Centrifuged-Concentrated Intravitreal Slurry Triamcinolone Acetonide: An Inexpensive, Easy, and Viable Alternative to Long-Term Steroid Delivery

Abstract

Purpose:

This work reports the duration, safety, and viability of intravitreal slurry triamcinolone acetonide (TA; 1.0 mL of 40-mg/mL TA centrifuge concentrated into a 0.1-mL pellet) to treat cystoid macular edema (CME).

Methods:

A retrospective, consecutive review was conducted of patients undergoing intravitreal slurry TA injections, July 2009 to December 2018.

Results:

In 143 eyes of 120 patients, slurry TA resolved CME for a mean of 327.15 (SD = 213.11) days, or 10.76 (SD = 7.00) months, per intravitreal injection (n = 466). In 100 eyes requiring multiple injections (n = 423), mean duration was 270.95 (SD = 177.14) days, or 8.91 (SD = 5.82) months, between injections. In 43 single-injection eyes, duration was 749.30 (SD = 483.17) days, or 24.63 (SD = 15.88) months. Mean duration decreased from 337.89 (SD = 210.46) days, or 11.11 (SD = 6.92) months, in nonvitrectomized eyes to 279.74 (SD 179.63) days, or 9.20 (SD = 5.91) months, in vitrectomized eyes (n = 74 injections, t = 2.24, P = .014, 1-tailed). Central foveal thickness as shown on optical coherence tomography decreased by 173.89μ (SD = 147.56μ), from 459.16μ (SD = 47.14μ) to 285.27μ (SD = 77.27μ; t = –25.31, P < .001), within 43.41 days (SD = 36.86). Visual acuity improved from 20/100 (logMAR 0.70, SD = 0.33) to 20/74 (logMAR 0.57, SD = 0.31; SD = 0.21; t = –11.01, P < .001), within 33.98 (SD 24.98) days. Fifteen of 31 phakic eyes (48.39%) underwent cataract extraction. Fifty-seven eyes (39.86%) developed a steroid response (> 10 mm Hg increase from baseline) 94.79 days (SD = 85.52 days), or 3.11 (SD = 2.81) months, following injection.

Conclusions:

A single injection of slurry TA lasted on average 10.76 months with significant improvement of CME and visual acuity. Adverse ocular effects were comparable to currently available, long-term, implantable steroids. Slurry TA appears to be an easily reproducible, safe, and cost-effective alternative to long-term intraocular steroid delivery.

Introduction

Long-term, intravitreal steroid delivery plays a vital role in the treatment of cystoid macular edema (CME) from a variety of conditions, including diabetes, vein occlusion, uveitis, and radiation retinopathy. Corticosteroids have anti-inflammatory, antiangiogenic, and antipermeability properties that make them an important therapeutic option for a variety of posterior-segment diseases. Based on experimental studies, corticosteroids have been shown to control gene expression of inflammatory mediators. This regulation influences the expression of vascular endothelial growth factor (VEGF), inhibits proinflammatory genes such as tumor necrosis factor α and other inflammatory chemokines, and induces the expression of anti-inflammatory factors such as pigment-derived growth factor. ^{1 -4} Machemer, among others, first suggested the intravitreal delivery of steroids to locally suppress intraocular inflammation, proliferation of cells, and neovascularization. ^5,6

Long-term steroid preparations include fluocinolone acetonide 0.59 mg (Retisert; Bausch + Lomb), dexamethasone (DEX) 0.7 mg (Ozurdex; Allergan, Inc), fluocinolone acetonide 0.19 mg (Iluvien; Alimera Sciences), and fluocinolone acetonide 0.18 mg (Yutiq; EyePoint Pharmaceuticals). In the United States, triamcinolone acetonide (TA) is available as a Food and Drug Administration (FDA)–approved, preservative-free (PF) TA intravitreal suspension (Triesence; Alcon) and TA injectable suspension (Kenalog; Bristol-Myers Squibb). According to DrugPatentWatch.Com, PF-TA is approved in 21 countries and DEX 0.7 mg implant is approved in 29 countries worldwide. ^7,8

Although TA carries a specific warning in the package insert against its use in or around the eye, because of the limited availability of other steroid preparations, it remains the most widely used intraocular steroid worldwide ⁹ and has been used for intravitreal injections since 2004. This formulation is FDA approved for only intramuscular and intra-articular use and is currently used off-label for intraocular injections. In Europe, Vitreal S (Sooft s.p.a.) is a medical device used intraoperatively to stain the vitreous during vitrectomy, and it is not registered as a drug for intraocular use. In the European Union, TA is contraindicated for intraocular injection, and PF-TA is not approved for any indication.

TA represents a commonly used steroid agent for the worldwide treatment of several retinal conditions. ¹⁰ This steroid has an anti-inflammatory potency 5 times higher than hydrocortisone. It appears as a white- to cream-colored crystalline powder and is practically insoluble in water and very soluble in alcohol. ^11,12 Because the formulation is insoluble, the steroid remains as a depot with a small, constant diffusion of the active medication. Injections of 4 mg of intravitreal TA can provide therapeutic effects for approximately 3 months. ^{13 -15} Because pure TA is insoluble, the larger the dose, the longer it is expected to be effective. To increase the duration of its effect, various methods to concentrate the steroid have been used. Sedimentation and Millipore filtration produce variable results. In 2013, Ober and Valijan ¹⁶ reported on the benefit of centrifuge-concentrated intravitreal triamcinolone (C-IVT) using PF-TA as another alternative for long-term, intraocular steroid delivery. The duration of effect was estimated based on the presence of visible steroid within the vitreous cavity in nonvitrectomized eyes.

