DIABETIC MACULAR EDEMA

ASSOCIATED WITH GLITAZONE USE

EDWIN H. RYAN JR, MD,* DENNIS P. HAN, MD,†
ROBERT C. RAMSAY, MD,* HERBERT L. CANTRILL, MD,*
STEVEN R. BENNETT, MD,* SUNDEEP DEV, MD,*
DAVID F. WILLIAMS, MD*

Purpose: To describe diabetic macular edema (DME) in patients who developed fluid
retention as a consequence of glitazone use.
Methods: A chart review identified 30 patients who used pioglitazone or rosiglitazone
and had both lower extremity edema and macular edema. Clinical reports, photographs,
and fluorescein angiograms were reviewed. Patients followed for ⬎3 months were ana-
lyzed separately.
Results: Seventeen patients took oral pioglitazone, 11 took rosiglitazone, and 2 took
both drugs at different times. Eleven patients were observed for ⬎3 months after cessation
of glitazones. Mean weight gain during drug administration in this group was 30 lb, and
mean weight loss after drug discontinuation was 19 lb. Rapid reduction in macular edema
off drug occurred in only 4 of 11 patients, but 8 of 11 had reduced edema over 2 years.
Mean visual acuity in this group at the initial visit was 20/60, and at the final visit, it was
20/85. Four eyes of three patients had resolution of diffuse macular edema with improved
vision after cessation of glitazones without laser treatment.
Conclusions: Fluid retention occurs in 5% to 15% of patients taking glitazones. In some
of these patients, glitazone use appears to be a cause of macular edema, and drug
cessation appears to result in rapid resolution of both peripheral and macular edema. Fluid
retention associated with glitazone use should be considered when assessing treatment
options for patients with DME, especially those with concomitant peripheral edema.
RETINA 26:562–570, 2006

G litazones (thiazolidinediones) are commonly used

oral agents that reduce insulin resistance in type 2
diabetes and include rosiglitazone and pioglitazone.
Glitazones are effective for glycemic control as mono-
therapy in combination with other oral agents or with
insulin.1 The two drugs currently on the market in-
clude pioglitazone and rosiglitazone, with troglitazone
having been withdrawn from the market in 2000 after
From *VitreoRetinal Surgery, P.A., Minneapolis, Minnesota;
and †The Eye Institute, Medical College of Wisconsin, Milwaukee,
Wisconsin.
Presented at the American Academy of Ophthalmology Annual
Meeting; Anaheim, CA; November 2003.
None of the authors have financial interest in the subject matter.
Reprint requests: Edwin H. Ryan Jr, MD, 7760 France Avenue
South, Suite 310, Minneapolis, MN 55435.
deaths due to liver toxicity.2 Fluid retention and pe-
ripheral edema are described in 5% to 7% of patients
using glitazones alone3–5 or in conjunction with other
oral agents or up to 15% of patients using pioglitazone
with insulin.5 This fluid retention is refractory to treat-
ment with diuretics.6 Systemic fluid retention, usually
caused by congestive heart failure or renal disease,
can aggravate diabetic macular edema (DME) and
complicate its management.7–9 We suspect that fluid
retention associated with glitazone use in some cases
leads to worsening of DME.
To our knowledge, this association has been iden-
tified in one case report.10 The two clinical objectives
of the current study were to identify patients in our
clinical practice who were noted to have both clini-

562
 Table 1. Glitazones and Peripheral Edema—Grouped by Diffuse Versus Focal Macular Edema
Initial
Examination Study Weight Weight Macular
Date (Date Entry Present Gain Loss Con- Edema
Glitazone Last Diabetes Vision Vision Drug on off gestive Other (Focal, Laser Additional
Use Examination Age Duration (OD/ (OD/ Glitazone Drug Start Stop Drug Drug Heart Hyper- Medical Diffuse, (Focal, Diabetes
Patient Noted) Date (y) Sex (y) OS)* OS)* Drug(s) Date Date (lb) (lb) Failure tension History Both) PRP) Medication

