Review

INTRAVITREAL CORTICOSTEROIDS IN DIABETIC

MACULAR EDEMA
Pharmacokinetic Considerations
YIT YANG, MBCHB, FRCOPHTH,* CLARE BAILEY, MD, FRCP, FRCOPHTH,†
ANAT LOEWENSTEIN, MD, MHA,‡ PASCALE MASSIN, MD, PHD§

Purpose: To review the relationship between kinetics, efficacy, and safety of several
corticosteroid formulations for the treatment of diabetic macular edema.
Methods: Reports of corticosteroid use for the treatment of diabetic macular edema
were identified by a literature search, which focused on the pharmacokinetics, efficacy, and
safety of these agents in preclinical animal models and clinical trials.
Results: Available corticosteroids for diabetic macular edema treatment include intra-
vitreal triamcinolone acetonide, dexamethasone, and fluocinolone acetonide. Because of
differences in solubility and bioavailability, various delivery mechanisms are used.
Bioerodible delivery systems achieve higher maximum concentrations than nonbioerodible
formulations. There is a relationship between visual gains and drug persistence in the
intravitreal compartment. Safety effects were more complex; level of intravitreal triamcin-
olone acetonide exposure is related to development of elevated intraocular pressure and
cataract; this does not seem to be the case for dexamethasone, where two different doses
showed similar mean intraocular pressure and incidence of cataract surgery. With
fluocinolone acetonide, rates of intraocular pressure elevations requiring surgery seem to
be dose related; rates of cataract extraction were similar regardless of dose.
Conclusion: Available corticosteroids for diabetic macular edema exhibit different
pharmacokinetic profiles that impact efficacy and adverse events and should be taken
into account when developing individualized treatment plans.
RETINA 35:2440–2449, 2015

D iabetic macular edema (DME) is a common micro-

vascular complication of diabetes, characterized by
a breakdown in the blood–retina barrier that causes
treally administered anti–vascular endothelial growth
factor (anti-VEGF) agents6–8 or corticosteroids.9,10
The efficacy of laser photocoagulation in reducing risk
intracellular and extracellular accumulation of excess of visual loss in patients with DME has been well
fluid and lipid exudates, leading to retinal thickening. established since it was found to reduce the risk of
Center-involved DME is the most common cause of moderate visual loss from 24% to 12% over 3 years in
moderate vision loss in patients with diabetes and is patients with DME without macular ischemia in the Early
the leading cause of visual impairment in the Treatment Diabetic Retinopathy Study in 1985.4 More
working-age population. The goal of therapy in DME recently, in a 2-year study comparing laser with intra-
is to preserve or improve retinal function and vision by vitreal triamcinolone acetonide (IVTA), moderate
reducing or resolving retinal thickening and edema. visual loss occurred in 14% of patients treated with
Therapeutic measures that have been proven to be laser.5 In 2010, the Diabetic Retinopathy Clinical
effective in reducing risk of visual loss in DME include Research Network (DRCR.net) study (Protocol I) re-
intensive glycemic and blood pressure control,1–3 feno- ported a moderate visual loss rate of 12% in eyes treated
fibrate therapy,3 laser photocoagulation,4,5 and intravi- with laser at 2 years; however, the proportion of