The present study reports the duration of effect of centrifuge-concentrated intravitreal slurry TA (ie, slurry Kenalog) based on the visible recurrence of macular edema by optical coherence tomography (OCT). The effect on visual acuity (VA) is also evaluated. An attempt is also made to further differentiate the duration of effect based on the clinical condition. Furthermore, this study reports the rate of cataract development and steroid response following injection. The status of the vitreous is also analyzed for any effect on steroid duration, including degree of posterior vitreous detachment (PVD) and previous vitrectomy.

Methods

This study was approved by the Western Institutional Review Board (Puyallup, Washington), tracking number 20172984. All records of consecutive patients undergoing injection of centrifuge-concentrated slurry TA from July 2009 through December 2018 were reviewed. Exclusion criteria included less than 9 months of follow-up from the time of last injection and uninterpretable or corrupted OCT data. Study patients did not receive any concurrent nor subsequent anti-VEGF therapy, other intravitreal steroid injections, nor laser treatments.

All intravitreal injections consisted of 0.1 mL of centrifuged slurry TA (40 mg/1.0 mL) concentrated to approximately 25 mg/0.1 mL (Figure 1). ¹⁷ All patients were treated by 1 surgeon (S.M.M.). Data were collected regarding date of slurry TA injection(s), ocular diagnosis, ocular history (including previous vitrectomy surgery), previous steroid and/or anti-VEGF treatments, baseline VA, lens status, degree of vitreous detachment as determined by clinical examination and OCT assessment, baseline and follow-up intraocular pressure (IOP), best follow-up VA and OCT central foveal thickness (CFT), and time to maximal improvement in VA and CFT. OCT data included baseline CFT, CFT at the time of most improved VA, thinnest CFT post injection, and the time to thinnest CFT. Reinjection was based on recurrence of macular edema (at least 50 μ) as identified by OCT. Indication for re-treatment was typically based on clearly visible, significant recurrence of macular edema to pretreatment (or near pretreatment) levels by OCT, usually accompanied by visual decline (Figure 2).

Figure 1.

Slurry triamcinolone acetonide injectable suspension visible in the inferior fundus 3 weeks following intravitreal injection.

Figure 2.

Case example of a 71-year-old African American woman with a 22-year history of insulin-dependent diabetes. Ocular history included uncomplicated pseudophakia 5 years ago, with preexisting history of primary open-angle glaucoma, well controlled with twice daily brimonidine tartrate 0.2%. Successful panretinal photocoagulation for proliferative diabetic retinopathy was performed 3 years previously. In the antecedent 19 months, 5 ranubizumab and 3 aflibercept injections, 1 dexamethasone 0.7 mg implant, and 2 focal laser treatments were given for diabetic macular edema. Edema persisted despite these interventions. Slurry triamcinolone acetonide (TA) therapy was initiated. (A) Visual acuity (VA) was 20/50, intraocular pressure (IOP) 18, and central foveal thickness (CFT) 513 µ. Intravitreal slurry TA injection 1 was given. (B) One month following injection 1, VA improved to 20/30 −2, IOP 20, and CFT 261 µ. (C) The patient was seen regularly every 3 to 4 months with no recurrent edema and stable VA and IOP. At 16.6 months following injection 1, the patient reported decline in vision: VA was 20/40 −2, IOP 16, with recurrent macular edema and CFT 572 µ. Slurry TA 2 given. (D) Three weeks following repeat injection 2, VA was 20/40 +2, IOP 16, and CFT 209 µ. (E) The patient was again seen regularly every 3 to 4 months with no recurrent edema and stable VA and IOP. At 9.3 months following slurry TA injection 2, VA declined to 20/100, IOP 17, and CFT 600 µ. Slurry TA 3 was given. (F) Three weeks following slurry TA injection 3, VA improved to 20/50 + 2, IOP 14, and CFT 201 µ.

In some cases, especially in patients with more severe baseline VA (eg, < 20/200), the recurrent edema was not associated with any subjective visual decline because of advanced retinal disease. In these cases, the re-treatment was based solely on the recurrence of thickening by OCT, typically to pretreatment (or near pretreatment) levels. During the study time frame, CFT measurements were obtained on both time domain and spectral domain OCT (Carl Zeiss Meditec). Any complications deemed to be the result of the injection, such as sterile or infectious endophthalmitis as well as retinal detachment, were recorded. Because only 1 physician performed all the injections, a fairly regular follow-up evaluation was available for most patients. The typical patient reevaluation post injection was an initial evaluation at 3 to 4 weeks post injection, then every 2 to 3 months, or sooner if the patient contacted the office with complaints of visual decline. Owing to the retrospective nature of the study, the presence of visible, far peripheral, intravitreal steroid was not deemed reliable as an indicator of efficacy or duration because patients typically did not undergo scleral depression on subsequent visits.

The centrifuge technique for concentrating TA has been elegantly described in detail by Ober and Valjian. ¹⁶ The entire contents (1 mL) of 1 vial of TA 40 mg/mL is withdrawn into a 1.0-mL tuberculin syringe. The plunger and flanges are excised. The syringe is then placed into a centrifuge with the sterile cap facing upward and spun for 30 seconds. To create a more uniform medication level, the syringe is then rotated 180°, placing the higher medication level toward the inside, and spun for an additional 30 seconds. A new plunger is inserted behind the severed plunger and the plunger is advanced to 0.1 mL for all injections. The intention is to remove as much of the clear supernatant as possible, while leaving the maximum amount of opaque TA in the syringe. Using a subconjunctival anesthetic of 2% lidocaine, the entire 0.1 mL is injected inferotemporally intravitreally via a firmly attached 30-gauge needle (video available on Eyetube at http://eyetube.net/video/uumnxen/ and YouTube at https://www.youtube.com/watch? v=37QI4aHDyu8).

Analysis of the data consisted of basic descriptive analyses as well as parametric difference of means (t tests and analysis of variance). Multivariate analyses were also used to examine effects of injection on change in CFT, VA, and duration of effect.