1 4/00 (4/03) 5/03 71 M 4 20/150 100/150 Rosiglitazone 4/02 7/03 Negative 20 Negative Positive Respiratory Both OU F&P Glucovance
insufficiency
2 (6/03) 6/03 51 F 3 100/70 70/50 Pioglitazone 2001 NA Positive NA Negative Negative Hyper- Focal OD, Focal None
thyroidism diffuse OS
3 7/02 (5/03) 9/03 38 F 1 40/20 60/150 Pioglitazone 9/02 5/03 36 10 Negative Negative — Both OU F&P Glucovance,
insulin
4 8/01 (4/03) 5/03 67 M 5 25/150 30/70 Pioglitazone 6/02 NA Unknown NA Positive Negative — Diffuse OU Focal Glipizide
5 (4/03) 6/03 71 M 20 60/40 50/30 Pioglitazone 2001 NA Negative NA Negative Positive Myocardial Diffuse OU Focal Insulin,
infarct amaryl
6 4/98 (5/03) 8/03 70 F 9 200/200 400/400 Rosiglitazone 2001 4/03 50 18 Negative Positive — Both OU Focal Insulin
7 3/02 (8/03) 3/04 65 M 2 20/60 70/400 Rosiglitazone 2000 NA 15 NA Negative Negative — Both OU Focal None
OU
8 (9/03) 9/03 66 M 10 200/70 200/70 Rosiglitazone Unknown 9/03 Unknown NA Positive Positive Myocardial Both OU Focal —
infarct
9 9/02 (6/03) 8/03 74 F 8 800/30 400/40 Pioglitazone 8/02 6/03 Positive Negative Negative Positive Colon Both OU Focal Glyburide,
cancer glucophage
10 5/01 (1/02) 10/03 61 M 10 60/400 60/HM Rosiglitazone, Rosiglitazone, 1/02 30 30 Negative Positive Renal failure Diffuse OU F&P Insulin
pioglitazone 11 mo; OU
pioglitazone,
3 mo
DME WITH GLITAZONE USE

11 4/00 (9/01) 10/03 60 F 40 60/70 300/100 Rosiglitazone 9/99 9/01 15 0 Positive Positive High Both OU F&P Insulin
●

cholesterol OU
12 (9/01) 2/04 52 F 16 40/25 150/40 Rosiglitazone, Rosiglitazone, 9/01 30 30 Negative Positive — Both OU Focal Glucophage,
pioglitazone 2 mo; OU glyburide
pioglitazone,
3 mo
13 (11/99) 1/00 46 M 19 50/80 25/40 Rosiglitazone 8/99 11/99 24 24 Positive Positive — Diffuse OU Prior Glucophage,
PRP insulin
14 (9/02) 3/03 65 F 11 60/50 50/50 Pioglitazone Unknown 7/02 Positive 12 Negative Positive Asthma, Diffuse OU Focal Insulin,
RYAN JR ET AL

colon amaryl
cancer
15 10/02 (7/03) 7/03 61 M 12 25/25 20/20 Pioglitazone 2001 5/03 30 15 Negative Negative Neuropathy Diffuse OU Focal Glucophage,
insulin
16 9/01 (7/02) 7/02 55 M 2 60/40 60/70 Pioglitazone 2002 NA Positive Negative Negative Positive Nephropathy, Diffuse OU Focal Glucophage
neuropathy
17 (8/03) 11/03 72 M 4 60/40 40/30 Rosiglitazone 3/03 8/03 15 28 Negative Positive Neuropathy Diffuse OU Focal Amaryl
18 7/02 (9/03) 11/03 62 M 7 40/80 60/150 Pioglitazone Unknown NA Unknown Negative Negative Positive Nephropathy, Diffuse OU Focal, Oral agent,
neuropathy PRP unknown
19 (8/03) 8/03 56 F 12 150/20 150/20 Pioglitazone Unknown NA Positive NA Negative Negative — Diffuse OD Focal Insulin
Diffuse — — Mean, — — Mean Mean — — — Mean, Mean, — — — — — —
leakage 62 visual visual 23 19
acuity, acuity,
20/60 20/84
(Table continues)
563
 Table 1. (Continued)
564

