The Effect of Phenylephrine
Henry F. Edelhauser, PhD; James E. Hine, MD; Harlan Pederson, MS; Diane L. Van Horn, PhD; Richard O. Schultz, MD
\s=b\ Rabbit corneas were treated with
three drops of phenylephrine hydrochlo-
ride with the epithelium intact or denuded.
Corneal thickness was measured before
and after drug treatment, and at various
times after treatment the corneas were
fixed for scanning and transmission elec-
tron microscopic observation. The results
of this study show that phenylephrine
caused a dramatic increase in corneal
thickness (drug-induced edema) and cel-
lular vacuolation within the keratocytes
and endothelial cells in the corneas with-
out the epithelium. Corneal thickness did
not change and the ultrastructural
changes were minimal following drug
application in those corneas with the
epithelium intact. Results of this study
also suggest that phenylephrine has a
cytotoxic effect on the corneal endothe-
lium and keratocytes when used in
corneas where the epithelium has been
removed. In corneas with intact epithe-
lium, the damage was less severe and
limited to the epithelium.
(Arch Ophthalmol 97:937-947, 1979)
"Dhenylephrine hydrochloride is
widely used to obtain mydriasis for
routine funduscopic examination, re¬
fraction, and during retinal surgery.
With its use there have been reports
describing acute systemic hyperten¬
sion,1-' and two studies have been
published comparing dose vs mydriat-
Accepted for publication July 3, 1978.
From the Departments of Physiology (Drs
Edelhauser and Van Horn) and Ophthalmology
(Drs Edelhauser, Hine, Van Horn, and Schultz),
The Medical College of Wisconsin, Milwaukee,
and Research Service, Veterans Administration
Center (Mr Pederson and Dr Van Horn), Wood,
Wis.
Reprint requests to Medical College of Wiscon-
sin, PO Box 26509, Milwaukee, WI 53226 (Dr
Edelhauser).
Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015
on the Cornea
ic effect. ,·ß Recently, in the national
drug registry7 two documented cases
have been listed of epithelial edema
and sloughing of the complete epithe¬
lium with the use of 10% phenyleph¬
rine hydrochloride.
Many of the newer techniques in
retina and vitreous surgery demand
maximal pupillary dilation for pro¬
longed periods of time. Repeated topi¬
cal application of mydriatics and
cycloplegics are often required to
insure continued mydriasis. Frequent¬
ly during the course of the surgery,
the corneal epithelium becomes ede¬
matous and eventually denuded by the
surgeon, yet the mydriatic drops are
continued. Clouding of the corneal
stroma is often observed during the
course of the operation. Machemer"
has noted that phenylephrine hy¬
drochloride should not be used for
vitrectomy because it causes endothe¬
lial haze that will impair the surgeon's
visibility of the fundus.
Therefore, the purpose of this study
was to investigate the effect of
phenylephrine on the cornea and to
describe the specific effect this agent
has on the corneal epithelium, stroma,
and endothelium.
MATERIALS AND METHODS
Commercial 2.5% and 10% phenylephrine
hydrochloride and 2.7% phenylephrine hy¬
drochloride prepared without preservative
were administered topically (one drop
every five minutes for three applications)
to the corneas of three groups of albino
rabbits. Each group contained 14 rabbits.
Prior to the application of phenylephrine,
one drop of 0.4% benoxinate hydrochloride
was placed in both eyes of each rabbit. The
corneal epithelium was then removed from
one eye and left intact in the other eye. For
controls the epithelium was scraped and no
additional drug or vehicle added (N 4) or
=
one drop of 0.9% sodium chloride was
administered topically every five minutes
for three applications (N 5).
=
Corneal thickness was measured with a
pachometer, modified according to Mishi-
ma and Hedbys," before topical application
of phenylephrine, after scraping the epi¬
thelium, and at 15, 30, 60, 120, 240 minutes,
and 24 hours. At 15 minutes, 1, 48, and 72
hours, the animals were killed and the
corneas excised for scanning and transmis¬
sion electron microscopy (SEM and TEM).
In another series of five rabbits, 10%
phenylephrine hydrochloride was adminis¬
tered to the scraped cornea (one drop every
hour for eight applications) for three
consecutive days. Corneal thickness was
measured and the animals killed one hour
after the final drop of phenylephrine.
The corneas were fixed for electron
microscopy in 2.7% glutaraldehyde buf¬
fered with phosphate (pH 7.1, 330 mOsm)
for at least eight hours at 4 °C. The corneas
were then cut in half and postfixed in 2%
osmium tetroxide for two hours. Small
pieces of one half of each cornea were
embedded in a low viscosity epoxy medium
for TEM. The other half of each cornea was
prepared for SEM of the endothelium
according to a modified method of Cleve¬
land and Schneider.1" They were pene¬
trated with the low viscosity resin, poly¬
merized overnight at 37 °C and for 48 hours
at 60°C, glued to stubs, sputtered-coated
with gold palladium, and viewed with a
scanning electron microscope.
RESULTS
Topical application of 2.5% or 10%
phenylephrine hydrochloride in eyes
with intact corneal epithelium did not
result in an increase in corneal thick¬
ness (Fig 1). Corneal ultrastructural
changes were present, however, and
were found to be most prominent in
the epithelium and anterior stroma.
At 15 minutes after application, TEM
demonstrated sloughing of the super¬
ficial, plate-like, squamous epithelial
cell layers. The remainder of the