Results

A total of 544 slurry TA injections were performed from July 2009 through December 2018. Of the total injections, 466 were eligible for analysis as defined by at least 9 months of follow-up from the most recent injection and OCT data that were interpretable and available for analysis. The 466 injections were given in 143 eyes from 120 patients. Of these 143 eyes, 100 eyes in 89 patients received repeated injections for a total of 423 injections. Forty-three eyes received only 1 injection without a second treatment.

Indications for treatment included CME due to diabetes (63.6%), central retinal vein occlusion (CRVO; total 21.7%, ischemic 12.6%, nonischemic 9.1%), uveitis (11.2%), branch retinal vein occlusion (BRVO; 2.1%), and radiation retinopathy (1.4%) (Table 1).

Table 1.

Number of All Injections and Eyes Stratified by Diagnosis.

Diagnosis	Injections (%)	Eyes (%)
BRVO	13 (2.8)	3 (2.1)
DME	276 (59.2)	91 (63.6)
Ischemic CRVO	63 (13.5)	18 (12.6)
Nonischemic CRVO	40 (8.6)	13 (9.1)
Uveitis	57 (12.2)	16 (11.2)
Radiation retinopathy	17 (3.6)	2 (1.4)
Total	466	143

Abbreviations: BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; DME, diabetic macular edema.

The average total time of study follow-up was 1226.26 (SD = 863.38) days, or 40.29 (SD = 28.39) months, per eye. No patient was treatment naive, having received a myriad combination of previous treatments with anti-VEGF agents, long-term steroid implants, laser treatments, and/or vitrectomy surgery. All eyes first received an initial 0.1-mL challenge dose of unconcentrated 4 mg/0.1 mL of TA. The average duration of effect, as defined by recurrence of CME on OCT, was 97.61 (SD = 36.80) days, or 3.21 (SD = 1.21) months, after this uncentrifuged, commercially available dose of TA.

Duration

For all eyes (including 1-injection and multiple-injection eyes), slurry TA lasted an average of 327.15 (SD = 213.11) days, or 10.76 (SD = 7.00) months, following a single intravitreal injection, as defined by recurrence of CME or continued absence of CME after more than 9 months of follow-up. One hundred eyes received multiple injections for a total of 423 injections. In the multiple-injection group, only the duration between injections was used for analysis (even if there was more than 9 months of follow-up from the last injection) because of the presumption that eventually the macular edema would recur after the last documented injection had the patient been followed longer than 9 months. In this multiple-injection group, the mean duration was 270.95 (SD = 177.14) days, or 8.91 (SD = 5.82) months, between injections, with no appreciable degradation in effect across subsequent injections. Forty-three eyes received only 1 injection for a mean duration of 749.30 (SD = 483.17) days, or 24.63 (SD = 15.88) months. Of the 43 eyes receiving only 1 injection, 39 patients never recurred, 2 patients developed sterile inflammation, 1 eye developed a severe steroid response, and 1 patient declined additional treatment because of continued poor vision despite documented OCT improvement.

In 25 eyes, prior vitrectomy (74 injections) significantly decreased the mean duration of treatment effect from 337.89 (SD = 210.46) days, or 11.11 (SD = 6.92) months, without vitrectomy to 279.74 (SD = 179.63) days, or 9.20 (SD = 5.91) months, with vitrectomy (t = 2.24, P =.014, 1-tailed). Five of these 25 eyes underwent vitrectomy during the course of the study (Figure 3).

Figure 3.

Vitrectomy (n = 74) decreased the overall average duration of effect of a single injection of centrifuge-concentrated slurry triamcinolone acetonide from 11.11 (n = 392) to 9.20 months (P = .014).

Duration by Diagnosis

The duration of effect for the multiple-injection group was evaluated by specific disease indication (Table 2 and Figure 4). The longest average duration of effect was in diabetic macular edema (DME) at an average duration of 12.15 (SD = 7.54) months and the shortest was in ischemic CRVO at 7.79 (SD = 3.32) months.

Table 2.

Average Duration Following Each Individual Injection of Slurry Triamcinolone Acetonide, Stratified by Diagnosis.^a

Diagnosis	Average duration, d/mo	SD, d/mo	No. of injections
BRVO	289.39/9.51	84.93/2.79	10
DME	369.67/12.15	229.47/7.54	215
Ischemic CRVO	236.80/7.79	101.36/3.32	49
Nonischemic CRVO	297.36/9.78	169.89/5.58	33
Noninfectious uveitis	271.75/8.93	194.97/6.41	44
Radiation retinopathy	266.13/8.75	20.18/0.66	15
Total	327.15/10.76	213.11/7.00	366

Abbreviations: BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; DME, diabetic macular edema.

^aMultiple injection eyes only.

Figure 4.

Average duration of effect from a single injection of centrifuge-concentrated slurry triamcinolone acetonide (multiple-injection eyes only), stratified by diagnosis.

In a pooled analysis of the multiple-injection group, there were significant aggregate differences in duration of effect between diagnoses (F = 5.01; P < .001), with the difference being driven mainly by the increased duration in patients with DME relative to those with ischemic CRVO (–132.87 days; P = .0005) and uveitis (–97.92 days; P = .039). However, the post hoc pairwise results should be considered with caution because of small sample size within certain diagnoses. Although the injections appeared to last longer in DME than in these other diagnoses, there was substantially greater variation in duration among this group (Table 3).

Table 3.