Initial
Examination Study Weight Weight Macular
Date (Date Entry Present Gain Loss Con- Edema
Glitazone Last Diabetes Vision Vision Drug on off gestive Other (Focal, Laser Additional
Use Examination Age Duration (OD/ (OD/ Glitazone Drug Start Stop Drug Drug Heart Hyper- Medical Diffuse, (Focal, Diabetes
Patient Noted) Date (y) Sex (y) OS)* OS)* Drug(s) Date Date (lb) (lb) Failure tension History Both) PRP) Medication

20 3/01(7/03) 9/03 66 M 10 50/20 30/30 Rosiglitazone 4/03 7/03 Negative Negative Negative Positive Stroke Focal OD Focal Insulin,
glucophage
21 (8/02) 10/03 73 F 20 150/20 100/20 Pioglitazone 5/02 8/02 40 10 Negative Positive — Focal OU Focal Glyburide,
OU glucophage
22 (8/03) 8/03 74 M 18 30/50 30/50 Pioglitazone Unknown NA Unknown NA Positive Negative Myocardial Focal OU Focal Insulin
infarct
23 3/03(9/03) 9/03 52 M 4 20/20 40/20 Pioglitazone 5/03 NA 40 Unknown Negative Positive — Focal OU Focal Glucophage,
insulin,
metformin
24 (8/03) 9/03 69 M 5 20/20 20/20 Pioglitazone Unknown 9/03 Unknown 3 Negative Positive — Focal OU None Glyburide
25 1/02(7/03) 10/03 84 F 10 25/25 50/30 Rosiglitazone 6/03 7/03 Negative 11 Negative Positive Myocardial Focal OS None Glyburide
infarct,
nephropathy
26 12/01(7/02) 4/03 41 M 12 70/70 60/80 Pioglitazone 2001 NA Positive Negative Negative Positive Nephropathy, Focal OU Focal, Insulin
neuropathy, PRP
alcohol
abuse
27 (12/02) 12/02 64 F 23 50/50 50/50 Pioglitazone 6/01 NA Positive Negative Positive Positive Myocardial Focal OS None Insulin
infarct,
thyroid
cancer,
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES

neuropathy
●

28 (8/03) 9/03 46 M 16 20/20 20/20 Rosiglitazone “Longtime” 8/03 100 Negative Negative Positive — Focal OS Focal Glucophage
29 (5/03) 8/03 49 M 15 20/25 20/30 Pioglitazone 4/03 NA 10 Negative Negative Positive Peripheral Focal OS Focal Glucophage,
2006

neuropathy amaryl
●

30 (3/03) 8/03 55 M 16 40/100 50/60 Rosiglitazone 3/02 3/03 Positive Positive Negative Positive Nephropathy, Focal OU Focal Insulin
neuropathy
Focal leakage† — — Mean, — — Mean Mean — — — — — — — — — — —
61 visual visual
acuity, acuity,
VOLUME 26

20/35 20/36
●

Overall† — — Mean, — — Mean Mean — — — — — — — — — — —

61 visual visual
acuity, acuity,
20/50 20/62
NUMBER 5

*Visual acuity shown as 20/(number) OD/20/(number) OS.

†Mean visual acuity shown as 20/(number) and includes both eyes.
PRP, panretinal photocoagulation; F&P, focal and peripheral; NA, not available; HM, hand motion.
 Table 2. Characteristics of 11 Patients Followed for ⬎3 Months After Cessation of Glitazones
Initial Study Macular Macular
Examination Entry Present Weight Weight Edema Macular Edema
Date (Date Months Vision Vision Gain on Loss off Congestive (Focal, Laser Edema Response to
Glitazone Followed Age Years of (OD/ (OD/ Glitazone Start Stop Drug Drug Heart Diffuse, (Focal, Response Drug
Patient Stopped) off Drug (y) Sex Diabetes OS)* OS)* Drug(s) Date Date (lb) (lb) Failure Hypertension Both) PRP) to Laser Discontinuation