1 .—Effect of topical phenylephrine applications on corneal thickness (in
Fig
Each point represents mean and brackets SEM. EPI indicates epithelium; —,
not scraped.
central and basal epithelial cells were
intact and displayed normal ultra-
structure (Fig 2, top). At one hour the
squamous cells were completely de¬
nuded, and many cytoplasmic vacuoles
were present within the central wing
cells (Fig 2, center). Basal epithelial
cells were elongated and narrow with
numerous cytoplasmic vacuoles that
tended to distort the nuclei (Fig 2,
bottom). Keratocytes located in the
anterior stroma were frequently
swollen and contained large cytoplas¬
mic vacuoles (Fig 3), while a majority
of the keratocytes seen in the posteri¬
or stroma were normal in appearance
with only a few showing smaller
vacuoles. Scanning electron micros¬
copy of the endothelium showed excel¬
lent preservation of the normal mosa¬
ic-like cellular pattern (Fig 4, top) and
TEM demonstrated normal ultra-
structure except for the presence of
occasional cytoplasmic vacuoles (Fig 4,
bottom).
Corneas that had only the epithe¬
lium scraped and those that had both
the epithelium scraped and the 0.9%
sodium chloride vehicle applied topi¬
cally demonstrated a slow but pro¬
gressive swelling rate (0.063 ± 0.014
mm/hr) at one hour and 0.090 mm/hr
at two hours. The corneal thickness
had increased to 0.16 ± 0.02 mm at
four hours and 0.2 mm after 24 hours
(Fig 1). Reepithelialization brought
about a progressive decrease in cor¬
neal thickness with a return to normal
in five days. Electron microscopy of
the keratocytes and endothelial cells
from corneas in this group showed
Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015
millimeters).
scraped; + ,
normal ultrastructure at both 15
minutes and one hour (Fig 5).
Corneas that were treated with 2.5%
or 10% phenylephrine hydrochloride
with the epithelium scraped demon¬
strated a dramatic increase in thick¬
ness at 15 minutes, 30 minutes, 1 hour,
and 2 hours, which was significantly
different from the denuded corneas
not receiving the drug (P < .02) (Fig
1). Although the 10% phenylephrine
hydrochloride caused a greater in¬
crease in corneal thickness, it was not
significantly greater than the 2.5%
solution.
Corneal ultrastructural changes
were associated with the sharp in¬
crease in thickness. At 15 minutes and
one hour, the stromal keratocytes
were grossly enlarged and displayed
numerous cytoplasmic vacuoles, dila¬
tion of the rough endoplasmic reticu-
lum, and distortion of the nuclei (Fig
6). Scanning electron microscopy dis¬
closed balled-up endothelial cells and
the breakdown of many junctional
complexes, resulting in a "cobble¬
stone" appearance of the endothelium.
Pits, representing areas where cyto¬
plasmic bridges have collapsed into
vacuoles just beneath the surface,
were prominent (Fig 7, top). Trans¬
mission electron microscopy of the
endothelial cells from the same cornea
displayed both intracellular and inter¬
cellular vacuoles, distortion of the
nuclei, disrupted junctions, and very
irregular outer plasma membranes
(Fig 7, bottom). At 48 hours, many
polymorphonuclear leukocytes were
found in close association with kerato-
cytes that had been damaged by the