Tukey Honestly Significant Difference Comparison of Duration (Days) from Single Injection Compared by Diagnosis.^a

Comparison by diagnosis	Difference, d	95% CI	P
BRVOb vs DMEc	80.27	–105.43 to 265.97	.82
BRVO vs ischemic CRVOd	52.60	–251.79 to 146.59	.97
BRVO vs nonischemic CRVOe	17.26	–190.71 to 225.229	> .99
BRVO vs uveitisf	–17.65	–218.75 to 183.459	> .99
BRVO vs radiation retinopathyg	–23.27	–257.62 to 211.089	> .99
DME vs ischemic CRVO	–132.877	–223.74 to –42.009	< .001
DME vs nonischemic CRVO	–63.017	–171.78 to 45.759	.56
DME vs uveitis	–97.92	–192.90 to –2.94	.04
DME vs radiation retinopathy	–103.54	–256.83 to 49.76	.38
Ischemic CRVO vs nonischemic CRVO	69.86	–60.61 to 200.33	.64
Ischemic CRVO vs uveitis	34.956	–84.27 to 154.18	.96
Ischemic CRVO vs radiation retinopathy	29.34	–140.05 to 198.72	> .99
Nonischemic CRVO vs uveitis	–34.90	–168.27 to 98.46	.98
Nonischemic CRVO vs radiation retinopathy	–40.52	–220.15 to 139.10	.99
Uveitis vs radiation retinopathy	–5.62	–177.25 to 166.01	>.99

Abbreviations: BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; DME, diabetic macular edema.

^a Duration difference was greatest for DME vs uveitis (–97.92 days, P = .039) and DME vs ischemic CRVO (–132.87 days, P < .001).

^b BRVO, n = 10.

^c DME, n = 185.

^d Ischemic CRVO, n = 45.

^e Nonischemic CRVO, n = 27.

^f Uveitis, n = 41.

^g Radiation retinopathy, n = 17.

Visual Acuity

For all injections, the mean VA improved from 20/100 (logMAR 0.70, SD = 0.33) to 20/74 (logMAR 0.57, SD = 0.31), a mean change of –0.11 (SD = 0.21; t = –11.01, P < .001), within an average of 33.98 (SD = 24.98) days.

In a pooled analysis across all injections, change in VA was significantly different between diagnoses (F = 3.37; P = .005), with the difference being driven mainly by the greater improvement in patients with BRVO relative to those with DME (0.18; P = .018) and ischemic CRVO (0.188; P = .025) (Table 4). Matched-pair difference of means tests within each diagnostic group for all injections, however, suggest significant VA improvements across all groups (Table 5).

Table 4.

Tukey Honestly Significant Difference Comparison of Visual Acuity Improvement (LogMAR) from a Single Injection Stratified by Diagnosis.^a

Comparison by diagnosis	Difference, logMAR	95% CI	P
BRVOb vs DMEc	0.18	0.02 to 0.34	.01
BRVO vs ischemic CRVOd	0.188	0.01 to 0.35	.02
BRVO vs nonischemic CRVOe	0.168	–0.02 to 0.34	.10
BRVO vs uveitisf	0.11	–0.06 to 0.28	.47
BRVO vs radiation retinopathyg	0.13	–0.07 to 0.34	.41
DME vs ischemic CRVO	0.004	–0.08 to 0.08	> .99
DME vs nonischemic CRVO	–0.02	–0.12 to 0.08	> .99
DME vs uveitis	–0.07	–0.16 to 0.01	.12
DME vs radiation retinopathy	–0.04	–0.18 to 0.09	.93
Ischemic CRVO vs nonischemic CRVO	–0.02	–0.14 to 0.10	> .99
Ischemic CRVO vs uveitis	–0.08	–0.18 to 0.03	.28
Ischemic CRVO vs radiation retinopathy	–0.05	–0.20 to 0.10	.94
Nonischemic CRVO vs uveitis	–0.06	–0.18 to 0.06	.75
Nonischemic CRVO vs radiation retinopathy	–0.03	–0.19 to 0.13	> .99
Uveitis vs radiation retinopathy	0.03	–0.12 to 0.18	> .99

Abbreviations: BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; DME, diabetic macular edema.

^aImprovement difference was greatest for BRVO relative to those with DME (0.18; P = .01), and ischemic CRVO (0.18; P = .02).

^b BRVO, n = 13.

^c DME, n = 246.

^d Ischemic CRVO, n = 58.

^e Nonischemic CRVO, n = 34.

^f Uveitis, n = 54.

^g Radiation retinopathy, n = 17.

Table 5.

Average Logarithm of Minimum Angle of Resolution Visual Acuity Preinjection and Postinjection, Stratified by Diagnosis.^a

Diagnosis	No. of injections	Initial logMAR (Snellen VA)	Initial mean SD	Ending/best logMAR (Snellen VA)	Ending/best mean SD	Diffb	t diffb	P b
BRVO	13	0.73 (20/108)	0.19	0.46 (20/58)	0.09	–0.27	–2.69	< .001
DME	276	0.57 (20/74)	0.33	0.42 (20/53)	0.28	–0.10	–8.26	< .001
Ischemic CRVO	63	1.08 (20/238)	0.28	0.99 (20/195)	0.34	–0.08	–3.55	< .001
Nonischemic CRVO	40	0.80 (20/126)	0.32	0.68 (20/97)	0.35	–0.10	–3.06	.002
Uveitis	57	0.77 (20/117)	0.43	0.61 (20/81)	0.47	–0.16	–3.96	< .001
Radiation retinopathy	17	1.15 (20/281)	0.20	1.01 (20/206)	0.18	–0.14	–3.85	< .001
Total	466	0.70 (20/101)	0.33	0.57 (20/74)	0.31	–0.11	–11.01	< .001

Abbreviations: BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; Diff, difference; DME, diabetic macular edema; VA, visual acuity.

^a Differences reported are matched-pair differences by cases, not group differences.

At the time of most improved visual acuity (VA), mean central foveal thickness (CFT) decreased by an average of 158.04μ (SD = 148μ), from 459.16μ (SD = 47.14μ) to 301.12μ (SD = 93μ; P < .001), within an average of 33.98 (SD = 24.98) days. The CFT decreased further by a maximum average of 173.89μ (SD = 147.56μ), from 459.16μ (SD = 47.14μ) to 285.27μ (SD = 77.27μ; t = –25.31, P < .001), within an average of 43.41 (SD = 36.86) days. Mean time to best VA (33.98 days, SD = 24.98) was significantly shorter than mean time to best CFT (43.41 days, SD = 36.86; t = 4.56; P < .001). There was no appreciable degradation of effect over time (see Figure 5).