2440
 CORTICOSTEROIDS FOR DME: PK, EFFICACY, SAFETY  YANG ET AL 2441

patients with $15-letter vision gain in the laser arm was Formulations of Triamcinolone Acetonide,
only 17%.11 In the same study, 29% of patients treated Dexamethasone, and Fluocinolone Acetonide
with ranibizumab and deferred laser gained $15 letters
at 2 years. Recently, other anti-VEGF agents, namely Triamcinolone acetonide (TA; Figure 1A) is a syn-
bevacizumab and aflibercept, have also been shown to thetic corticosteroid formulated as an injectable
have similarly improved outcomes compared with laser, suspension and has a 7.5-fold higher anti-
but all three agents require monthly or bimonthly in- inflammatory potency than cortisone.16 Because TA
jections to maintain efficacy.7,12 is water insoluble, it can remain in the vitreous for
Another approach to treatment of DME has been to much longer than other corticosteroids, which are usu-
use intravitreal corticosteroids. Whereas VEGF inhibitors ally eliminated within a few days. Available formula-
act to reduce excess vascular permeability by acting only tions include Kenalog, Trivaris, and Triesence, which
on VEGF, corticosteroids act on several pathways are indicated in the United States for ocular inflamma-
involved in the pathogenesis of DME to reduce the tory conditions unresponsive to topical corticoste-
action of inflammatory cytokines and VEGF, reduce roids.17–19 In the European Union, Kenalog is
leukostasis, and improve other physiologic and structural contraindicated for intraocular injection, and Trivaris
changes, such as tight junction integrity. Corticosteroids and Triesence are not approved for any indication.
that have been extensively evaluated in DME include Dexamethasone (DEX; Figure 1B), which is a less-
IVTA, dexamethasone (DEX), and fluocinolone aceto- potent glucocorticoid receptor agonist than TA, has
nide (FAc).5,9,10,13–15 These three agents are potent and higher solubility in water than TA and requires a slow
selective glucocorticoid receptor agonists with very lim- release delivery system to maintain therapeutic doses
ited activity on the mineralocorticoid receptor. It is likely in the eye.20 The DEX 700-mg posterior segment drug
that intravitreal corticosteroids will become an increas- delivery system (Ozurdex) is a bioerodible implant
ingly important element in our armamentarium for ther- delivered intravitreally using a 22-gauge injector. It
apy in DME. The purpose of this article was to review is indicated in the United States, the European Union,
the published data on these three corticosteroid agents to and other countries worldwide for the treatment of
highlight the differences in potency, solubility, and phar-
DME, as well as for macular edema following branch
macokinetic profiles and to consider the relevance of
retinal vein occlusion or central retinal vein occlusion
these differences to the observed differences in therapeu-
and for inflammation of the posterior segment of the
tic effects seen in clinical studies on DME.
eye presenting as noninfectious uveitis.21,22
Fluocinolone acetonide (FAc; Figure 1C), a more
potent glucocorticoid receptor agonist than TA, has
From the *Royal Wolverhampton Hospitals NHS Trust, Wolver- slightly higher solubility in water than TA and, similar
hampton, United Kingdom; †University Hospitals Bristol NHS
Foundation Trust, Bristol, United Kingdom; ‡Sackler Faculty of to DEX, requires a delivery system for sustained ocu-
Medicine, Tel Aviv University, Tel Aviv, Israel; and §Department lar dosing.20 Two sustained-release devices containing
of Ophthalmology, Lariboisière Hospital, Paris, France.
The authors received editorial assistance from MediTech Media,
FAc, Retisert and ILUVIEN, are indicated for ocular
Hamilton, NJ, funded by Alimera Sciences. use. Retisert contains 0.59 mg of FAc and has an
Yit Yang reports personal fees from Alimera Sciences, Allergan, initial release of approximately 0.6 mg/day in the vit-
Alcon, Bayer, Novartis, and Thrombogenics, during the conduct of
the study. Clare Bailey reports the following: Bristol Eye Hospital reous. It is surgically implanted into the posterior seg-
undertakes clinical trial work funded by Novartis pharmaceuticals, ment of the eye through a pars plana incision.23 In the
honoraria and occasional lectures for attending occasional advisory United States, it is indicated for the treatment of
board meetings; Bristol Eye Hospital took part in the FAME stud-
ies, therefore received commercial funding for recruitment into this chronic noninfectious uveitis affecting the posterior
trial from Alimera Sciences; Bristol Eye Hospital received research segment of the eye. ILUVIEN is a nonbioerodible
funding for clinical trials work from Bayer. Anat Loewenstein re- insert implanted in the eye via injection through the
ports personal fees from Allergan, Alcon, Alimera Sciences, Bayer,
La Roche, Novartis, Teva, during the conduct of the study; reports pars plana using a 25-gauge needle. ILUVIEN con-
personal fees from ForsightLabs, Orabio, Lumenis, Notal Vision, tains 0.19 mg of FAc. It is indicated in Europe for
outside of the submitted work. Pascale Massin reports consultancy the treatment of vision impairment associated with
for Novartis, Allergan, Alimera Sciences, Bayer, and Sanofi, during
the conduct of the study. chronic DME considered insufficiently responsive to
This is an open access article distributed under the terms of the available therapies and in the United States for the
Creative Commons Attribution-NonCommercial-NoDerivatives Li-
cense 4.0 (CC BY-NC-ND), which permits downloading and shar- treatment of DME in patients who have been previ-
ing the work provided it is properly cited. The work cannot be ously treated with a course of corticosteroids and did
changed in any way or used commercially. not have a clinically significant rise in intraocular pres-
Reprint requests: Yit Yang, MD, Royal Wolverhampton Hospi-
tals NHS Trust, New Cross, Wolverhampton, WV10 0QP, United sure (IOP).24,25 In the development of ILUVIEN, 2
Kingdom; e-mail: [email protected] doses were studied, one having an initial release of
2442 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES  2015  VOLUME 35  NUMBER 12

Fig. 1. Chemical structures of

(A) TA,17–19 (B) DEX,21,22 and
(C) FAc.24,25

approximately 0.2 mg/day and the other having an in the anterior chamber at all time points (Figure 2). In
initial release of 0.5 mg/day (similar to Retisert). The the vitreous, the concentrations of triamcinolone were
lower dose elutes FAc from one end of a nonbioerod- highest the first day (14,434.0 ± 10,768 mg/L),
ible polyimide tube, whereas the higher dose elutes whereas in the anterior chamber, the highest concen-
drug from both ends of the polyimide tube. Only the trations were observed 3 days after injection (28.9 ±
lower dose implant has been approved for the treat- 24.5 mg/L). In both sections of the eye, triamcinolone
ment of DME.24,25 levels followed a 2-compartment model of elimina-
tion: an exponential decrease in concentration within
the first 4 weeks (burst period), followed by a steady
Intravitreal Release Kinetics decline over the following months (duration period).
At 8 months, concentrations were 3.3 ± 1.6 mg/L in
Steroids are known to have complex dose–response the anterior chamber and 70.7 ± 37.0 mg/L in the
curves. Efficacy of steroids depends not only on the vitreous, approximately 200 times lower than the con-
potency and dose of drug administered but also on the centration in the burst period. In another similar study,
ability to maintain drug availability in the posterior Ye et al28 examined the pharmacokinetics of IVTA in
segment of the eye. This is dependent on the pharma- rabbits that received 4 mg or 8 mg of IVTA and then
cokinetic profiles produced by the different release had vitreous drug concentrations measured using high-
kinetics of the short- and sustained-release formulations performance liquid chromatography. They found that
in terms of burst height and duration of release. An both doses demonstrated peak at the time of initial
increase in dose of steroid released does not necessarily concentration measurement (burst period) and
result in a corresponding increase in efficacy.10,26 decreased gradually over time (duration period). Fol-
Release kinetics differ between the various formulations lowing injection of 4 mg of IVTA, a concentration of
of intravitreal corticosteroid agents. Evidence from 0.0524 mg/mL was detected in the vitreous after 130
animal and human studies demonstrates the pharmaco- days; with 8 mg of IVTA, 0.2606 mg/mL was present
kinetic profiles of TA, DEX, and FAc. at 150 days after injection. Of note, no significant
Similar release kinetics of IVTA in rabbits’ eyes differences in IOP were seen between the two dosing
have been reported in several studies. Kaamppeter groups, and neither group showed any signs of retinal
et al27 studied 18 white New Zealand rabbits, each damage.
receiving an injection of 6 mg and followed up for 8 The effects of vitrectomy on the pharmacokinetics of
months. The authors found that TA concentrations IVTA and DEX have also been studied extensively. In
were significantly higher in the vitreous samples than vitrectomized rabbit eyes, triamcinolone concentration
decreased 1.5 times more rapidly versus nonvitrectom-
ized rabbit eyes; half-life was also shorter in vitrectom-
ized versus nonvitrectomized eyes (1.57 days vs. 2.89
days).29,30 The release kinetics of the DEX implant
were examined in 25 vitrectomized and 25 nonvitrec-
tomized Dutch-belted rabbits. In nonvitrectomized eyes,
time maximal concentration was 22 days for both retina
and vitreous humor; by Day 31, 5.0 ± 3.3% of DEX
remained in the implant. Results were similar for non-
vitrectomized eyes, with time to maximal concentration
being shorter in the retina (15 days); 4.2 ± 5.4% of
DEX remained after 31 days in vitrectomized eyes.
Fig. 2. Triamcinolone in the vitreous and anterior chamber of rabbits Bhagat et al31 also demonstrated that the release profile
after injection of 6 mg of TA.27 of DEX in rabbit eyes was similar with an intact (one-
 CORTICOSTEROIDS FOR DME: PK, EFFICACY, SAFETY  YANG ET AL 2443