17 (8/03) 3 72 M 4 60/40 40/30 Rosiglitazone 3/03 8/03 15 28 Positive Positive Both OU Focal OD Improved Improved
Diffuse
14 (9/02) 8 65 F 11 60/50 50/50 Pioglitazone Unknown 2002 Positive 12 Negative Positive OU Focal OU Improved Improved
Diffuse Slow
10 5/01 (1/02) 21 61 M 10 60/400 60/HM Both 11/01 2/02 30 30 Negative Positive OU F&P OU decrease Slow decrease
Slow
11 4/00 (9/01) 25 60 F 40 60/70 300/100 Rosiglitazone 12/99 12/01 15 0 Positive Positive Both OU F&P OU decrease Slow decrease
Decreased
21 (8/02) 14 73 F 20 150/20 100/20 Pioglitazone 3/02 8/02 40 27 Negative Positive Focal OU Focal OU edema Slow decrease
12 (9/01) 30 52 F 16 40/25 150/40 Both 11/01 4/02 30 30 Negative Positive Both OU Focal OU No effect Slow decrease
DME WITH GLITAZONE USE

Diffuse Prior PRP,

●

13 (11/99) 3 46 M 19 50/80 25/40 Rosiglitazone 8/99 11/99 24 24 Positive Positive OU no focal NA Resolved
25 1/02 (7/03) 3 84 F 10 25/25 50/30 Rosiglitazone 6/03 7/03 Negative 12 Negative Positive Focal OS None NA Resolved
Increased
6 4/98 (5/03) 3 70 F 9 200/200 400/400 Rosiglitazone 2001 4/03 50 18 Negative Positive Both OU Focal OU edema No change
Increased
3 7/02 (5/03) 4 38 F 1 40/20 60/150 Pioglitazone 9/02 5/03 36 10 Negative Negative Both OU F&P OU edema Unknown
Edema
30 (3/03) 5 55 M 16 40/100 60/50 Rosiglitazone 3/02 3/03 Positive Positive Negative Positive Focal OU No worse No effect
RYAN JR ET AL

6 F,
Average† — 10.6 61.5 5M 19 20/60 20/85 — — — 30 19 — — — — — —

*Visual acuity shown as 20/(number) OD/20/(number) OS.