Figure 5.

Central foveal thickness (CFT) change over time. Note the rapid, and repetitive, effect on improving the CFT after each injection. CFT continues to decrease further even after the best final VA improvement is achieved. Baseline # indicates the sequence number of the injection; Best VA # indicates the CFT at the time of most improved visual acuity following each injection; Max # indicates the maximum improvement in CFT following each injection.

Predictors of CFT Change, VA Improvement, and Duration

A multivariate logistic regression for best CFT was performed. The model controlled for vitrectomy, epiretinal membrane, subretinal fluid, cystic change by OCT, starting vision, and degree of PVD as evidenced by OCT and initial CFT. Only starting CFT was significantly associated with any prediction of a decrease in CFT for injection 1 (β = –.87; P < .001; adjusted R ² = .83). Importantly, whereas the total model accounted for more than 80% of the variance in CFT change, starting CFT accounted for almost 90% of that explained variance.

When the same model was regressed on VA change, only starting VA (β = –.29; P = .004) and starting OCT thickness (β = –.35; P < .001) were significant, although the aggregate model explained only approximately 27% of the total variance, and these 2 variables combined accounted for only about half of that explained variance.

Finally, when the same model was regressed on duration of effect of the first injection, none of the independent variables manifested a significant association with duration of effect.

Effect of PVD on Duration

Analysis of variance was performed to examine whether a more liquified vitreous might also result in a shorter duration of effect of TA, as seen after vitrectomy surgery. The degree of posterior hyaloid detachment on OCT was used as a proxy for the extent of vitreous liquefaction. The level of PVD was graded from 0 to 4 based on OCT evaluation on a 9-mm horizontal scan through the fovea and disc and defined as follows:

• PVD 0: No evidence of detachment of the posterior hyaloid from the retinal surface

• PVD 1: Any elevation of the posterior hyaloid off the retinal surface, but the posterior hyaloid is not completely detached between the fovea and disc and/or fovea and temporal retina

• PVD 2: Persistent attachment of the posterior hyaloid at the fovea and disc with complete parafoveal detachment

• PVD 3: Persistent attachment of the posterior hyaloid at the disc only

• PVD 4: Complete detachment of the posterior hyaloid over the disc and macula

Eyes with an incomplete PVD (0, 1, 2, and 3) showed shorter duration (mean = 285.3 days; SD = 171.25) compared with those with a complete PVD (4) (mean = 332.58 days; SD = 309.39). The perceived difference in duration was not statistically significant (F = .681; P = .412), although the variances relative to the means were high for both subgroups.

Cataract Development

At the time of initial injection, 113 eyes were pseudophakic and 31 were phakic. Twenty-six phakic eyes (84%) had some progression of their cataracts. Fifteen of 31 phakic eyes (48.39%) underwent cataract extraction.

Steroid Response

Fifty-seven eyes (39.86%) developed a steroid response, defined as an increase of 10 mm Hg or more from baseline, at an average of 94.79 days (SD = 85.52), or 3.11 (SD = 2.81) months, following injection. The average IOP at the time of the initial steroid response was 30.65 mm Hg (SD = 5.82 mm Hg) with an average maximum peak IOP of 33.96 mm Hg (SD = 6.45 mm Hg).

Of the eyes developing a steroid response, 48 (84.21%) were controlled on 2 or fewer drops, 6 eyes (10.50%) required a trabeculectomy, and 3 eyes (5.26%) were monitored without any treatment.

A multivariate logistic regression including underlying diagnosis, lens status, history of glaucoma and/or glaucoma surgery, and number of injections demonstrated that none of these factors had any significant association with likelihood of steroid response even when controlling for each other factor. This may be because of the relatively small sample size.

Noninfectious and Infectious Endophthalmitis

No cases of infectious endophthalmitis occurred, with 2 cases of self-resolving, limited, sterile inflammation. This sterile inflammation resolved without treatment or sequelae within 2 weeks of injection and still produced a positive treatment effect. Sterile inflammation was associated with an almost immediate (within 3 days) painless significant drop in vision in the 20/400 to counting fingers level, vitritis with opacification, and hypopyon. These eyes demonstrated no chemosis or injection and there was absolutely no associated pain. These cases were observed with patient reassurance, without any additional treatment, and resolved within 10 to 14 days. The desired treatment effect of macular edema resolution still occurred in these instances, despite the development of sterile inflammation. Interestingly, these patients also developed sterile inflammation with subsequent injection of PF-TA. One patient with uveitis developed a rhegmatogenous retinal detachment 7 weeks following a second injection, which may or may not have been related to the injection and possibly to the uveitic process and other factors (Table 6).

Table 6.

Summary of Key Results.

Results	Injections (N = 466)
Eyes/patients	143/120
Mean duration overall of slurry TA injections	10.76 mo
Mean duration with vitrectomy (n = 74 injections)	9.20 mo
Mean duration without vitrectomy (n = 390 injections)a	11.11 mo
Steroid response (≥ Δ10 mm Hg from baseline)	57/143 eyes (39.86%)
Cataract progression	26/31 eyes (83.87%)
Cataract extraction	15/31 eyes (48.39%)
Sterile inflammation	2/466 injections (0.43%)
Retinal detachment	1/466 injections (0.21%)
Infectious endophthalmitis	0/466 injections (0.00%)

Abbreviation: TA, triamcinolone acetonide.

^a Two data points unavailable for analysis.