into two hemispheres, with and without the DEX

implant; VH indicates vitreous humor.
The release kinetics of FAc sustained-release im-
plants have been studied in rabbit vitreous with both
the Retisert and the ILUVIEN delivery systems. The
FAc-releasing Retisert implant was surgically placed
into the eyes of anesthetized rabbits via sclerotomy
and suture.33 Near zero–order kinetics were noted with
vitreous humor levels relatively constant from the first
time point (2 hours) through 1 year after implanta-
tion.33 Plasma levels were below the lower limit of
quantitation (200 pg/mL for Retisert) at all times.
Release kinetics of ILUVIEN FAc implants were stud-
ied in pigmented rabbits.34 After intravitreal injection
Fig. 3. Concentration of DEX in monkey eyes.32 Republished with
permission of the Association for Research in Vision and Ophthal- of different doses of FAc implants, exposure in the
mology, from Chang-Lin JE, et al. Invest Ophthalmol Vis Sci vitreous humor was assessed (Figure 4). Quantifiable
2011;52:80–86; permission conveyed through Copyright Clearance FAc concentrations were not observed at any dose
Center, Inc. Adaptations are themselves works protected by copyright.
So in order to publish this adaptation, authorization must be obtained level in the plasma (lower limit of quantitation, 100
both from the owner of the copyright in the original work and from the pg/mL for ILUVIEN). The ILUVIEN delivery system
owner of copyright in the translation or adaptation. and the Retisert system exhibited near zero–order
release kinetic profiles but differed in intraocular expo-
piece) DEX implant and a three-piece implant, suggest-
sure to FAc based on observed vitreous concentra-
ing no clinical significance would be associated with
tions. The overall burst heights of corticosteroid
fragmentation of the DEX implant.
concentration in the vitreous found with FAc implants
The release kinetics of DEX also were studied in 34
were orders of magnitude lower than either TA or the
monkeys after bilateral implantation of 0.7 mg of
bioerodible system releasing DEX (1–20 ng/g for FAc
sustained-release DEX implants.32 Chang-Lin et al32
vs. .1100 ng/g for DEX and .10,000 ng/g for TA).
showed that this bioerodible dose formulation releases
with a significant burst—for approximately 2 months
—before an exponential decline in release (Figure 3).
In Vivo Pharmacokinetics in Human Eyes
After 6 months, DEX was below the limits of detection
for this assay (0.001 ng/mL in the vitreous humor,
Intravitreal Triamcinolone Acetonide
0.001 ng/mL for the retina, and 0.200 ng/mL for
plasma). The difference between intravitreal injection A small study examined the release kinetics of a single
of a bolus dose of TA and a biodegradable release injection of 4 mg of IVTA in human eyes (N = 5).35
form of DEX is most visible in the duration of the They determined from aqueous humor samples that the
burst phase, during which the highest concentration pharmacokinetics followed a 2-compartment model and
of drug is released. the half-life ranged from 76 hours to 635 hours. The
Because the implant was difficult to separate from the mean elimination half-life was 18.6 days in
vitreous humor, the globes were bisected transversely

Fig. 4. Mean (±standard error of the mean) FAc concentration in the

vitreous humor of rabbits.34 Asterisk indicates that a second randomized Fig. 5. Central macular thickness curves calculated from four individ-
treatment was administered during Week 52. ual CMT values after a single injection of IVTA.36
2444 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES  2015  VOLUME 35  NUMBER 12