extreme cytoplasmic vacuolization.
Multivesicular vesicles and residual
bodies were evident in many of these
injured cells (Fig 8, top). Other kerato¬
cytes were found to be completely
disrupted and were apparently being
phagocytosed by WBCs (Fig 8, bot¬
tom). Scanning electron microscopy of
the corneas from the same group
showed that the endothelium had
returned to a more normal mosaic-like
pattern when compared with the one-
hour postapplication group. The cells,
however, remained slightly swollen,
and numerous apical flaps could be
seen at the cellular junctions (Fig 9,
top). Transmission electron micros¬
copy showed that cytoplasmic vacuoles
were no longer present within the
endothelial cells and that the junction¬
al complexes between cells were
restored. However, residual effects,
such as clarification of the basal cyto¬
plasm and a deposition of electron-
dense material in posterior Desce¬
met's membrane, still remained (Fig
9, bottom).
Scraped corneas that received a
topical application of 10% phenyleph¬
rine hydrochloride of one drop per
hour for an eight-hour period for three
days showed a dramatic increase
above pretreatment (0.5 mm) in
corneal thickness of 0.25 mm and were
edematous on slit-lamp examination.
Keratocytes from corneas in this
group appeared completely necrotic
and contained no recognizable orga-
nelles when examined with TEM (Fig
10). Endothelial SEM showed that the
cells had swollen and disrupted the
normal hexagonal pattern. Several
blebs, which proved to be WBCs under
TEM, could be seen in close associa¬
tion with the microvilli that project
from the posterior plasma membranes
of the endothelial cells (Fig 11, bottom
left). Endothelial TEM demonstrated
clarification of the basal cytoplasm
and a deposition of electron-dense
material in posterior Descemet's
membrane. White blood cells were
found entwined with microvilli pro¬
jecting from the endothelial cell
surfaces (Fig 11, top right). Other
areas from the same cornea displayed
polymorphonuclear leukocytes lying
within the endothelial cell layer (Fig
11, bottom right).
COMMENT
The results of this study show that,
if the epithelium is intact, there is no
increase in corneal thickness from any
of the concentrations of phenyleph¬
rine. There is, however, epithelial
damage to the outer cell surface that
 Fig 2—Transmission electron
microscopy. Top, Superficial
epithelium of intact rabbit cor¬
nea fixed 15 minutes after topi¬
cal application of phenylephrine
shows sloughing of superficial
plate-like squamous cell layers.
Center, Superficial epithelium of
intact rabbit cornea fixed one
hour after topical application of
phenylephrine shows denuded
squamous cells and cytoplasmic
vacuoles present with central
wing cells. Bottom, Basal epithe¬
lial cells from same cornea as
Fig 2, center, display very elon¬
gated narrow cells with distorted
nuclei. Numerous variable-sized
vacuoles fill cytoplasm (original
magnification 6,300).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 Fig 3.—Transmission electron
microscopy of anterior stroma of
intact rabbit cornea fixed one
hour after topical application
of phenylephrine. Many kera¬
tocytes are swollen and
contain large cytoplasmic
vacuoles (original magnification
X11,800).