Conclusions

Long-term implantable steroids play a vital role in the treatment of chronic and recurrent macular edema, including DME, ¹⁸ vein occlusion, ^19,20 and uveitis. ²¹ Although long-term steroid preparations are available, they demonstrate the typical adverse effects of all intraocular steroids, including cataract development and IOP elevation. The flucinolone acetonide 0.59 mg implant requires a procedure in the surgical operating room. Both the DEX 0.7 mg and fluocinolone acetonide 0.19 mg implant can occasionally migrate into the anterior chamber, causing vision-threatening complications that include permanent corneal decompensation. ^{22 -26} The perils and difficulty of removing these implants from the anterior chamber have been well described. ^27,28

Furthermore, these long-term steroid delivery devices are each FDA approved for only a narrow range of indications, limiting their insurance reimbursement for other diagnoses. PF-TA is FDA approved for only intravitreal injection in the treatment of sympathetic ophthalmia, temporal arteritis, uveitis, and ocular inflammatory conditions unresponsive to topical steroids. PF-TA is not specifically approved for macular edema due to diabetes, CRVO, BRVO, or postradiation retinopathy. Fluocinolone acetonide 0.19 mg is approved for only DME. DEX 0.7 mg is approved for retinal vein occlusion, uveitis, and DME. Fluocinolone acetonide 0.18 mg and 0.59 mg implants are approved for chronic, noninfectious uveitis. Both fluocinolone acetonide 0.19 mg and DEX 0.7 mg implants require previous steroid challenge before initiating therapy. TA frequently receives reimbursement for a broad range of disorders, making it useful in a variety of clinical settings.

Methods to Concentrate Steroid

Methods to concentrate intraocular steroid have previously yielded variable results. Jonas et al described a method to concentrate TA and remove preservative at the same time. The procedure required a rather involved process to inject TA into a Millipore filter attached to a 3-way stopcock and “rinse” the resulting crystals with a balanced salt solution 1 or more times to remove the preservative. ²⁹ They estimated that this technique yielded 20 mg to 25 mg per 0.1 mL. Attempts to measure the actual dose of TA by duplicating this method yielded highly variable results with a range of 8.6 mg to 14.8 mg per 0.2 mL. ¹³ The variability may be multifactorial; however, the Millipore filter likely plays a large role by retaining a significant (and variable) amount of crystals during the rinsing process. Jonas and Rensch recorded visible triamcinolone in the vitreous up to 9 months after injection using their concentration and preservative removal procedure for patients with retinal vein occlusion. ³⁰

Chin and colleagues ³¹ used sedimentation to concentrate TA with a tuberculin syringe. After 30 minutes, the concentration average was 19.8 mg/0.1 mL and after 40 minutes, the average concentration was 20.12 mg/0.1 mL. Song et al ³² also described a technique to concentrate TA using sedimentation to a maximum of 25.7 mg/0.1 mL after 2 hours. Both of these studies independently concluded that shorter sedimentation times yielded lower concentrations. Passive sedimentation uses the force of gravity to concentrate TA.

Hernaez-Ortega and Soto-Pedre ^33,34 first described the centrifugation of triamcinolone, which they initiated for the purpose of reducing the preservative in the vehicle. García-Arumí et al later reported an identical technique. ³⁵ In their procedure, they placed an unaltered vial of triamcinolone in a centrifuge for 5 minutes, used a syringe to withdraw the supernatant, and then resuspended the remaining crystals with an equal volume of balanced salt solution from a separate syringe. This significantly reduced the benzyl alcohol preservative by 90%, leaving a concentration of triamcinolone roughly equal (42.70 mg/mL) to that in the original vial. None of these articles or those referencing the technique used centrifugation to concentrate the medication for injection of a larger dose in a smaller volume.

The centrifuge technique potentiates the effect of sedimentation by more than 1000 times the force of gravity. The entire centrifuge technique takes less than 5 minutes, which is shorter than the sedimentation method, which can take up to 2 hours for a similar, final dose. The centrifuge technique allows for preparation “on the fly” while seeing patients in the office or clinic, minimizing medication waste and not slowing patient flow. This technique may also theoretically decrease the rate of contamination because the steroid is used immediately and not forced to sediment in a syringe for many minutes or hours after breaching the sterile vial.

In 2013, Ober and Valjian ¹⁵ reported success by using the centrifuge technique to concentrate PF-TA. They reported 69 injections of 0.5 mL of C-IVT and 15 injections of 1.0 mL C-IVT over a 1-year time frame. The duration of effect was estimated based on the presence of visible steroid in the vitreous cavity. Visible steroid was seen on average 5.0 ± 2.4 months (median 5 months) after 0.05 mL of C-IVT and 8.3 ± 4.0 months (median 8 months) after 0.1 mL of C-IVT in nonvitrectomized eyes. Efficacy based on VA and CFT, as well as duration in vitrectomized eyes, was not reported.

Slurry TA Effect on Vision and CFT

As with other steroid preparations, slurry TA had a quick and lasting effect both on vision and OCT thickness. Despite the advanced and recalcitrant nature of disease in all of these eyes, which had failed other therapies or required frequent serial treatments, slurry TA improved the VA from 20/100 (logMAR 0.70, SD = 0.33) to 20/74 (logMAR 0.57, SD = 0.31), a mean change of –0.11 (SD = 0.21; t = –11.01, P < .001), within an average of 33.98 (SD = 24.98) days.

Slurry TA also had a profound and rapid effect on CFT. The best VA was achieved prior to reaching the greatest reduction in CFT, that is, mean CFT continued to improve even after best VA was first achieved. At the time that the best VA was achieved (33.98 days, SD = 24.98), mean CFT decreased by an average of 158.04μ (SD = 148μ), from 459.16μ (SD = 47.14μ) to 301.12μ (SD = 93μ) (P < .001). Mean CFT further decreased to a maximum average of 173.89μ (SD = 147.56μ), from 459.16μ (SD = 47.14μ) to 285.27μ (SD = 77.27μ; t = –25.31, P < .001), within an average of 43.41 (SD = 36.86) days. Mean time to best VA (33.98 days, SD = 24.98) was significantly shorter than mean time to best CFT (43.41 days, SD = 36.86; t = 4.56; P < .001).