nonvitrectomized eyes, suggesting that measurable con- Fluocinolone Acetonide

centrations of triamcinolone would be expected to last for
The pharmacokinetics of FAc in human aqueous
approximately 3 months (93 ± 28 days). Audren et al36
was examined in the Fluocinolone Acetonide in
examined the pharmacokinetics of IVTA in a larger
Human Aqueous study, which compared the levels
group of patients with diffuse DME (51 injections
of FAc released in patients with DME receiving the
in 37 eyes). In this study, they used population phar-
ILUVIEN delivery system.14 In this study, 37 patients
macokinetic–pharmacodynamic modeling without
with DMO and 7 with uveitis had aqueous levels of
triamcinolone concentration measurements. They
FAc measured after receiving either a 0.2 mg/day
found that the pharmacokinetic profile of the effect
insert or a 0.5 mg/day insert. Follow-up assessments
of triamcinolone on central macular thickness (CMT)
were done at 1 week after treatment and then again at
was characterized by a curve in 3 phases: a fast
1, 3, 6, 9, 12, 15, 18, 21, 24, 30, and 36 months. A
decrease, a steady state, and a relapse (Figure 5).
companion study assessed the aqueous levels of FAc
Similar to results of the smaller study, the mean esti-
associated with Retisert in patients with noninfectious
mated half-life of triamcinolone ± standard deviation
posterior uveitis.37 Overall, results were similar to
(SD) was 15.4 ± 1.9 days. The mean (±SD) maximum
what was seen in the rabbit studies, with the highest
duration of the effect of a single injection was 140 ±
burst obtained with the Retisert implant (Figure 6).37
17 days, slightly longer than previously observed.
The 0.5-mg/day FAc implant had a higher mean burst
They suggested that the difference could be because
than the 0.2-mg/day ILUVIEN implant. The initial
of the interval between the elimination of TA and
release was followed by a relatively flat profile of
cessation of its pharmacologic activity, which is typ-
release, with a similar submicrogram dose released
ical of corticosteroid agents. The authors further sug-
during the follow-up period for the two doses studied
gest that functional measurements such as mean CMT
with the ILUVIEN system. Through Month 24, the
may be even more useful in the clinical field than
0.5-mg/day implant released at a slightly higher level
measuring the TA concentration because this mea-
than the 0.2-mg/day implant. The 0.2-mg/day implant
surement reflects both the concentration and activity
continued to release FAc through 36 months (end of
of TA.
follow-up). Both Retisert and ILUVIEN were below
Measured and estimated CMT after first injection
the lower limit of quantitation in plasma (200 pg/mL
are after the first injection of IVTA in the right and left
and 100 pg/mL, respectively).23,37 The observed phar-
eye of the same patient. Measured and estimated CMT
macokinetic properties of various formulations of
after first and second injection are after the first and
these corticosteroids are presented in Table 1.
second injection in the same eye of another patient.

Dexamethasone
Impact of Pharmacokinetics on Efficacy in DME
There are no published studies on the pharmaco-
kinetics of DEX 700-mg implants in the human eye. Intravitreal Triamcinolone Acetonide
Plasma levels in the majority of patients were below
Intravitreal triamcinolone acetonide has been eval-
the lower limit of quantitation (50 pg/mL) on Days 7,
uated for the treatment of DME in numerous small and
30, 60, and 90 after administration but were detect-
a few large clinical trials. The majority of data
able in 10 of 73 samples at a range of 52 pg/mL to
regarding the efficacy of IVTA in patients with
94 pg/mL.22
DME suggests that efficacy is greatest 1 month after
injection and that these effects are no longer significant
at 6 months. A single injection of ITVA (4 mg/0.1 mL)
or bevacizumab (1.5 mg/0.06 mL) was examined in
patients with refractory DME, defined as having at
least 1 previous macular laser photocoagulation treat-
ment.38 The single intravitreal injection resulted in
peak functional and anatomical improvement (best-
corrected visual acuity [BCVA] and CMT) between
Weeks 4 and 8, after which these parameters returned
to baseline by Week 24 (Figure 7). Similarly, in a pro-
spective controlled trial of IVTA for refractory diffuse
DME, CMT decreased after injection and was signif-
Fig. 6. Mean human aqueous concentrations of FAc implants.37 icantly lower at Weeks 4 and 12; however, this was no
 CORTICOSTEROIDS FOR DME: PK, EFFICACY, SAFETY  YANG ET AL 2445

Table 1. Observed Pharmacokinetic Properties of Approved Ocular Steroid Formulations

Duration of Therapeutic
Corticosteroid Formulation Total Dose Initial Daily Release Procedure Window
Ozurdex 700 mg Unknown in humans Injectable 2–4 months
IVTA (Triesence Trivaris Kenalog-40) 4 mg Unknown Injectable 1–3 months
Retisert 0.59 mg 0.6 mg/day Incision and suture 30 months
ILUVIEN 0.19 mg 0.2 mg/day Injectable Up to 36 months

longer significant at Week 24 because of recurrence of laser, −2 with 1 mg of IVTA, and −3 with 4 mg of
macular edema in 5 of 12 eyes.39 IVTA).5 Although loss of visual gains in these patients
Bonini-Filho et al40 examined intravitreal injection could occur through a variety of mechanisms, the tim-
versus sub-Tenon’s infusion of TA for refractory DME ing of both the visual gains and the loss are concordant
in a randomized clinical trial of 28 eyes. Visual acuity with the timing of the measured intravitreal concen-
significantly improved from baseline among patients trations of IVTA in rabbit eyes.27 This may reflect
receiving IVTA at 4, 8, and 12 weeks after injection a correlation between intravitreal concentration of
but was no longer significant at Week 24. These im- IVTA and therapeutic effect.
provements, and the improvements in CMT, were
Dexamethasone
greater with IVTA versus sub-Tenon’s infusion, sug-
gesting that intravitreal injections may be more effec- The clinical effect of the pharmacokinetics of DEX
tive in terms of both functional and anatomical intravitreal implants can be seen in the PLACID study,
outcomes.40 A meta-analysis of visual acuity after IV- which compared DEX intravitreal implants plus laser
TA for DME refractory to laser treatment indicated to laser alone in patients with DME, most of whom
that the maximum improvement in BCVA was found had not received previous medical treatment for
1 month after injection.41 Finally, in the large, ran- DME.42 The observed pattern in mean change in
domized DRCR.net Protocol B study, patients with BCVA after the first implant reflects the kinetics of
center-involved DME received either focal/grid laser drug release in vivo (Figure 8). Mean BCVA
photocoagulation (control group), 1 mg of IVTA, or 4 improved by slightly more than 6 letters at 1 month
mg of IVTA, with eligibility for retreatment at 4- and subsequently declined through Month 6. After that
month intervals if persistent or recurrent DME was point, two thirds of patients were retreated, resulting in
present. The mean number of injections received by a second spike in the BCVA curve.27 Similarly, in
patients in the 4-mg IVTA group was 3.1 over 2 years. vitrectomized eyes, the largest increase in BCVA (6
At the first follow-up visit (4 months), the patients letters) was noted 8 weeks after implant; by 26 weeks,
randomized to the 4-mg IVTA group had the greatest mean increase in BCVA had dropped to 3 letters.43
improvement in visual acuity (mean difference com- Similar to the mean BCVA profile in the PLACID
pared with laser, adjusted for prior photocoagulation study, results of the Phase 3 Macular Edema:
and baseline visual acuity = +3.8 letters); however, at Assessment of Implantable Dexamethasone in Dia-
the Month 24 visit, patients randomized to laser had betes (MEAD) study of Ozurdex for DME showed
the greatest improvement in visual acuity (+1 with improvement after injection at Months 1.5 and 3,

