.56 million HA filler procedures were performed globally, making HA fillers the second most common nonsurgical procedure after botulinum toxin injections.¹

HA fillers have a wide range of applications, with the most popular areas for treatment including the perioral region, periocular region, nasolabial folds, malar fat pads, marionette lines, jawline, and lips. The surge in demand has led to the development of a diverse range of HA-based injectable fillers, each with distinct biochemical and rheological properties.² HA dermal fillers are modified, naturally occurring HA polymers consisting of repeating disaccharide units linked by a glucuronidic β bond.³ The physicochemical structure of different filler preparations determines properties such as cohesiveness, elasticity, lift capacity, and durability.⁴

Magnetic resonance imaging (MRI) is highly effective in detecting soft tissue changes associated with high water content, which is helpful in many clinical contexts, including inflammation, abscess formation, and foreign materials within soft tissues with high water content. MRI is therefore well adapted to identifying the nature, distribution, and density of HA fillers. Additionally, MRI technology enables the creation of 3-dimensional (3D) renderings of filler distribution, volume, thickness, and patterns. This comprehensive information aids practitioners in understanding the behavior of fillers over time, anticipating potential complications, and guiding effective treatment strategies.

There are only a limited number of studies describing 3D MRI characteristics of facial dermal fillers. Girolamo et al⁵ studied a series of 26 patients with clinically diagnosed filler-related complications using contrast-enhanced MRI, focusing on filler localization without volume correlation. Tal et al⁶ analyzed MRI scans of 14 patients to determine the likely injected substance and complications. Kadouch et al⁷ reported on complications in a larger cohort of 32 patients who received MRI after filler complications.

Traditionally, it was believed that fillers gradually dissolve over time. HA fillers, in particular, are designed to have a medium-term duration, lasting between 3 and 12 months. However, recent data have suggested that fillers persist for significantly longer periods of time.^8–10

Wilde et al¹¹ examined a large cohort of 90 patients and reported on posthyaluronidase syndrome, identifying an mean duration of filler in situ of 30 months, with the upper range extending to 156 months for patients presenting for dissolution. There are also isolated case reports on dermal filler longevity that have been assessed using MRI technology. Master⁹ recently published a study evaluating facial filler longevity using MRI in 33 patients, showing detectable filler for at least 2 years, with 1 patient demonstrating persistence for 15 years. This study aims to evaluate the volumetric behavior and persistence of HA fillers over time using 3D MRI, hypothesizing that fillers demonstrate measurable expansion and prolonged detectability beyond their expected duration.¹²

METHODS

This retrospective cross-sectional study was conducted at Moorfields Eye Hospital and The Ezra Clinic (London, United Kingdom) between January 2020 and June 2024, using anonymized patient records and MRI data to assess the relationship between the injected filler volume and the corresponding MRI-measured volume at a single time point. This project was registered with the Moorfields Eye Hospital NHS Foundation Trust Clinical Audit and Effectiveness Committee as a retrospective review of anonymized clinical records and imaging (ID: MEH/2024/1868). All data were fully anonymized before analysis, and the study adhered to institutional governance and the principles of the Declaration of Helsinki.

Medical records of patients with known injected volume and location of HA filler who had undergone facial MRI with 3D reconstructions were included. Exclusion criteria included missing demographic data and any previous hyaluronidase injections. All patients underwent facial MRI for clinical indications, most commonly for evaluation of swelling or irregular contour following filler injection. No MRI scans were performed solely for research purposes.

For each patient, the following data were recorded: age, sex, date of last filler injection, date of MRI, injected filler volume (mL), and measured filler volume (3D MRI). Global volume measurements were taken, and a subanalysis of filler thickness present in different facial areas was conducted. Different facial areas were defined according to the most common areas for placement of filler: the tear trough, zygomatic arch, temple, malar region, midcheek, nasolabial line, masseter, jowl, upper and lower lip, perioral area, and mentum. Because this was a retrospective chart review, detailed information on the specific filler brands and formulations was not consistently available. This was a retrospective imaging-based study. No standardized clinical photographs were available for publication, and written consent for the use of identifiable images was not obtained. Therefore, no patient photographs are included in this report.

Statistical analysis was conducted with the Student t test using Microsoft Excel (version 16.88, Microsoft 2024).

Generating a 3D MRI Model

To generate a 3D MRI model, the following process was used:

1. Guiding sequence: The base T2 Dixon sequence was used to determine the position and extent of the filler on both axial and coronal images. Although this sequence provided a reliable reference for correlation, its nonvolumetric nature precludes its use in direct 3D modeling.

2. Segmentation process: Filler segmentation was performed on the proton density fat-suppressed 3D sequence, primarily using the thresholding tool within the 3D Slicer software. This step required careful attention, particularly in areas with susceptibility artifacts or inhomogeneous fat suppression, which were more pronounced around the periphery of the scan.

3. Creation of 3D model: Once satisfied with the 3D segmentation’s alignment with the T2 Dixon images, the software was used to convert the segmented data into a 3D model. The anatomical compartments of the face were then used as a guide to separate the 3D segmentation into distinct anatomical sites in accordance with the reference diagram.