†Mean visual acuity shown as 20/(number) and includes both eyes.
PRP, panretinal photocoagulation; HM, hand motion; F&P, focal and peripheral; NA, not available.
565
566 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
cally significant macular edema and systemic fluid
retention during glitazone use and to review the course
of both weight loss and DME status in patients fol-
lowed for ⱖ3 months after cessation of glitazone use.
This retrospective study was not able to test whether
glitazone causes worsening of macular edema; rather
it was intended to explore whether an association
exists.
Patients and Methods
This study was a noncontrolled, nonrandomized,
retrospective chart review. Institutional review board
exemption was obtained for this study. Thirty patients
with type 2 diabetes were identified who were using
pioglitazone or rosiglitazone and had the characteris-
tics of identifiable lower extremity edema that in-
creased since starting the drugs and clinically signif-
icant macular edema. There were no controls. Patients
were not matched for age or sex. Patients were iden-
tified by a computer search of patient correspondence
for mention of “Actos” (pioglitazone) or “Avandia”
(rosiglitazone). Patients were included in the study if
they had the following characteristics: clinically sig-
nificant macular edema in at least one eye and lower
extremity edema thought to be associated with glita-
zone use. Patients who had glitazone use without
macular edema and glitazone use with no history of
peripheral edema were excluded.
The setting was the clinical practice of a single
specialty multiphysician retina-only referral group.
Clinically significant macular edema was documented
by clinical examination and review of the color pho-
tographs and fluorescein angiograms. Fluid retention
and peripheral edema were noted to be present or
absent. Response to laser treatment was assessed by
clinical examination. Response to cessation of glita-
zones was measured by reported weight loss, clinical
estimation of lower extremity edema, visual acuity
change, and change in macular edema. The use of
other glycemic control agents (insulin or oral) was
reported by the patients. Therapeutic ocular interven-
tions included focal laser treatment.
Therapeutic systemic interventions included cessa-
tion of glitazones, diuresis, and dialysis (one case).
Adverse effects were assessed using visual acuity
measurements and macular edema grading. Weight
gain or loss was reported by the patients. Clinical
outcome assessments were made by the authors. Pa-
tient follow-up was as clinically dictated. Visual acu-
ities were converted from Snellen readings to log-
MAR for calculations of averages and then
reconverted to Snellen readings.
A 2-week inventory of our clinic population to
● 2006 ● VOLUME 26 ● NUMBER 5
determine the total number of diabetic patients seen
versus the total number of patients taking glitazones
was performed. To calculate the number of patients
taking glitazones who were likely to be seen over a
9-month period, we used the following formula:
([number of glitazone-using patient visits per week/2
weeks] ⫻ [39 weeks/9 months]) ⫼ (average of 2 to 3
visits/patient). We then divided this number by the
number of glitazone-using patients seen in 2003 who
had macular edema, peripheral edema, and/or weight
gain.
Results
This study identified 30 patients seen in the period
November 1999 through September 2003 who had
clinically significant macular edema and peripheral
edema during glitazone use (Table 1). A subgroup of
11 patients was observed at least 3 months after ces-
sation of glitazones. Nineteen patients were male, and
11 were female. The mean age of the patients was 61
years (range, 38 – 84 years). All patients had type 2
diabetes (average duration since diagnosis, 8.3 years).
Hypertension was found in 23 patients, congestive
heart failure, in 6, and renal failure, in 1.
The first patient with the possible association of
glitazones, fluid retention, and macular edema was
seen in late 1999. Twelve patients were seen for the
first time during January to September 2003, while
another eight previously established patients were first
recognized to have peripheral edema in 2003. Seven-
teen patients took pioglitazone, 11 took rosiglitazone,