Fig. 7. Central macular thickness

(A) and BCVA (B) after a single
injection of triamcinolone or bev-
acizumab.38 *P , 0.05 and **P ,
0.01 for within-group changes
from baseline. Reproduced from
Paccola L, et al. Br J Ophthalmol
2008;92:76–80, with permission
from BMJ Publishing Group Ltd.
Adaptations are themselves works
protected by copyright. So in order
to publish this adaptation, authori-
zation must be obtained both from
the owner of the copyright in the
original work and from the owner
of copyright in the translation or
adaptation.
2446 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES  2015  VOLUME 35  NUMBER 12

In the FAME trials, which examined the safety and

efficacy of ILUVIEN in patients with DME, the
greatest treatment effect was seen in patients with
chronic DME.17 To further examine efficacy in the
context of the pharmacokinetic results, the mean
change in BCVA was determined for patients with
chronic DME who received only a single ILUVIEN
implant during FAME. Figure 9A demonstrates that
efficacy was maintained through Month 36 in these
patients (data on file, Alimera Sciences). A majority
of patients who were phakic at baseline (80%) required
cataract surgery; however, those patients who under-
went cataract surgery during the study still experi-
enced a visual benefit. In patients who were
Fig. 8. Mean change from baseline BCVA with DEX implant versus
sham in the PLACID study.42 Reprinted from Callanan DG, et al. pseudophakic at baseline (Figure 9B; data on file, Ali-
Ophthalmology 2013;120:1843–1851, with permission from Elsevier. mera Sciences), there was stability of visual acuity on
Adaptations are themselves works protected by copyright. So in order to follow-up, particularly in a subgroup of patients with
publish this adaptation, authorization must be obtained both from the
owner of the copyright in the original work and from the owner of chronic DME and pseudophakia at baseline. Overall,
copyright in the translation or adaptation. these data suggest that a sustained, low dose of steroid
delivered to the eye can result in long-term efficacy.
followed by a decline through Month 6 when patients
were eligible for reinjection.9 The decline after the 3- Impact of Pharmacokinetics on Safety in DME
month time point was evident even among patients
who were pseudophakic at baseline. Overall, these The most common complications associated with
studies suggested that a retreatment interval shorter ocular steroids are elevated IOP and cataract develop-
than 6 months may be needed to provide sustained ment. Elevated IOP associated with corticosteroid
benefit with DEX implants, consistent with the phar- injection results from the interaction of the steroid with
macokinetic profile. the trabecular meshwork of the anterior chamber of the
eye.44 It is possible that minimizing steroid exposure in
Fluocinolone Acetonide the front of the eye may decrease the interaction with
the trabecular meshwork and thereby decrease the inci-
Results of the 3-year Retisert in DME trial demon- dence of IOP-related events. Risk factors for developing
strated that differences in retinal thickness and visual ocular hypertension in response to steroid exposure
acuity were significant for Retisert versus standard of include documented or suspected glaucoma, family his-
care at 2 years, but these differences were no longer tory of glaucoma, and younger age.45,46 The potential
significant at 3 years.15 The timing of these results is impact on the release profile of IVTA, DEX, and FAc
consistent with the 30-month duration of release from were examined in published trials.
the implant. Similarly, consistent with the sustained
release of drug through the end of the 36-month
Intravitreal Triamcinolone Acetonide
Fluocinolone Acetonide in Human Aqueous study,
0.2-mg/day FAc ILUVIEN implants resulted in sus- Elevated IOP usually occurs 1 to 2 months after
tained efficacy through 36 months.10,13,14,37 injection of IVTA but has been found to occur

Fig. 9. Mean change in BCVA

with a single treatment (sham
control or FAc implant) in the
FAME study among all patients
with chronic DME (A) and pa-
tients with chronic DME who
were pseudophakic at baseline
(B) (data on file, Alimera
Sciences).
 CORTICOSTEROIDS FOR DME: PK, EFFICACY, SAFETY  YANG ET AL 2447

