The use of basic fibroblast growth factor to improve vocal function: A systematic review and meta‐analysis

This systematic review and meta‐analysis examines if intralaryngeal injection of basic fibroblast growth factor 2 (FGF2) can improve voice outcomes in those with vocal disability.

there is no established treatment to restore viscoelastic tissue within the vocal fold. Experimental therapies using biomaterials, growth factors and/or mesenchymal stem cells have been tested in animal models with some reported success. [1][2][3] However, few of these have reached clinical trials and significant challenges remain in terms of scalability and manufacturing cost. A notable exception is the use of basic fibroblast growth factor (FGF2), which has been trialled in human subjects since 2008. 4 FGF2 is a growth factor that stimulates the migration and proliferation of vocal fold fibroblasts and has been shown to stimulate the secretion of proteins important for viscoelasticity and suppress the deposition of fibrous protein. 5 FGF2 is relatively inexpensive to manufacture and can be injected into the vocal fold using routine in-office procedures via a trans-oral or transcutaneous route. This provides a significant advantage over more complex biotherapies and greatly increases its potential to be adopted as a new globalised clinical therapy for chronic vocal disability.

| Objectives
The objectives of this article were to establish whether intralaryngeal injection of FGF2 improve voice outcomes in those with an existing vocal disability. The review will also examine improvements in mucosal wave and glottic closure following injection and examine the safety profile of this therapy.

| METHODS
A systematic review of literature was performed by one reviewer with specific reference to the use of basic FGF2 to restore vocal function following the PRISMA reporting guideline (Figure 1).

| Ethical considerations
This review did not require ethical committee approval.

| Eligibility criteria
All original human studies reporting voice outcomes following injection of FGF2 were included. The inclusion criteria were: Studies written in English, studies examining vocal disability caused by vocal fold atrophy, vocal fold scarring, vocal fold sulcus or vocal fold immobility. Objective outcome measurements were acoustic and aerodynamic analysis of voice and measurements of glottic gap. Subjective outcomes were the voice handicap index (VHI) questionnaire, the perceptual voice grading system, GRBAS and assessments of vocal fold mucosal wave during phonation. Exclusion criteria were: Articles not written in English, animal studies and articles lacking outcome measures before and after injection of FGF2.
The primary outcome measure for this study was maximum phonation time (MPT). Secondary outcome measures were acoustic analysis, patient-reported VHI scores, the perceptual voice grading system GRBAS and improvements in mucosal wave and glottic closure.

| Data selection and collection
One reviewer identified relevant articles. Reference lists were scanned for further relevant articles. Identified international experts in

Key points
• Mean maximum phonation time increased significantly after injection of FGF2 in those with vocal fold atrophy in the meta-analysis.
• Acoustic analysis outcomes were more variable but may represent difficulties with assessing global voice improvement with the methods used. • To date, injection of FGF2 appears to be safe with no major adverse events recorded in 390 cases examined. Full-text articles were reviewed for eligibility and those not relevant were excluded.

| Study risk of bias assessment
The risk of bias was determined using the ROBINS-I tool for determining risk of bias in non-randomised intervention studies. 6 A funnel plot and Egger's test, 7 were used to investigate the possibility of publication bias.

| Synthesis methods
The mean and standard deviation of MPT were recorded before injection and 3-6 months after injection. The values were recorded or calculated from data provided in the reports or extrapolated from the error bars on graphs. Information (i.e., standard deviation) was not available on the differences between values. Given the heterogeneity in the underlying cause of vocal disability, only MPT data derived from vocal fold atrophy were pooled using a random effects meta-analysis (with τ 2 estimated using REML). Forest plots were used to summarise the results and heterogeneity was quantified using the I 2 statistic.

| Study selection and characteristics
A total of 1023 articles were found matching the search criteria.
Records after duplicates removed were 807. Fifteen of which fulfilled the entry criteria. One article was excluded as it did not report on the assessment of voice outcomes following injection of FGF2. One article was found from scanning reference lists. Fifteen studies were taken forward for analysis ( Figure 1). 4,[8][9][10][11][12][13][14][15][16][17][18][19][20][21] The total number of study participants across all 15 reports was 390 (median: 17, IQR: 11-31). Conditions treated were vocal fold atrophy (n = 186), vocal cord paralysis (n = 74), vocal fold scar (n = 74) and vocal fold sulcus (n = 56) ( Table 1). Data collection was prospective (n = 13) and retrospective (n = 2) and involved a single treatment arm in all cases. Follow-up ranged from 3 months postinjection to 36 months. A transoral method of injection was used in six studies, a transcutaneous method of injection via the cricothyroid or thyrohyoid membrane was used in eight studies and in one study, injection was administered via the side port of a flexible nasolaryngoscope.