4. Final model and quantification: After the segmentation was finalized, a 3D model video was generated by rendering the background facial tissues transparent within the software, thereby highlighting the segmented components. The software’s quantification tool was subsequently used to calculate the volume of each segmented component.

Figure 1 shows an example of the generated 3D image, demonstrating the distribution and thickness of the injected filler material.

Fig. 1.

Three-dimensional MRI example.

Data Analysis

To assess the relationship between the volume of filler injected and the volume measured by MRI, the correlation between the injected filler volume and the measured MRI volume was evaluated, producing a graph with an R² value, indicating the strength of the correlation. A paired t test was also used to evaluate the statistical significance of this relationship.

Additionally, a secondary analysis was conducted to determine whether there was a correlation between the time since the last filler injection and any observed volume changes on MRI. This was assessed using the Pearson correlation coefficient. A P value of less than 0.05 was considered statistically significant.

RESULTS

Table 1 summarizes demographic data and filler volume metrics for a group of 14 patients. The mean age of the patients was 44 years (SD 12.1 y), with 71% (10) being female. The mean age of this cohort suggests that most patients fell within a broad age range, from their early 30s to mid-50s.

Table 1.

Patient Demographics and Filler Volume

Characteristic	Value
M:F	4:10
Age at presentation, y, mean ± SD	44 ± 12.1
Time between last filler injection and MRI, mo, mean ± SD	37.43 ± 51.8
Volume injected, mL, mean ± SD	3.32 ± 3.19
Volume measured in 3D MRI, mL, mean ± SD	12.26 ± 13.19
MRI volume/volume injected, mL, mean ± SD	3.2 ± 1.4

The average time between the last filler injection and the MRI scan was 37.4 months (SD 51.8 mo), ranging from 6 to 180 months, showing substantial variability in injection-to-scan duration. The mean volume of filler injected was 3.32 mL (Fig. 2), whereas the mean volume measured with MRI was 12.26 mL, with a statistically significant difference (P = 0.006).

Fig. 2.

Boxplot of injected versus measured filler volume.

The scatter plot in Figure 3 shows the relationship between injected and measured filler volume:

Fig. 3.

Correlation between injected and measured HA filler.

1. Linear trend: The graph demonstrates a positive linear relationship between injected volume and measured volume, indicated by the trend line. As injected volume increases, measured volume also increases (R² = 0.79). This indicates a strong positive correlation.

2. Regression equation: The equation of the trend line suggests that for every unit increase in injected volume, measured volume increases by approximately 2.8 units. However, the negative y intercept (−8.7978) indicates that at very low injected volumes, measured volume may initially be underrepresented.

3. Data points: Most of the data points cluster near the trend line, but there are a few that deviate more from the line, especially at higher injected volumes. This may indicate variability in filler expansion when initial injected volume is higher.

The negative y-intercept (−8.7978) represents an extrapolated statistical artifact of the linear model rather than a physiological volume, as no patient had zero injected volume. The correlation remains valid within the observed data range.

Volume Changes Measured With MRI Scanning

The graph reveals that, in most cases, the volume of filler measured by 3D MRI is significantly higher than the volume initially injected. This discrepancy is particularly evident in patients who received larger initial injections, where the measured volume tends to be substantially greater than the injected amount. This observation suggests that the filler volume increases, most likely due to water complexing resulting from the HA hydrophilic properties. Pearson correlation analysis showed no significant relationship between injection-to-MRI interval and the degree of expansion (r = 0.13, P = 0.64), indicating that filler expansion plateaus early and remains stable.

Individual Variability

The graph also highlights variability among patients, particularly those with larger initial injections. For patients 12, 13, and 14, the measured filler volume far exceeds the amount injected, indicating that the effect of the filler becomes more pronounced with larger doses. In contrast, patients who received smaller injections (patients 1–5) show a smaller difference between the injected and measured volumes, suggesting that smaller volumes may result in more predictable outcomes.

Table 2 presents data on the thickness of dermal fillers used in various facial locations, expressed as the mean thickness in millimeters, along with the SD. It also shows the number of cases (n) and the percentage of the total cases for each location.

Table 2.

Filler Location and Presentation

Location	n	Percentage	Thickness, mm, Mean ± SD
Tear trough	6	42.86	7.17 ± 3.06
Zygoma	7	50.00	7.57 ± 2.07
Temporalis	1	7.14	3
Malar region	3	21.43	7 ± 5
Midcheek	5	35.71	8.2 ± 5.36
Nasolabial region	3	21.43	9 ± 3.46
Masseter	2	14.29	6 ± 7.07
Jowl	2	14.29	8.5 ± 0.71
Upper and lower lips	2	14.29	5.5 ± 0.71
Perioral	2	14.29	6
Mentum	1	7.14	6

Key observations include the zygoma and tear trough regions being the most commonly injected, with filler thickness generally between 7 and 8 mm. The nasolabial region exhibited the highest mean filler thickness at 9 mm, whereas the temple had the lowest at 3 mm; notably, only 1 case had temple filler in our series. Significant variability in filler thickness was noted in the malar region and midcheek areas, suggesting varying treatment strategies or patient-specific considerations. Consistent filler application was observed in the jowl and upper/lower lips, with minimal variability in thickness. The tear trough region showed the greatest proportional increase in thickness (mean 7.17 ± 3.06 mm), possibly due to the tight fibrovascular density and lower subcutaneous fat content in this area, resulting in visible tissue displacement.