anywhere between 1 and 13 weeks postinjection.44

Following a single injection of IVTA in patients with
refractory DME in the Intravitreal Triamcinolone Ace-
tonide Versus Intravitreal Bevacizumab for Refractory
Diabetic Macular Edema (IBEME) study, the highest
IOP was observed by Week 4, followed by a decline to
baseline over time, consistent with the timing of max-
imum efficacy.38 In the DRCR.net Protocol B trial,
IOP elevations of $10 mmHg were observed in
33% of patients receiving 4 mg of IVTA, and 30%
of these patients required IOP-lowering medications.5
These results are based on 2 years of follow-up and Fig. 10. Mean change from baseline IOP in the FAME study (data on
a mean of 3.1 injections of 4 mg of IVTA. A single file, Alimera Sciences).
dose of IVTA in the IBEME study did not result in
cataract progression during the 24-month follow-up38; recorded in 61.4% of implanted eyes, and 33.8%
however, a dose–response relationship was observed required surgery for ocular hypertension by 4 years.15
between IVTA exposure and cataract formation in the Fewer IOP-related events were associated with ILU-
DRCR.net Protocol B study, where 51% of patients VIEN, particularly IOP-lowering surgery (over 3
receiving 4 mg of IVTA injections required cataract years, 4.8% of patients in the 0.2-mg/day FAc group
surgery compared with 23% in the 1-mg IVTA arm and 8.1% of those in the 0.5-mg/day FAc group).10
(P , 0.001).5 As with IVTA, timing of IOP elevation seemed to
mirror the release of steroid from the FAc implant
Dexamethasone (Figure 10). The dose of FAc did not seem to impact
Elevated IOP was seen following treatment with the rate of cataract, with cataract extraction occurring
Ozurdex treatment in the PLACID study, with 16.8% in 91% of phakic eyes treated with Retisert, 87.2%
and 4% of eyes experiencing an IOP of $25 mmHg treated with 0.5 mg/day FAc, and 80% treated with
and $35 mmHg, respectively, at some point during ILUVIEN 0.2 mg/day FAc.10,15
the study.42 By Month 12, no eyes had IOP $25 Overall, the timing of IOP elevation with the
mmHg. Haller et al47 observed that most cases of ele- steroids studied suggests that elevated IOP mirrors
vated IOP occurred during the first 60 days after treat- drug release in the eye and returns to baseline as drug
ment, during the initial burst phase of release. In the is depleted. Cataract incidence is not related to the
MEAD study, the highest mean change in IOP in the pharmacokinetics observed with a single short-term
overall population occurred within 3 months after release corticosteroid but rather to sustained dosage
Ozurdex implant and subsequently declined by Month over an extended time. This was seen both with
6. These results occurred for each retreatment through- sustained-release steroid implants and with repeated
out the study.9 Mean IOP was similar in the 700-mg injections of short-release steroids formulations.
and 350-mg dose groups. In studies with shorter
follow-up, such as the PLACID study, rates of cataract
Discussion
surgery were not greater in patients treated with Ozur-
dex implants than in patients treated with laser alone;
The importance of corticosteroids in the treatment of
however, cataract-related adverse events were more
DME has been established, including emerging evi-
frequent in the DEX group versus the laser group
dence from Phase 3 trials supporting the need for
(22.2% vs. 9.5%; P = 0.017).42 In studies with longer
multifactorial treatment in patients with longer dura-
follow-up, such as the 3-year MEAD study, cataract
tion disease who do not respond to highly specific
surgery was performed on 59.2% of patients receiving
treatments (i.e., VEGF inhibitors).10,42 The corticoste-
700 mg of DEX and 52.3% receiving 350 mg of DEX
roids that have been examined for treatment of DME
(35.9% and 33.7% of patients discontinued the study
include TA, DEX, and FAc, which differ in potency,
in the 700-mg and 350-mg DEX arms, respectively).9
solubility, and pharmacokinetic profile. Because of the
insolubility of TA, a bolus injection is used, which
Fluocinolone Acetonide
forms a crude depot inside the eye that slowly dis-
With FAc treatment, dose seemed to have an effect solves. Dexamethasone, the most soluble of the corti-
on the incidence of IOP-related adverse events. With costeroids, must be delivered via a bioerodible insert.
Retisert treatment of DME, IOP $30 mmHg was Currently, only FAc is delivered via a nonbioerodible
2448 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES  2015  VOLUME 35  NUMBER 12