Fig 4—Top, Scanning electron

microscopy of endothelium of
intact rabbit cornea fixed one
hour after topical application of
phenylephrine shows preserva¬
tion of normal mosaic-like cellu¬
lar pattern (original magnifica¬
tion 1,300). Bottom, Transmis¬
sion electron microscopy of en¬
dothelial cells from same cornea
show that normal ultrastructure
is maintained except for pres¬
ence of occasional cytoplasmic
vacuoles (original magnification
6,800).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 Fig 5—Top, Transmission elec¬
tron microscopy of stroma of
rabbit cornea fixed one hour
after epithelium was scraped. No
drug was applied topically. Kera¬
tocytes between collagen lamel¬
lae appear normal (original mag¬
nification x 10,400). Center,
Scanning electron microscopy
of endothelium from same cor¬
nea shows normal mosaic-like
cellular pattern (original magnifi¬
cation 1,300). Bottom, Trans¬
mission electron microscopy of
endothelial cells from same cor¬
nea shows normal ultrastruc¬
ture (original magnification
19,800).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 Fig 6.—Transmission electron
microscopy of stroma of scraped
rabbit cornea fixed after topical
application of phenylephrine.
Keratocytes are grossly en¬
larged and display numerous
cytoplasmic vacuoles, dilation of
rough endoplasmic reticulum,
and distortion of nuclei (original
magnification 11,800).

Fig 7.—Top, Scanning electron

microscopy of endothelium from
same cornea as in Fig 6. Cells
are balled up and pulled apart.
Pits are present in posterior
plasma membrane, and many of
junctional complexes appear to
have collapsed (original magnifi¬
cation 1,300). Bottom, Trans¬
mission electron microscopy of
endothelial cells from same cor¬
nea shows both intracellular and
intercellular vacuoles, distortion
of nuclei, straight cellular junc¬
tions, and very irregular outer
plasma membranes (original
magnification 6,300).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 Fig 8.—Transmission electron
microscopy. Top, Stroma of
scraped rabbit cornea fixed 48
hours after topical application of
phenylephrine. Polymorphonu-
clear leukocyte is seen in close
association with damaged kera-
tocyte containing multivesic-
ular vesicles and residual
bodies (original magnification
x 10,500) Bottom, Stroma from
same cornea. A completely dis¬
rupted keratocyte is being pha-
gocytosed by a polymorphonu-
clear leukocyte (original magnifi¬
cation 20,600).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015


may be related to the epithelial
vacuolization that results from the use
of phenylephrine. Sloughing of the
corneal epithelium has been noted
with the use of this drug.7 This may be
secondary to the vacuole formation
noted in this study leading to cell lysis.
Another possible cause, as noted by
Pfister and Burstein," is the epithelial
damage caused by benzalkonium chlo¬
ride, which is used as a preservative in
commercial preparations of this
drug.
Of particular clinical importance is
the drug-induced edema that occurs in
corneas with the epithelium removed.
This edema appears to be the result of
the drug penetrating through the
cornea. Since it does not occur with
the epithelium intact, one can con¬
clude that the epithelium acts as a
Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015
barrier to the drug. This is consistent
with past studies where removal of
the epithelium will increase the drug
penetration by three- to fivefold for
steroids14 and antivirale,'4 respective-
ly.
Since corneal edema and vacuole
formation occurs regardless of drug
concentration, these changes would
not appear to be related to the osmo-
larity of the drug (1,062 mOsm for
10%, 375 mOsm for 2.5%, and 300
mOsm for 2.7%). The drug vehicle can
also be ruled out as a factor because
the same degree of corneal edema and
vacuolization occurs with 2.7% phenyl¬
ephrine hydrochloride that is prepared
to isotonicity in distilled water. There¬
fore, these corneal effects would
appear to be related to the phenyleph¬
rine itself or possibly a breakdown
Fig 9.—Top, Scanning electron
microscopy of endothelium from
same cornea as in Fig 8. Cells
have returned to a more normal
mosaic-like pattern but remain
slightly swollen and display nu¬
merous apical flaps at cellular
junction (original magnification
1,300). Bottom, Transmission
electron microscopy of endothe¬
lial cells from same cornea show
an absence of cytoplasmic vac¬
uoles and restoration of junc¬
tional complexes. Note deposi¬
tion of electron-dense material
in posterior Descemet's mem¬
brane (original magnification
7,400).
product or a metabolite of phenyleph¬
rine. It has been reported that
buffered phenylephrine does contain
two breakdown products: 1,2,3,4-tetra-
hydro-4,6-dihydroxy-2-methyliso-
quinoline and a 4,8-dihydroxy analo¬
gue,'4 which also could act osmotically
within the cornea to increase its thick¬
ness.
The vacuolization that occurs within
the keratocytes and endothelium is
most likely the result of the phenyl¬
ephrine, its breakdown product or
metabolite causing an increase in
intracellular osmolarity. As the water
is osmotically drawn into the cells, the
water vacuoles form. A similar obser¬
vation has been reported for sucrose
in the proximal tubular cells of the
rat,45 and in the ciliary epithelium
following intravenous administration
 Fig 10.—Transmission electron
microscopy of stroma of scraped
rabbit cornea fixed after topical
application of 10% phenyleph¬
rine, one drop per hour for an
eight-hour period for three days.
Keratocytes are completely ne-
crotic and display no recogniza¬
ble organelles (original magnifi¬
cation x 7,900).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 Fig 11.—Bottom left, Scanning
electron microscopy of endothe¬
lium from same cornea as in Fig
10. Cells are swollen and show
greatly increased number of sur¬
face microvilli. Several polymor-
phonuclear leukocytes are seen
lying on endothelial cell mem¬
branes (original magnification
1,300). Top right, Transmis¬
sion electron microscopy of en¬
dothelial cells from same cornea
display clarification of basal cy¬
toplasm and narrow band of
electron-dense material in pos¬
terior Descemet's membrane. A
WBC is entwined with numerous
microvilli on endothelial cell sur¬
face (original magnification
10,000). Bottom right, Trans¬
mission electron microscopy
from same cornea shows two
polymorphonuclear leukocytes
lying within endothelial cell
layer (original magnification
x 11,400).