| Assessment of bias
Bias was assessed using the ROBINS-I tool (Appendix 2). Bias in selection of participants into the study and in classification of interventions

| Voice outcomes following FGF2 injection
A range of outcome measures were used (Table 2) The perceptual voice grading system, GRBAS, was used in 7 out of 15 cases in the short-term and 5 out of 7 cases in the longterm. Mean GRBAS was 5.7 pre-injection, 2.5 within 6 months of injection and 1.8 a year or more after injection. A significantly reduced mean GRBAS score was recorded for all studies. To control inter-rater reliability, 2-3 raters were used in four of the studies. 10,11,17,20 Of these four studies, two reported using independent assessors. 11,20 In three studies, no method to control inter-rater reliability was described. 8,12,15 Mucosal wave and glottic closure were assessed using either an independent four-point scale (Reports 5 and 8) or motion analysis software (Reports 3, 10 and 11). In all cases, a significant improvement in glottic closure and mucosal wave characteristics were reported. Acoustic analysis of voice showed more variable improvement in both short-term and long-term timepoints following injection (Table 2).

| Sub-group analysis
Out of the six studies that investigated different vocal fold pathologies, two provided a sub-group analysis. Hirano et al. 11 grouped sulcus and scar patients and graded them as mild, moderate or severe. VHI-10 scores improved more in the mild and moderate cases following FGF2 injection compared to severe cases. Improvements in VHI-10 were also significantly greater amongst patients with vocal fold T A B L E 1 Fifteen studies were included for analysis. Note: Studies examined patients with vocal cord palsy (P), vocal fold atrophy (A), vocal fold scarring (Sc) or vocal fold sulcus (Su). Injections were delivered via a trans-oral route using a curved injection needle, a trans-nasal route using a flexible endoscope with side channel or percutaneously (perc) via the thyrohyoid or cricothyroid membrane.
T A B L E 2 The outcome measures used before and after injection of FGF2. atrophy compared to patients with scar or sulcus. Kanazawa et al. 19 ranked improvement in MPT and found the greatest improvement in vocal fold atrophy followed by paralysis and then scar and sulcus. The impact of age was also explored and improvements in MPT following injection were found to have an inverse relationship to age. However, Okui et al. 13 found no effect of age on MPT or VHI when comparing those over 70 years of age and those under 70 years with vocal fold atrophy.

| Injection method and adverse events
All studies used recombinant human FGF2 (Fiblast, Kaken Pharmaceutical, Tokyo, Japan). Four reports used a dose of 10 mcg given four times at 1-week intervals and repeated at 3 months if no improvement was found (Table 1). Three reports used a single dose of 10 mcg.
Four reports used a single dose of 50 mcg, which in one report, was repeated up to three times if no improvement was recorded after 4 months. All other studies used a single injection ranging between .25, 2-4 and 20-30 mcg. Given the variability in study methodology, data synthesis was not possible to calculate dose effect. As shown in Table 1, similar outcomes were observed irrespective of dose and timing. Ten studies commented on the incidence of adverse effects following injection. 8,11,12,[14][15][16][18][19][20][21] Out of these 10, no major adverse events were reported. Transient hyperemia of the vocal fold with associated transient vocal hoarseness was consistently reported. In Hirano et al., 11 where 100 patients were followed up, hyperemia occurred in 72% and was graded as mild in 66% and severe in 6%. In all cases, the hyperemia had resolved by 2 months post-injection.

| DISCUSSION
Out of 390 cases examined, no major adverse events were reported.
This indicates intralaryngeal injection of FGF2, to date, is safe although evaluation of further cases is needed to identify rarer events that may arise.

| CONCLUSION
To date, the injection of intralaryngeal FGF2 to treat vocal dysfunction appears safe and there is evidence supporting efficacy in improving voice outcomes. Future studies should aim to include a control group and should deploy multivariate acoustic measurements to achieve a more global assessment of voice quality.

FUNDING INFORMATION
No external funding or support was provided for this review.