and 2 took both drugs at different times. Two patients
took glitazones as monotherapy, 12 took other oral
agents in addition to glitazones, 7 took insulin alone
with glitazones, 7 took insulin plus other oral agents in
addition to glitazones, and 2 had no information avail-
able. Fluid retention and peripheral edema were
present in 20 patients; they were categorized as severe
in 10 patients (Fig. 1). Weight gain was quantified in
13 patients (mean, 33 lb; range, 10 –100 lb) and was
present but not quantified in all but 4 others. Clinically
significant macular edema was bilateral in 24 patients
and was unilateral in 6. Twenty-six of 30 patients
(48/60 eyes) received macular laser photocoagulation
(ⱖ2 times in 22/60 eyes). Panretinal photocoagulation
was performed on 15 of 60 eyes. Mean visual acuity
overall at the first visit was 20/50, and at the last visit,
it was 20/62.
Macular edema was judged by the examining phy-
sician based on results of clinical evaluation and flu-
orescein angiography to be diffuse in at least 1 eye of
19 of 30 patients, with 17 of 30 patients having
bilateral diffuse edema. Because fluid overload from
 DME WITH GLITAZONE USE
Fig. 1. A 65-year-old man with a 40-lb weight gain and moderate
pitting edema while taking pioglitazone.
other causes often leads to diffuse macular edema, we
thought that fluid retention associated with glitazone
use might be provoking macular edema in these pa-
tients. Patients with diffuse edema tended to have the
most severe vision loss (Fig. 2). This group of patients
had mean initial visual acuity of 20/60 and mean final
visual acuity of 20/84, compared with 20/35 and 20/
36, respectively, for the patients with focal edema
only. The average weight gain on drug was 23 lb, and
the average weight loss off drug was 19 lb. Two
patients used glitazones as monotherapy, 7 took other
oral agents in addition to glitazones, 5 took insulin
alone with glitazones, 4 took insulin plus other oral
agents in addition to glitazones, and 1 had no infor-
mation available. Macular laser treatment was per-
formed on 31 eyes of 18 patients (⬎1 time in 16 eyes).
Of the 11 patients followed for ⬎3 months after
cessation of glitazones, 10 had weight loss after drug
cessation, and decreased lower extremity edema oc-
curred in all 11. Weight loss was generally quite rapid,
as was resolution of peripheral edema, and both were
mostly complete by 2 months. Weight gain on drug in
this group averaged 30 lb (range, 15–50 lb), and
weight loss averaged 19 lb (range, 0 –30 lb) (Table 2).
Of the 11 patients, 10 were hypertensive, 3 developed
renal failure while being followed, and 3 had conges-
● RYAN JR ET AL 567
tive heart failure. Reduction in macular edema oc-
curred in ⬍3 months in only 4 of 11 patients, but it
was ultimately seen in both eyes of 8 of 11 patients
over a 1- to 2-year period. The three patients who had
no decrease in macular edema had been off glitazones
for ⬍6 months. The average visual acuity at the initial
visit was 20/60, and at the final visit, it was 20/85.
Although most patients had a progressive decrease in
macular edema over time, visual improvement did not
necessarily follow. Three patients developed renal
failure and recurrent fluid retention months after drug
cessation. One patient began dialysis and had rapid
resolution of macular edema. The other two patients
developed recurrent worsening of macular edema after
having had a transient decrease after stopping glitazone.
Most patients had several macular laser treatments,
and some were believed by the examiners to have
edema that was quite refractory to treatment (Fig. 3).
Several patients had retinal pigment epithelium atro-
phy or gliosis from chronic macular edema contribut-
ing to visual loss. Three patients (four eyes) did have
resolution of diffuse macular edema with improved vi-
sion after cessation of glitazones without laser treatment.
We identified 12 patients with peripheral edema
and/or weight gain and macular edema who were seen
for the first time in the first 9 months of 2003. Eight
other established patients were first noted to have this
association during this period. To calculate an esti-
mate of the percentage of patients who were at risk for
macular edema in association with glitazone-related
fluid retention, we conducted an inventory of patients