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015

 of urea."4 Once equilibrium has been
reached between the drug and cornea,
the vacuolization will decrease, as
illustrated by the three-day phenyl¬
ephrine treatment (Fig 10). Recent
studies in our laboratory have shown,
however, that these corneas do not go
unscathed. Phenylephrine can cause a
decrease in corneal endothelial glucose
metabolism,'7 and, as indicated in Fig
10, the endothelium has been induced
to produce an electron-dense material
between the endothelium and Desce¬
met's membrane.
These studies provide evidence that
phenylephrine can have a cytotoxic
effect on the corneal endothelium and
keratocytes when used in corneas
where the epithelium has been re¬
moved. In corneas with the epithelium
intact, the damage was less severe and
limited to the epithelium. These obser¬
vations are not species specific, for the
same results occur with 10% phenyl¬
ephrine in cats and isolated perfused
human cornea.
Of major clinical concern is the
prolonged use of 10% phenylephrine
for mydriasis following retinal reat-
tachment surgery and its use during
vitrectomy following removal of the
corneal epithelium, which could lead to
corneal decompensation. The corneas
of diabetic patients may possess
epithelial ulcération as well as a slow
rate of reepithelialization.18 Use of
phenylephrine in these corneas could
be particularly hazardous.
Recent studies have shown very
small amounts of cycloplegics in an
ointment vehicle to provide adequate
mydriasis and cycloplegia.1" Perhaps
use of this vehicle would allow safer
delivery of phenylephrine, or phenyl¬
ephrine could be delivered as a precur¬
sor as has been done with dipive¬
frin.-"
This investigation was supported in part by
National Eye Institute research grants EY-
00933, EY-o"l436, Ophthalmic Research Center
grant I-P30-EY01931, and by the Medical
Research Service of the Veterans Administra¬
tion.
Name and
Nonproprietary
Trademarks of Drug
Phenylephrine hydrochloride—Almefrin,
Consdrin, Derizene, Efricel, Neo-Syn-
ephrine Hydrochloride, Nose Drops-
Spray.

