Full text of DOI:10.1007/s004170100285

Graefe’s Arch Clin Exp Ophthalmol
(2001) 239:367–372
C L I N I C A L I N V E S T I G AT I O N
DOI 10.1007/s004170100285
Thomas Reinhard
Alexander Reis
Daniel Böhringer
Michael Malinowski
Adina Voiculescu
Peter Heering
Systemic mycophenolate mofetil
in comparison with systemic cyclosporin A
in high-risk keratoplasty patients:
3 years’ results of a randomized prospective
clinical trial
Erhard Godehardt
Rainer Sundmacher
Abstract Background: With the use grafts were rejection-free in the
Received: 25 October 2000
Revised: 14 February 2001
Accepted: 28 February 2001
Published online: 16 May 2001
© Springer-Verlag 2001
of systemic cyclosporin A (CsA),
graft prognosis after high-risk pene-
trating keratoplasty has improved
considerably. However, the applica-
MMF group and 73% in the CsA
group (Kaplan Meier P=0.46, log-
rank test). Eight endothelial immune
reactions were observed in the
tion of CsA is limited owing to a va- MMF group (three under systemic
riety of severe side effects. In this
prospectively randomized study
mycophenolate mofetil (MMF), a
safe and efficient immunosuppres-
immunosuppression/five thereafter;
six mild/two severe) and five in the
CsA group (three under systemic
immunosuppression/two there-
sive medication after renal transplan- after; three mild/two severe). Side
tation, was compared with CsA after effects occurred in six patients
high-risk penetrating keratoplasty.
under MMF and 11 under CsA.
Methods: Twenty-nine high-risk ker- Conclusions: Concerning efficacy,
atoplasty patients were treated with
MMF 2× 1 g daily; another 27 pa-
no statistically significant difference
between systemic MMF and system-
Presented as free paper during the annual
convention of the German Ophthalmo-
logical Society (DOG) in Berlin,
21-24 September 2000
tients received CsA, aiming at blood ic CsA administered for 6 months af-
trough levels of 120–150 ng/ml. Sys- ter high-risk penetrating keratoplasty
temic immunosuppression was
scheduled for 6 months. In both
groups oral corticosteroids (fluo-
cortolone 1 mg/kg) were adminis-
tered for 3 weeks postoperatively.
Results: During the first year after
operation, no graft failure was re-
corded. Two years postoperatively
could be shown. Systemic MMF was
proven to be at least as safe as CsA.
The main mechanism in improving
graft survival is a shift from severe
to milder endothelial immune reac-
tions, as already demonstrated for
CsA. Thus, MMF may become an
alternative to CsA for immunosup-
T. Reinhard ( ) · A. Reis · D. Böhringer
✉
M. Malinowski · R. Sundmacher
Eye Hospital, Heinrich Heine University,
40225 Düsseldorf, Germany
e-mail: thomasreinhard@uni-duesseldorf.de
Tel.: +49-211-8118795
Fax: +49-211-8118796
A. Voiculescu · P. Heering
92%/82% and 3 years postoperative- pression after penetrating high-risk
Clinic for Nephrology and Rheumatology,
Heinrich Heine University, Düsseldorf,
Germany
ly 74%/69% of grafts were clear in
the MMF and CsA group, respec-
tively (Kaplan Meier P=0.33, log-
rank test). In total, two graft failures
were recorded in the MMF group
and four in the CsA group. Three
years postoperatively 53% of the
keratoplasty. About 2 years postop-
eratively, pharmacologically induced
relative immunological tolerance
slowly decreases. Therefore, long-
term administration of systemic
MMF should be evaluated in further
studies.
E. Godehardt
Department of
Biostatistics/Cardiovascular Surgery,
Heinrich Heine University, Düsseldorf,
Germany
[7, 8, 14]. However, side effects such as nephrotoxicity,
hepatotoxicity and arterial hypertension occur in about
Introduction
Short-term administration of systemic cyclosporin A 10% of patients treated with CsA as systemic mono-
(CsA) has led to considerable improvement in graft immunomodulation for 6–12 months postoperatively
prognosis for high-risk penetrating keratoplasty patients [14]. In order to minimize the risk of side effects, CsA

368
cause of recurrence of the underlying dystrophies. In all six of
these patients superficial keratectomy had been performed had at
least once prior to rekeratoplasty, and in all six patients repeat ker-
atoplasty took the form of limbokeratoplasty. Steroid glaucoma in
combination with bullous keratopathy was regarded a high-risk
situation since no topical steroids can be applied postoperatively –
a steep increase in the incidence of immunological events in such
a situation is known [22, 23].
Patients with a history of malignant tumors and those with
acute or chronic systemic infections, acute peptic ulcer disease,
pregnancy or insufficient contraceptive measures were excluded,
as were those under 18 years of age. Herpetic eye disease or any
other kind of acute corneal infection also led to exclusion. All pa-
tients underwent physical and laboratory examination prior to sur-
gery.