with diabetes seen in our office over a 2-week period
and recorded the number of patients taking rosiglita-
zone or pioglitazone; 334 such patients were seen, and
61 were taking glitazones. Because 20 patients had
been identified with macular edema, weight gain, and
peripheral edema associated with glitazone use during
a 9-month period, we estimated the risk of this to be
3.4% to 5% for patients taking glitazones: ([61 glita-
zone-using patient visits/2 weeks] ⫻ [39 weeks/9
months]) ⫼ (average of 2 to 3 visits/patient/9 months)
⫽ 559 to 839 glitazone-using patients, divided by 20
at-risk patients ⫽ 3.4% to 5%.
Discussion
Fluid overload is reported to contribute to worsen-
ing of DME, and laser treatment can be relatively
ineffective until the fluid overload is adequately treat-
ed.7–9 Congestive heart failure and renal disease are
two common causes of this. In patients seen early in
this study, these two entities were first suspected as
the cause of macular edema associated with peripheral
edema, and consultation with the primary physician
568 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
was obtained. After congestive heart failure and renal
disease were ruled out, fluid retention due to glitazone
use was recognized as a possible cause after review of
product information.4,5 Once we became aware of this
possible association, we identified many more cases,
in part by paying more careful attention to lower
extremity edema. We think that the much larger num-
ber of cases recognized in 2003 (20/30 patients) rep-
resents more likely better surveillance due to dissem-
ination of information regarding glitazones and fluid
retention than a greater frequency of occurrence.
Worsening of congestive heart failure, as well as
pulmonary edema, is described in some patients using
glitazones.6,11 Functional improvement (increased ex-
ercise tolerance) is seen when the fluid overload is
corrected, shortly after glitazones are discontinued.
Fluid overload rather than a direct toxic effect on the
heart is implicated as the cause.12 Some evidence
suggests that glitazone-associated fluid retention is
related to increased endothelial cell permeability.3
Glitazones also increase plasma concentrations of vas-
cular endothelial growth factor,13 which could also
lead to increased vascular permeability. Fluid accu-
mulation in the retina might be caused by several
mechanisms leading to both increase in plasma vol-
ume and increased permeability.
● 2006 ● VOLUME 26 ● NUMBER 5
Fig. 2. A 61-year-old man
taking pioglitazone who had a
30-lb weight gain and pro-
nounced lower extremity
edema. A, Color photograph
of right eye; vision 20/200
and diffuse edema. B, Color
photograph of left eye; vision
2/200 and severe diffuse
edema. C, Transient-phase
fluorescein angiogram of
right eye showing telangiecta-
sia and early leakage. D,
Late-phase fluorescein angio-
gram of left eye showing se-
vere diffuse leakage.
For patients in this study who were followed for ⱖ3
months after discontinuation of glitazones, two prom-
inent patterns developed. First, weight loss and reduc-
tion or resolution of peripheral edema occurred rap-
idly in nearly all cases. Several patients said that they
felt better and had less difficulty breathing. In our
patients, the resolution of macular edema was often
very slow after drug stoppage and reversal of fluid
retention, for reasons unknown. In 3 of 11 patients (4
eyes), macular edema also resolved substantially and
promptly, without laser treatment, and with visual
improvement. In the other cases, macular edema re-
solved slowly, requiring laser treatments. In some of
these cases, visual acuity remained poor due to mac-
ular scarring from chronic edema. It is unclear why
the peripheral edema cleared rapidly and the macular
edema cleared slowly in so many cases. It is possible
that some diabetic patients who are prone to the vas-
cular leakage that would cause diffuse macular edema
are likewise prone to peripheral edema associated with
glitazone use and that discontinuation of the drug
lessens the fluid overload but does not affect the
preexisting tendency for macular leakage.
Patients with worsening of DME associated with
fluid overload have diffuse edema.7 In the study group
as a whole, 19 (63%) of 30 patients had diffuse
 DME WITH GLITAZONE USE ● RYAN JR ET AL 569