References

1. Lansche RK: Systemic reactions to topical
epinephrine and phenylephrine. Am J Ophthal-
mol 61:95-98, 1966.
2. McReynolds WV, Havener WH, Henderson
JW: Hazards of the use of sympathomimetic
drugs in ophthalmology. Arch Ophthalmol 56:176\x=req-\
179, 1956.
3. Solosko D, Smith RB: Hypertension follow-
ing 10% phenylephrine ophthalmic. Anesthesiolo-
gy 36:187-189, 1972.
4. Wilensky JT, Woodward HJ: Acute systemic
hypertension after conjunctival installation of
phenylephrine hydrochloride. Am J Ophthalmol
76:156-157, 1973.
5. Haddad NJ, Mayer NJ, Riley FC: Mydriatic
effect of phenylephrine hydrochloride. Am J
Ophthalmol 70:729-733, 1970.
6. Brian Smith R, Read S, Oczypok PM:
Mydriatic effect of phenylephrine. Eye Ear Nose
Throat Monthly 55:133-134, 1976.
7. Case reports, in National Drug Registry of
Drug-Induced Ocular Side Effects. Little Rock,
University of Arkansas for Medical Sciences,
1978.
8. Machemer R: Vitrectomy\p=m-\A Pars Plana
Approach, Machemer R, Benson WE (eds). New
York, Grune & Stratton, 1975, p 51.
9. Mishima S, Hedbys BO: Measurement of
corneal thickness with the Haag-Streit pachom-
eter. Arch Ophthalmol 80:710-713, 1968.
10. Cleveland PH, Schneider CW: A simple
method of preserving tissue for scanning elec-
tron microscopy. Vision Res 9:1401-1402, 1969.
11. Pfister RR, Burstein N: Effects of ophthal-
mic drugs, vehicles and preservatives on corneal
epithelial surface: Scanning electron microscope
study. Invest Ophthalmol 15:246-259, 1976.
12. Hull DS, Hine JE, Edelhauser HF, et al:
Permeability of the isolated rabbit cornea to
corticosteroids. Invest Ophthalmol 13:457-459,
1974.
13. O'Brien WJ, Edelhauser HF: The corneal
penetration of ionfluorothymidine, adenine ara-
binoside, and idoxuridine: A comparative study.
Invest Ophthalmol 16:1093-1103, 1977.
14. Millard BJ, Priaulx DJ, Shatton E: The
stability of aqueous solutions of phenylephrine at
elevated temperatures: Identification of the
decomposition products. J Pharm Pharmacol
25(suppl):24-31, 1973.
15. Janigan DT, Santamaria A: A histochemi-
cal study of swelling and vacuolation of proximal
tubular cells in sucrose nephrosis in the rat. Am J
Pathol 39:175-185, 1961.
16. Shabo AL, Maxwell DS, Kreiger AE: Struc-
tural alterations in the ciliary process and the
blood-aqueous barrier of the monkey after
systemic urea injections. Am J Ophthalmol
81:162-172, 1976.
17. Geroski DH, Edelhauser HF: Corneal
metabolism: Endothelial response to epithelial
scraping, topical anesthesia, and phenylephrine.
Invest Ophthalmol 17(suppl):211, 1978.
18. Hyndiuk RA, Kazarian EL, Seidman S:
Neurotropic corneal ulcers in diabetes mellitus.
Arch Ophthalmol 95:2193-2196, 1977.
19. Cable MK, Hendrickson RO, Calvin H:
Evaluation of drugs in ointment for mydriasis
and cycloplegia. Arch Ophthalmol 96:84-86,
1978.
20. Kaback MB, Podos SM, Harbin TS Jr, et al:
The effects of dipivalyl epinephrine on the eye.
Am J Ophthalmol 81:768-772, 1976.

CORRECTION

Error in Letter to the Editor.\p=m-\Inthe letter to the editor titled "Possible Pineal-
Suprachiasmatic Clock Regulation of Development and Life Span," published in the
February Archives (97:359, 1979), the second sentence in the second paragraph should
have read, "Blind persons younger than age 65 with retrolental fibroplasia were found to
have a significantly better ten-year survival rate (P < .05) than other blind persons
younger than age 65."

Downloaded From: http://archopht.jamanetwork.com/ by a University of California - San Diego User on 12/22/2015