The 56 patients were randomized to receive CsA (27 patients)
or MMF (29 patients). MMF was administered in a daily dose of
2× 1 g; the CsA dose was adjusted according to blood trough lev-
els, with a target of 120–150 ng/ml (monoclonal TDx, Abbott Ex-
sym). After 6 months, immunosuppressive medication was tapered
within 2 weeks. Additionally, all patients except those with
steroid-induced glaucoma received corticosteroids systemically
(1 mg/kg body weight fluocortolone, tapered within 3 weeks post-
operatively) and topically (5 drops prednisolone acetate 1% daily
after epithelial consolidation, tapered within 5 months).
blood trough levels should not exceed 150 ng/ml. There-
fore, expensive laboratory monitoring is necessary.
Mycophenolate mofetil (MMF), the morpholinoethy-
lester of mycophenolenic acid (MPA), is a potent inhibi-
tor of inosine monophosphate dehydrogenase. MPA in-
hibits the proliferation of T and B lymphocytes [12]. The
efficacy and safety of MMF in a daily dose of 2× 1 g
combined with CsA and corticosteroids has been shown
after kidney transplantation [4, 5, 24].The efficacy of
MMF after penetrating keratoplasty has been demon-
strated in the rat model [16]. In a prospective random-
ized study, short-term results after mean follow-up of
about 10 months showed MMF to be an effective and
safe potential alternative to CsA for therapy after high-
risk penetrating keratoplasty [17]. Now, efficacy and
safety data for a larger group of patients with a much
longer follow-up of up to 3 years can be presented.
Patients and methods
Postoperatively, all patients were monitored for efficacy (clear
graft survival, immune reactions, endothelial cell loss) and drug
side effects (hepato-, nephro-, neuro- and gastroenterotoxicity).
Patient selection, treatment and follow-up
Fifty-six patients undergoing high-risk penetrating keratoplasty Postoperative visits were scheduled after 1, 3, 6, 9, 12, 24 and 36
were enrolled in this single-center study. All patients gave written months. At every visit, the patients underwent evaluation of visual
informed consent after obtaining approval by the local ethics com- acuity, slit-lamp examination, specular microscopy of the graft, es-
mittee. The study was performed in accordance with the Declara- timation of intraocular pressure and examination of the retina.
tion of Helsinki. High risk was defined by the presence of deep Furthermore, physical examination and blood tests were carried
vascularization in three or four quadrants, a history of previous out. Between attendances the patients were closely examined by
keratoplasty, position of the graft close to the limbus, transplanta- their general practitioners.
tion of a highly immunogenic graft (limbokeratoplasty), severe
Surgery was performed using grafts preserved in organ culture
atopic dermatitis or steroid response glaucoma (Table 1). In eight with an endothelial cell density of at least 2000 cells/mm2
keratoconus patients with rekeratoplasty (three in the CsA, five in (Table 2). The HLA types of donor and recipient could be deter-
the MMF group) first keratoplasty had been performed elsewhere. mined in 15 cases in the CsA group and in 18 cases in the MMF
Rekeratoplasty was judged necessary because of endothelial fail- group; in the remaining cases HLA typing of the donor failed
ure (two patients in the CsA, two in the MMF group), scarring of (Table 2). HLA typing was performed using serological tests for
the graft due to atopic blepharokeratoconjunctivitis (no patients in class I and PCR for class II. There was no statistically significant
the CsA, two in the MMF group), and for refractive reasons (one difference between the two groups concerning distribution of HLA
patient in the CsA, one in the MMF group). In six patients with mismatches (χ2 test, P=0.59). A 10.0 nylon double-running cross-
granular and lattice dystrophy rekeratoplasty was performed be- stitch suture was used for graft fixation, and, if necessary cataract
Table 1 Data of graft recipi-
ents
CsA
MMF
Stat. test (P)
1. Patients
27
29
2. Male/female
13/14
14/15
3. Mean age (years)
57.7 (28–87) 52.5 (18–86) t-test
19.3 (7–36) 19.8 (6–36) t-test
χ2 test
(0.30)
(0.80)
(0.80)
4. Follow-up (months)
5. High risk factor
a) Graft failure after previous keratoplasty
Keratokonus
15
3
3
3
3
2
1
4
2
3
3
0
19
5
3
4
5
2
0
2
3
1
2
2
Granular/lattice dystrophy
Scars
Bullous keratopathy
Fuchs’ dystrophy
Bacterial ulcer
b) Steroid induced glaucoma (bullous keratopathy)
c) Severe atopic dermatitis (keratokonus/keratitis)
d) Limbo-keratoplasty (granular/lattice dystrophy)
e) Deep vascularization in 3/4 quadrants (scars)
f) Graft position close to limbus
(eccentric keratokonus)

369
Table 2 Graft and HLA data
CsA
61.1 (32–84)
17.7 (0.5–72.0)
17.3 (10–32)
2400 (2088–3000)
MMF
55.1 (19–87)
11.5 (0.5–43.5)
17.0 (12–27)
2410 (2050–3060)
Stat. test
(P)
1. Donor age (years)
t-test
t-test
t-test
t-test
(0.19)
(0.17)
(0.72)
(0.72)
2. Time death/graft preparation (hours)
3. Organ culture period (days)
4. Endothelial cell density, preoperative (cells/mm2)
5. Graft diameter
7.7 mm (n)
22
5
4/23
21
8
5/24
χ2 test
χ2 test
(0.32)
(0.54)
8.2–9.7 mm (n)
6. Triple procedures/penetrating keratoplasty
7. HLA mismatches
0
2
2
0
2
1
2
0
1
3
1
2
4
4
3
5
4
5
6
2
5
Donor HLA-untyped
12
11
χ2 test
(0.59)
surgery was performed simultaneously. Postoperatively, immune
reactions were diagnosed by endothelial precipitates adhering to
graft endothelium with (severe endothelial immune reactions) or
without (mild endothelial immune reactions) stromal edema or by
the presence of non-infectious stromal infiltration (stromal im-
mune reactions) [1, 13, 15, 19]. These patients were treated with
prednisolone acetate 1% eye drops hourly and, in severe cases, ad-
ditionally with 1 mg/kg body weight fluocortolone.