For eyes that received multiple injections, the effectiveness of slurry TA did not appear to diminish over time. Over serial injections, the robustness of both VA and OCT improvement was maintained and very predictive of future response in any given patient.

Dose vs Duration

The appropriate dose of intravitreal TA remains a subject of debate. Audren et al and Jonas et al both showed that the use of a 4-mg dose of intravitreal TA does not have enough advantages over the lower 1-mg or 2-mg dose. ^14,37 However, Jonas et al published a comparison between 4-mg and 8-mg doses and showed that the higher dose had a more sustained effect on both VA and central macular thickness, although with a trend to more ocular complications. By using a dose of about 20 mg of TA, the increase in VA was most marked during the first 3 to 6 months after injection and benefit was observable for a period of about 6 to 9 months. ^37,38 In contrast, by using a dose of 4 mg, the duration in the reduction of macular thickness as measured by OCT was less than 6 months. ³⁹

Previous studies have suggested that increasing the dose of intravitreal steroid increases the duration of its effect. Beer and colleagues estimated a 4-mg injection of TA duration to be 3 months in a nonvitrectomized eye based on formal kinetic data. ¹² Chang and Wu used sedimentation to concentrate the TA to approximately 20 mg and found that it lasted in the vitreous for up to 5 months. ³⁶ Anterior chamber measurements by Jonas and Schlichtenbrede demonstrated the presence of TA from 4.1 weeks to 210 weeks following the injection of 20 to 25 mg of TA. ³⁷ They were even able to measure drug levels up to 1.5 years after injection and concluded that a greater injection dose resulted in increased duration of intraocular availability. In a separate study, Jonas estimated the clinical efficacy of concentrated TA to last 7 to 8 months in patients with DME. ³⁸ Spandau and colleagues also demonstrated a direct correlation between the dosage of TA and duration of its effect. ³⁹

Previous vitrectomy affected the duration of efficacy of slurry TA. Prior vitrectomy (74 injections) decreased the mean duration of treatment effect from 11.11 months without vitrectomy to 9.20 months with vitrectomy, a 17% reduction in treatment effect.

The effects of vitrectomy on the pharmacokinetics of intravitreal triamcinolone and DEX implant have also been studied extensively. In vitrectomized rabbit eyes, triamcinolone concentration decreased 1.5 times more rapidly vs nonvitrectomized rabbit eyes. The half-life was also shorter in vitrectomized vs nonvitrectomized eyes at 1.57 days vs 2.89 days, respectively. ^40,41 On the other hand, vitrectomy does not appear to affect the duration of efficacy of the DEX implant. The release kinetics of the DEX implant were examined in 25 vitrectomized and 25 nonvitrectomized Dutch-belted rabbits. There was no statistically significant difference in DEX concentration at any time point. For both groups, the maximum concentration in retina and vitreous was achieved at day 22. By day 31, a similar amount of residual DEX steroid remained in both groups: 5.0 ± 3.3% in nonvitrectomized vs 4.2 ± 5.4% in vitrectomized eyes (P = .47). Clinical observational studies have suggested no difference in the duration of the DEX implant between vitrectomized and nonvitrectomized eyes, on the order of approximately 4 to 5 months ^{42 -44} ; however, slurry TA appears to last almost twice as long at 9.2 months in vitrectomized eyes.

Because vitrectomy is known to decrease the duration of triamcinolone, the degree of PVD on OCT (as a proxy for the degree of vitreous liquefaction) was evaluated as a possible variable for duration of efficacy. Interestingly, there was a trend toward a more syneretic vitreous (PVD 4) to produce a longer duration of effect than a more formed vitreous (PVD 0, 1, 2, or 3), but the great variability within the groups rendered the difference not statistically significant.

Adverse Ocular Effects

Intravitreal steroids are known to have adverse ocular side effects, primarily cataract development and steroid response. In the present study, most eyes were initially already pseudophakic. Of the 31 phakic eyes, 48% required cataract extraction. Although it is difficult to make a direct comparison, this is somewhat lower than the rates seen with fluocinolone acetonide 0.19 mg (75%), ⁴⁵ flucinolone acetonide 0.59 mg (93%), ⁴⁶ DEX 0.7 mg (68%), ⁴⁷ and fluocinolone acetonide 0.18 mg (70%-74%) ⁴⁸ over a 2- to 3-year time frame in pivotal studies. Overall rates of cataract progression or extraction were not reported in the 1-year Ober and Valijan study. ¹⁶

It is difficult to make a direct comparison of the severity of steroid response of slurry TA to currently available, FDA-approved, long-term steroid preparations, which were evaluated in a controlled clinical fashion. In the present study, 57 eyes (39.86%) developed a steroid response (defined as ≥ 10-mm Hg change in IOP from baseline). The average IOP at the time of the initial steroid response was 30.65 (SD = 5.82) with an average maximum-peak IOP of 33.96 (SD = 6.45). Forty-eight eyes (84.21%) were controlled on 2 or fewer drops, 6 eyes (10.50%) required a trabeculectomy, and 3 eyes (5.26%) were monitored without any treatment. In the PeriOcular vs. INTravitreal corticosteroids for uveitic macular edema (POINT) Trial, which compared periocular triamcinolone, intravitreal triamcinolone, and intravitreal DEX implant for the treatment of uveitic macular edema, 39% of eyes developed a steroid response 24 weeks following injection of nonconcentrated, intravitreal triamcinolone. ⁴⁹

In the present study, the average time to IOP response was 13.52 weeks postinjection, suggesting that the steroid response occurs earlier following injection with slurry TA; therefore, more diligent monitoring may be indicated in the 3 to 4 months following injection. The reported rates of steroid response were fluocinolone acetonide 0.19 mg (18%), ⁴⁵ flucinolone acetonide 0.59 mg (59%), ⁴⁶ DEX 0.7 mg (32%), ⁴⁷ and fluocinolone acetonide 0.18-mg (27%) ⁴⁸ in pivotal studies. It should be noted, however, that the criteria to define steroid response were not equivalent among these drug trials. Whereas the criteria were identical for fluocinolone acetonide 0.59 mg and 0.18-mg implants compared with the present study, the criteria were different for fluocinolone acetonide 0.19 mg and DEX 0.7 mg, with a recorded pressure greater than 30 mm Hg and greater than 25 mm Hg, respectively.