Fig. 3. Progression of dia-

betic macular edema in a 60-
year-old woman taking ros-
iglitazone. A, October 19,
2000; vision 20/60. B, Febru-
ary 8, 2001, vision 20/80. C,
May 23, 2001; vision 20/100.
D, September 20, 2001; vi-
sion 20/200 and rosiglitazone
discontinued. E, April 18,
2002; vision 20/300. F, Au-
gust 7, 2003; edema resolved
with subfoveal glial scarring
and vision 20/150.

macular edema in at least one eye. We think that the
fluid overload due to glitazone use likely contributed
to this. These patients had poorer vision than the
group as a whole (initial visual acuity, 20/60; final
visual acuity, 20/84; compared with 20/35 and 20/36,
respectively). In the three eyes of two patients with
rapid resolution of both peripheral and macular edema
once off the drug, we thought that there was a direct
cause-and-effect relationship. Because our estimate of
the risk of macular and peripheral edema associated
with glitazone use is 2.4% to 3.6% and 63% of these
patients had diffuse edema, we think that the inci-
dence of worsening of macular edema due to glitazone
use is ⬇1.5% to 2.6%. This estimate may be low due
to missed cases, but it could be high due to the referral
nature of our practice and would be less accurate than
an estimate obtained prospectively.
There are a number of problems inherent in this
study due to its retrospective nature. A retrospective
study cannot prove whether glitazone use–related fluid
retention worsens DME. This study can only survey
the incidence of glitazone-related fluid retention and
macular edema. A prospective study would be nec-
essary to test causality, and optical coherence to-
mography could be used to quantify macular edema.
This study can only suggest a possible association
between fluid retention from glitazone use and
worsening of DME.
570 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
We did not think it was ethical to rechallenge pa-
tients with glitazones to see if peripheral and macular
edema worsened again. We did not review in detail
the charts of glitazone-using patients who had no
reported weight gain or macular edema. Because these
subjects were referral patients, we did not have the
opportunity to check visual acuity or evaluate for
macular edema before the start of glitazone adminis-
tration; therefore, we are presuming that they devel-
oped worsening of edema after starting the drugs.
Other difficulties include dependence on patient
reporting for weight gain and loss and drug start and
stop dates. In most cases, however, patients were very
certain about weight gain and loss recollection. In
addition, many of the study patients had coexisting
medical problems such as congestive heart failure,
which can worsen fluid retention. In addition, most of
the patients underwent laser photocoagulation as part
of their treatment; therefore, the relative contribution
of drug cessation versus laser treatment on fluid res-
olution is hard to discern.
Glitazones are useful in the management of diabetes
mellitus and are commonly used in current practice,
with an estimated $5 billion in sales for 2003.14 Ap-
proximately 5% to 10% of patients taking these drugs
develop fluid retention, and some of these patients are
at risk for the fluid retention aggravating DME. It is
recognized that fluid retention complicates the man-
agement of DME. It is our impression that the fluid
retention associated with glitazone use can aggravate
DME in susceptible patients. It is currently recom-
mended that patients with congestive heart failure
exercise caution in using glitazones.15 We think that
these drugs should be used with caution in patients
with DME. Medical doctors should understand that
fluid retention secondary to glitazone use could aggra-
vate macular edema. Ophthalmologists should under-
stand that patients with macular edema should be
queried regarding glitazone use and examined for
lower extremity edema.
● 2006 ● VOLUME 26 ● NUMBER 5
Key words: diabetes mellitus, diabetic retinopathy,
fluid retention, glitazones, macular edema, pioglita-
zone, rosiglitazone, thiazolidinediones.
References
1. Stumvoll M, Hfaring HU. Glitazones: clinical effects and
molecular mechanisms. Ann Med 2002;34:217–224.
2. Graham DJ, Green L, Senior JR, Nourjah P. Troglitazone-
induced liver failure: a case study. Am J Med 2003;114:299–
306.
3. Niemeyer NV, Janney LM. Thiazolidinedione-induced
edema. Pharmacotherapy 2002;22:924–929.
4. Avandia Package Insert (see adverse reactions section).
http://us.gsk.com/products/assets/us_avandia.pdf.
5. Actos Package Insert (see adverse reactions section). http://
www.fda.gov/medwatch/SAFETY/2003/03Jan_labels/
Actos_PI.pdf.
6. Wang CH, Weisel RD, Liu PP, Fedak PW, Verma S. Glitazones
and heart failure: critical appraisal for the clinician. Circulation
2003;107:1350.
7. Bresnick GH. Diabetic macular edema. A review. Ophthal-
mology 1986;93:989–997.
8. Tokuyama T, Ikeda T, Sato K. Effects of haemodialysis on
diabetic macular leakage. Br J Ophthalmol 2000;84:1397–
1400.
9. Perkovich BT, Meyers SM. Systemic factors affecting dia-
betic macular edema. Am J Ophthalmol 1988;105:211–212.
10. Coluciello M. Vision loss due to macular edema induced by
rosiglitazone treatment of diabetes mellitus. Arch Ophthal-
mol 2005;123:1273–1275.
11. Kermani A, Garg A. Thiazolidinedione-associated congestive
heart failure and pulmonary edema. Mayo Clin Proc 2003;
78:1088–1091.
12. Kennedy F. Do thiazolidinediones cause congestive heart
failure? Mayo Clin Proc 2003;78:1076–1077.
13. Emoto M, Anno T, Sato Y, et al. Troglitazone treatment
increases plasma vascular endothelial growth factor in dia-
betic patients and its mRNA in 3T3-L1 adipocytes. Diabetes
2001;50:1166–1170.
14. Glitazones and the Oral Diabetes Market. http://www.ims-
global.com/insight/news_story/news_story_990723.htm. Ac-
cessed 2 February 2003.
15. FDA Drug Alert—Actos and Avandia and CHF, April 2002.
http://www.fda.gov/medwatch/SAFETY/2002/
safety02.htm#thiazo. Accessed 5 June 2003.