system. These different delivery mechanisms contrib- some advantage in clinical practice to select a particular
ute to the different therapeutic durations, which range corticosteroid for a particular clinical scenario based
from months to years. For drugs with a high burst on our understanding of how the differences in phar-
(IVTA, DEX implant), the greatest effect is seen dur- macokinetic profiles can be considered for the individ-
ing this initial phase, followed by waning during the ual needs of patients.
duration phase. For those with a low burst and near Key words: diabetic macular edema, intravitreal
zero–order kinetics (FAc implant), efficacy is main- corticosteroids, pharmacokinetics, bioerodible im-
tained throughout the duration phase. There could be plants, triamcinolone acetonide, dexamethasone, fluo-
many explanations for this effect, one of which is that cinolone acetonide, best-corrected visual acuity,
a very high initial dose could “shock the system,” intraocular pressure, cataract.
making cells dependent on a high concentration of
drug for continued effect. References
Because of the known adverse events related to
corticosteroid use in the eye, one of the goals of 1. Holman RR, Paul SK, Bethel MA, et al. Long-term follow-up
after tight control of blood pressure in type 2 diabetes. N Engl J
corticosteroid administration is maximizing efficacy Med 2008;359:1565–1576.
while minimizing cataract development and IOP 2. UK Prospective Diabetes Study Group. Tight blood pressure con-
elevation. Minimization of exposure of the trabecu- trol and risk of macrovascular and microvascular complications in
lar meshwork through posterior positioning and type 2 diabetes: UKPDS 38. UK Prospective Diabetes Study
a low dose of corticosteroid may ameliorate the Group. BMJ 1998;317:703–713.
3. ACCORD Study Group; ACCORD Eye Study Group;
IOP events, as evidenced by the difference seen with Chew EY, Ambrosius WT, Davis MD, et al. Effects of medical
the two FAc delivery systems, Retisert and ILU- therapies on retinopathy progression in type 2 diabetes. N Engl
VIEN. The rate of IOP-lowering surgery was 7-fold J Med 2010;363:233–244.
lower with ILUVIEN than Retisert, possibly because 4. Early Treatment Diabetic Retinopathy Study Research Group.
of the lower dose and more posterior positioning of Photocoagulation for diabetic macular edema. Early Treatment
Diabetic Retinopathy Study report number 1. Arch Ophthalmol
ILUVIEN. 1985;103:1796–1806.
The studies described herein also have limitations 5. Diabetic Retinopathy Clinical Research Network. A random-
in the interpretation of results. The patients in each ized trial comparing intravitreal triamcinolone acetonide and
trial had different baseline characteristics, including focal/grid photocoagulation for diabetic macular edema. Oph-
some who were treatment-naive and some who were thalmology 2008;115:1447–1449; 1449.e1–1449.e10.
6. Cunningham ET Jr., Adamis AP, Altaweel M, et al. A phase II
refractory to laser. Differences in the resiliency of
randomized double-masked trial of pegaptanib, an anti-
their photoreceptor systems could impact cellular vascular endothelial growth factor aptamer, for diabetic macu-
response to steroid exposure. Some trials included lar edema. Ophthalmology 2005;112:1747–1757.
only the use of a steroid injection, whereas in others, 7. Rajendram R, Fraser-Bell S, Kaines A, et al. A 2-year prospective
patients received laser or other treatments, which randomized controlled trial of intravitreal bevacizumab or laser
could further impact the efficacy and confound therapy (BOLT) in the management of diabetic macular edema:
24-month data: report 3. Arch Ophthalmol 2012;130:972–979.
results. Data in all cases are not directly comparable 8. Brown DM, Nguyen QD, Marcus DM, et al. Long-term out-
because of differing use of standard of care treat- comes of ranibizumab therapy for diabetic macular edema: the
ments; however, the totality of evidence across trials 36-month results from two phase III trials: RISE and RIDE.
can give insights to the duration and clinical profile of Ophthalmology 2013;120:2013–2022.
various steroid treatments. 9. Boyer DS, Yoon YH, Belfort R Jr., et al. Three-year, random-
ized, sham-controlled trial of dexamethasone intravitreal
In conclusion, the published data suggest that there implant in patients with diabetic macular edema. Ophthalmol-
are distinct differences in pharmacokinetic profiles ogy 2014;121:1904–1914.
between IVTA formulations, the bioerodible DEX 10. Campochiaro PA, Brown DM, Pearson A, et al. Sustained
implant, and nonbioerodible FAc implants. These delivery fluocinolone acetonide vitreous inserts provide benefit
differences may well account for the differences in for at least 3 years in patients with diabetic macular edema.
Ophthalmology 2012;119:2125–2132.
duration of treatment response and adverse events, 11. Diabetic Retinopathy Clinical Research Network; Elman MJ,
such as cataract formation and elevated IOP. Given the Aiello LP, Beck RW, et al. Randomized trial evaluating rani-
presence of multiple pathogenic mechanisms that drive bizumab plus prompt or deferred laser or triamcinolone plus
the development of macular edema in a variety of prompt laser for diabetic macular edema. Ophthalmology
retinal diseases, it is likely that corticosteroids will 2010;117:1064–1077.e35.
12. Nguyen QD. Two-year outcomes of the VISTA/VIVID trials
form an important part of our armamentarium in the
of intravitreal aflibercept injection in diabetic macular edema.
therapeutic management of patients with such retinal 2014. AAO Meeting Abstracts 2014; [abstract PA083]. Avail-
conditions. Given these corticosteroids’ different pro- able at: https://secure.aao.org/apps/MeetingArchive/tabid/433/
files of efficacy, duration, and safety, it may be of Default.aspx. Accessed July 28, 2015.
 CORTICOSTEROIDS FOR DME: PK, EFFICACY, SAFETY  YANG ET AL 2449