Figure 4 illustrates the correlation between the time elapsed since the last filler injection and the degree of expansion. No correlation was observed between the time elapsed since the last filler injection and the degree of expansion. Notably, even in the patient whose last filler injection occurred 180 months before their scan, an expansion of 3.5 times was recorded.

Fig. 4.

Correlation between time since the last HA filler injection (mo) and the degree of volume expansion measured on 3D MRI. Each data point represents 1 patient (n = 14). No significant correlation was found (R² = 0.0177), suggesting that volumetric expansion occurs early after injection and remains stable thereafter.

The R² value of 0.0177 suggests a weak correlation between expansion and time, suggesting that the expansion process occurs early after the injection is administered. Although Figure 3 demonstrates a strong correlation between injected and measured filler volume, Figure 4 shows that this expansion is not time-dependent but occurs early after injection.

DISCUSSION

This is the first study to compare the initial volume of injected HA filler to the volume detected using 3D MRI. We observed 2 key phenomena related to HA filler behavior: (1) the expansion of the filler material beyond the injected volume and (2) the prolonged persistence of the filler, detectable even after 180 months. These findings challenge traditional understandings of filler behavior, particularly regarding its volume dynamics and longevity.

The significant expansion of the filler, on average 2.8 times its initially injected volume, was found to correlate with the amount initially injected. This may be attributed to the hydrophilic properties of HA, which draw in water and contribute to a larger final volume than initially expected. This raises important considerations for clinical practice, suggesting that overcorrection can occur, especially when larger amounts are injected initially. This may also explain the tendency for swelling, especially in superficial areas. The pronounced changes in the tear trough likely result from the unique anatomical characteristics of the region. The dense orbicularis retaining ligaments and limited subcutaneous space may concentrate volumetric swelling, leading to more prominent tissue displacement.

Remarkably, our MRI analysis shows detectable filler presence even after 180 months, a duration much longer than the commonly expected lifespan of HA fillers, which is often estimated at 6–12 months. This suggests that HA fillers, particularly those used in certain anatomical areas, may offer extended cosmetic benefits, thus altering the frequency of maintenance treatments. These findings are consistent with recent evidence suggesting that HA fillers may persist significantly longer, especially in low-motion or lower metabolic activity areas, underscoring the need for individualized treatment planning.

Our study supports the scientific literature that presents an expanding body of evidence regarding the long-term persistence of HA fillers. Several studies have recorded instances of fillers lasting for many years.^12–14

Clinical Implications

The expansion and prolonged longevity of HA fillers have critical implications for clinical practice. The expansion phenomenon suggests that clinicians should adopt a conservative approach when selecting injection volumes to avoid overcorrection and aesthetic complications, especially in areas prone to swelling. The extended longevity observed challenges current assumptions for filler maintenance, suggesting that patients may not require frequent follow-up treatments as often as previously thought.

Patients receiving larger volumes should be informed of the possibility of increased filler spread or other changes that could impact the final aesthetic result. Understanding these trends can help guide the amount of filler administered and inform follow-up care to achieve the desired outcome. The absence of correlation between time since injection and expansion suggests that hydration-related swelling occurs early—within days to weeks postinjection—and remains stable thereafter.

Although this study offers valuable insights into filler behavior, it is constrained by its sample size and retrospective design. Another limitation is that the research was conducted in a highly specialized clinic where patients commonly present with filler-related complications, most frequently swelling from previously injected material. This setting may introduce selection bias. Because many patients were referred for MRI due to suspected filler-related swelling or contour irregularities, this patient group may be particularly susceptible to filler expansion. A population-based characterization of the behavior of filler is therefore beyond the scope of this study. Additionally, a subanalysis comparing patients with and without clinical edema was not possible due to incomplete documentation of symptoms.

Future studies with larger cohorts and prospective designs are necessary to validate these findings and explore the biological mechanisms underpinning the observed filler expansion and prolonged longevity. Additionally, the role of different HA formulations, injection techniques, and anatomical sites warrants further investigation.

CONCLUSIONS

In summary, this study demonstrates that 3D MRI often detects a substantially greater filler volume than was initially injected, with a clear linear relationship between injection volume and measured expansion. The wide range of intervals between injection and imaging suggests that this expansion occurs early and that HA filler persists longer than commonly assumed. These findings highlight the importance of precise treatment planning and informed patient counseling. Further longitudinal studies are needed to characterize how filler volume evolves over time and to compare different HA filler types.

DISCLOSURE

The authors have no financial interest to declare in relation to the content of this article.

ACKNOWLEDGMENT

The authors grant permission for the use of the illustrations created for this submission, which have not been used elsewhere.