Although the number of eyes requiring a trabeculectomy was higher in our study than in some controlled trials (previously reported rates: 2%-37%), ^{21,46 -49} it remained small and may have been falsely elevated by the nature of this study. The present treatment was administered in a live clinical environment, primarily in patients with advanced, recalcitrant disease, who had previously failed numerous treatments. In the Ober and Valijan series, 1 eye (n = 69) required glaucoma-filtering surgery after a C-IVT injection during that 1-year study; however, 2 additional eyes required filtering surgery after C-IVT injections given before or after the covered study period. Overall rates of steroid response were not reported in the Ober and Valijan study. ¹⁶

Spandau et al concluded that higher doses of intravitreal steroid yield a longer treatment response without necessarily a dose-dependent IOP response in patients with DME. ³⁹ Tammewar and colleagues also concluded that a 20-mg injection of TA (concentrated by sedimentation) did not yield a greater risk of IOP elevation than a standard 4-mg injection. ⁵⁰ This may be due to saturation of the receptors at lower doses, which then does not produce any greater steroid effect. Perhaps it is the slow release of the steroid that does not cause large overall increases in IOP.

However, the Standard Care vs. COrticosteroid for REtinal Vein Occlusion (SCORE) study concluded that increasing the dose of triamcinolone from 1 mg to 4 mg increased the incidence and severity of glaucoma and cataracts without improving efficacy for patients with venous occlusive disease. ⁵¹ The SCORE trial used a uniform gel formulation of triamcinolone, Trivaris from Allergan, which never became commercially available and had different fundamental pharmacokinetic properties from either TA or PF-TA.

Acute, self-resolving, limited sterile inflammation has been reported with all preparations of TA. ^{52 -54} This typically presents as an acute, profound loss of vision to the 20/400 or worse level. Patients typically report this event within days of receiving the injection. Evaluation will demonstrate an absence of pain, conjunctival injection, or chemosis. A hypopyon is typically present. The vitreous shows dense opacification and inflammation. The inflammation, as in this series, typically settles within 2 weeks and the macular edema still resolves.

Although the prevailing explanation is that this is a reaction to the preservative benzyl alcohol in the TA-injectable suspension, this does not explain why patients who receive PF-TA also experience sterile inflammation. This was indeed the case in and out of the present series, in which patients who developed sterile inflammation to slurry TA also developed the same phenomenon following intraocular PF-TA. In vitro and in vivo studies more strongly suggested that sterile inflammation is the result of an immune-mediated response directly to the steroid particles themselves or mechanical interaction. ⁵⁵ The particles of TA in PF-TA are smaller than in TA-injectable suspension, which may make them more likely to incite an acute inflammatory response. ⁵⁶ In a large series, Dodwell et al reported more cases of sterile inflammation with intravitreal PF-TA and PF-formulated TA than with TA-injectable suspension. ⁵² Sterile inflammation, therefore, does not appear to be related solely to the preservative in TA-injectable suspension. The exact etiology of the sterile inflammation is not completely understood and may be multifactorial. No cases of human retinal toxicity to intravitreal TA-injectable suspension have ever been reported.

Although no cases of infectious endophthalmitis occurred in this series, there may be a theoretical increased risk of infection because the medication is drawn up and concentrated in a clinical setting without the benefit of sterile manufacturing conditions and preparation (eg, sterile hood, mask, and gloves). To decrease this possible risk, the slurry should always be prepared immediately before injection, with adherence to expected and available sterile clinical techniques. Time between preparation and injection should be kept to a minimum.

Cost

Centrifuged slurry TA is cost-effective. At current US pricing, assuming approximately 3 injections, the cost of 2 to 3 years of therapy would be approximately $39.

Owing to the relatively long duration of action, centrifuge-concentrated slurry TA may also be more cost-effective by decreasing the treatment burden both to the patient and physician.

Limitations

An important limitation of this study includes, first and foremost, the use of an off-label version of TA. Other limitations of this study include the uncontrolled, retrospective nature, variability in CFT measurements over time (spectral-domain vs time-domain technologies), the small subsample groups of certain study variables (eg, diagnoses and steroid responders), and the corresponding high relative variance on some variables. Interpretation of the data and, therefore, expected clinical expectations, should be considered in light of the large variances in the data. There are also possible effects of clustering with respect to using eyes instead of unique patients as the sampling units; however, relatively few patients underwent bilateral injections.

Summary

Slurry TA is a quick, easily reproducible, safe, and cost-effective alternative to more expensive, long-term intraocular steroid delivery. A single injection lasts on average 10.8 months. Although slurry TA is associated with the typical adverse effects of all long-term, intraocular steroids, these appear to be manageable and tolerable for most patients. Sterile inflammation occurs infrequently and is self-limiting. Owing to the low cost, duration, and effectiveness of slurry TA, this treatment should be considered as a viable option for patients with otherwise recalcitrant disease, intolerance to frequent intraocular injections, limited financial means, or difficult access to health care. At times, shortages of other currently available preparations may necessitate the use of this alternative treatment. ⁵⁷ More, the relatively long duration of Slurry TA may be beneficial during restrictions on health care delivery imposed by public health emergencies. ⁵⁸ Future prospective, controlled studies, using larger numbers of patients with specific diagnoses, can further define the worldwide benefit of this alternative treatment in treating patients affected by macular edema.