13. Campochiaro PA, Brown DM, Pearson A, et al. Long-term vitrectomized and nonvitrectomized eyes. Invest Ophthalmol
benefit of sustained-delivery fluocinolone acetonide vitreous Vis Sci 2011;52:4605–4609.
inserts for diabetic macular edema. Ophthalmology 2011; 31. Bhagat R, Zhang J, Farooq S, Li XY. Comparison of the
118:626–635.e2. release profile and pharmacokinetics of intact and fragmented
14. Campochiaro PA, Hafiz G, Shah SM, et al. Sustained ocular dexamethasone intravitreal implants in rabbit eyes. J Ocul
delivery of fluocinolone acetonide by an intravitreal insert. Pharmacol Ther 2014;30:854–858.
Ophthalmology 2010;117:1393–1399.e3. 32. Chang-Lin JE, Attar M, Acheampong AA, et al. Pharmacoki-
15. Pearson PA, Comstock TL, Ip M, et al. Fluocinolone acetonide netics and pharmacodynamics of a sustained-release dexameth-
intravitreal implant for diabetic macular edema: a 3-year mul- asone intravitreal implant. Invest Ophthalmol Vis Sci 2011;52:
ticenter, randomized, controlled clinical trial. Ophthalmology 80–86.
2011;118:1580–1587. 33. Driot JY, Novack GD, Rittenhouse KD, et al. Ocular pharma-
16. Mansoor S, Kuppermann BD, Kenney MC. Intraocular cokinetics of fluocinolone acetonide after retisert intravitreal
sustained-release delivery systems for triamcinolone acetonide. implantation in rabbits over a 1-year period. J Ocul Pharmacol
Pharm Res 2009;26:770–784. Ther 2004;20:269–275.
17. TRIESENCE (triamcinolone acetonide injectable suspension) 34. Kane FE, Green KE. Ocular pharmacokinetics of fluocinolone
40 mg/mL [package insert]. Fort Worth, TX: Alcon acetonide following iluvien implantation in the vitreous humor
Laboratories; 2013. Available at: http://ecatalog.alcon.com/pi/ of rabbits. J Ocul Pharmacol Ther 2015;31:11–16.
Triesence_us_en.pdf. Accessed July 28, 2015. 35. Beer PM, Bakri SJ, Singh RJ, et al. Intraocular concentration
18. TRIVARIS (triamcinolone acetonide injectable suspension) 80 and pharmacokinetics of triamcinolone acetonide after a single
mg/mL [package insert]. II, Irvine, CA: Allergan, Inc. 2008. intravitreal injection. Ophthalmology 2003;110:681–686.
Available at: http://www.drugs.com/pro/trivaris.html. Ac- 36. Audren F, Tod M, Massin P, et al. Pharmacokinetic-pharma-
cessed July 28, 2015. codynamic modeling of the effect of triamcinolone acetonide
19. KENALOG-40 INJECTION (triamcinolone acetonide inject- on central macular thickness in patients with diabetic macular
able suspension, USP) [package insert]. Princeton, NJ: edema. Invest Ophthalmol Vis Sci 2004;45:3435–3441.
Bristol-Myers Squibb Company; 2011. Available at: http:// 37. Campochiaro PA, Nguyen QD, Hafiz G, et al. Aqueous levels
packageinserts.bms.com/pi/pi_kenalog-40.pdf. Accessed July of fluocinolone acetonide after administration of fluocinolone
28, 2015. acetonide inserts or fluocinolone acetonide implants. Ophthal-
20. Edelman JL. Differentiating intraocular glucocorticoids. Oph- mology 2013;120:583–587.
thalmologica 2010;224:S25–S30. 38. Paccola L, Costa RA, Folgosa MS, et al. Intravitreal triamcin-
21. electronic Medicines Compendium (eMC). Ozurdex. Available olone versus bevacizumab for treatment of refractory diabetic
at: http://www.medicines.org.uk/emc/medicine/23422. Ac- macular oedema (IBEME study). Br J Ophthalmol 2008;92:
cessed October 23, 2014. 76–80.
22. Ozurdex (dexamethasone intravitreal implant) [package insert]. 39. Massin P, Audren F, Haouchine B, et al. Intravitreal triamcin-
Irvine, CA: Allergan, Inc; 2014. Available at: http://www. olone acetonide for diabetic diffuse macular edema: prelimi-
allergan.com/assets/pdf/ozurdex_pi.pdf. Accessed October 6, 2014. nary results of a prospective controlled trial. Ophthalmology
23. Bausch & Lomb. RETISERT (fluocinolone acetonide intravi- 2004;111:218–224; discussion 224–225.
treal implant) 0.59 mg [package insert]. Rochester, NY: 40. Bonini-Filho MA, Jorge R, Barbosa JC, et al. Intravitreal injec-
Bausch & Lomb; 2012. Available at: http://www.bausch. tion versus sub-tenon’s infusion of triamcinolone acetonide for
com/Portals/107/-/m/BL/United%20States/USFiles/Package% refractory diabetic macular edema: a randomized clinical trial.
20Inserts/Pharma/retisert-prescribing-information.pdf. Ac- Invest Ophthalmol Vis Sci 2005;46:3845–3849.
cessed July 28, 2015. 41. Rudnisky CJ, Lavergne V, Katz D. Visual acuity after intra-
24. ILUVIEN summary of product characteristics. Aldershot, vitreal triamcinolone for diabetic macular edema refractory to
United Kingdom: Alimera Sciences F Limited; 2013. Avail- laser treatment: a meta-analysis. Can J Ophthalmol 2009;44:
able at: https://www.medicines.org.uk/emc/medicine/27636. 587–593.
Accessed July 28, 2015. 42. Callanan DG, Gupta S, Boyer DS, et al. Dexamethasone intra-
25. ILUVIEN (fluocinolone acetonide intravitreal implant) [package vitreal implant in combination with laser photocoagulation for
insert]. Alpharetta, GA: Alimera Sciences, Inc; 2014. Available the treatment of diffuse diabetic macular edema. Ophthalmol-
at: http://iluvien.com/downloads/iluvien-prescribing-information. ogy 2013;120:1843–1851.
pdf. Accessed October 6, 2014. 43. Boyer DS, Faber D, Gupta S, et al. Dexamethasone intravitreal
26. Hauser D, Bukelman A, Pokroy R, et al. Intravitreal triamcin- implant for treatment of diabetic macular edema in vitrectom-
olone for diabetic macular edema: comparison of 1, 2, and 4 ized patients. Retina 2011;31:915–923.
mg. Retina 2008;28:825–830. 44. Razeghinejad MR, Katz LJ. Steroid-induced iatrogenic glau-
27. Kamppeter BA, Cej A, Jonas JB. Intraocular concentration of coma. Ophthalmic Res 2012;47:66–80.
triamcinolone acetonide after intravitreal injection in the rabbit 45. Armaly MF. Effect of corticosteroids on intraocular pressure
eye. Ophthalmology 2008;115:1372–1375. and fluid dynamics. II. The effect of dexamethasone in the
28. Ye YF, Gao YF, Xie HT, Wang HJ. Pharmacokinetics and glaucomatous eye. Arch Ophthalmol 1963;70:492–499.
retinal toxicity of various doses of intravitreal triamcinolone 46. Davies TG. Tonographic survey of the close relatives of pa-
acetonide in rabbits. Mol Vis 2014;20:629–636. tients with chronic simple glaucoma. Br J Ophthalmol 1968;
29. Chin HS, Park TS, Moon YS, Oh JH. Difference in clearance 52:32–39.
of intravitreal triamcinolone acetonide between vitrectomized 47. Haller JA, Kuppermann BD, Blumenkranz MS, et al. Random-
and nonvitrectomized eyes. Retina 2005;25:556–560. ized controlled trial of an intravitreous dexamethasone drug
30. Chang-Lin JE, Burke JA, Peng Q, et al. Pharmacokinetics of delivery system in patients with diabetic macular edema. Arch
a sustained-release dexamethasone intravitreal implant in Ophthalmol 2010;128:289–296.