Endothelial cell loss was assessed only in patients with at least
three postoperative endothelial cell density values. Patients with
endothelial immune reactions were excluded from this calculation.
The individual mean loss of endothelial cells per day and per
square millimeter was derived from the postoperatively acquired
endothelial values of each patient individually. This was done by
calculating the slope of the regression line for each scatter plot of
endothelial cell density values plotted against time [2, 11].
Statistical methods
Treatment groups were compared using the t-test for independent
samples (patient age, follow-up, donor age, period between do-
nor’s death and graft preparation, organ culture period, preopera-
tive and postoperative endothelial cell density) or the χ2 test (indi-
cation for surgery, graft diameter, HLA data, distribution of side
effects). Efficacy parameters were clear graft survival and occur-
rence of clinically detectable immune reactions. Calculation was
done using the Kaplan-Meier estimator, evaluation for statistical
significance by means of the log-rank test. A P value below 0.05
was considered to show a statistically significant difference.
Fig. 1 Kaplan-Meier clear graft survival curves (log-rank test:
P=0.33)
Efficacy
In both groups, no graft loss was observed during the
first postoperative year. During the whole follow-up,
four grafts failed in the CsA group and two grafts in the
MMF group (Table 3). Three years postoperatively, 69%
of the grafts were clear in the CsA group and 74% in the
Results
Demography
There was no statistically significant difference between MMF group. Graft survival did not show a statistically
the 2 treatment groups concerning patient age, follow- significant difference between the treatment groups
up, indication for surgery, donor age, period between do- (Fig. 1). Reasons for graft failure were glaucoma
nor’s death and graft preparation, organ culture period, (2 cases in the CsA group, 1 in the MMF group), graft
preoperative endothelial cell density, graft diameter and infection (1 in the CsA group) and endothelial immune
HLA data (Tables 1, 2).
reactions (1 in the CsA group, 1 in the MMF group).

370
Five patients in the CsA group and eight in the
MMF group experienced endothelial immune reactions
(Table 3). In both groups the majority of these reactions
were mild. Only one graft (in the CsA group during im-
munosuppressive therapy) was rejected twice. Three
years postoperatively, 73% of the grafts in the CsA
group and 53% of those in the MMF group were free of
immune reactions. However, this difference was not sta-
tistically significant (Fig. 2). The respective error of the
second kind (beta) was 0.42 for the sample sizes tested.
In both treatment groups most immune reactions were
reversible (4/5 in the CsA, 7/8 in the MMF group). Nei-
ther the two CsA patients without HLA mismatches nor
the two CsA patients with six HLA mismatches experi-
enced immune reactions or graft failure. In the MMF
group, one of the five patients with six HLA mismatches
experienced two mild, reversible endothelial immune re-
actions, another one graft failure due to glaucoma de-
compensation. Concerning endothelial cell loss, no sta-
tistically significant difference between the two groups
was found (Table 3).
Fig. 2 Absence of immune reactions calculated according to Ka-
plan-Meier (log-rank test: P=0.46)
Safety
diagnosis of Hodgkin’s lymphoma 1 month after kerato-
plasty, one of dermatological problems in a patient with
severe atopic dermatitis). Side effects were observed in
11 patients in the CsA group and 6 patients in the MMF
group (Table 3).
In the CsA group, premature withdrawal of immunosup-
pressive prophylaxis was performed in two patients (one
case of severe gingiva hyperplasia, one of hepatotoxici-
ty), and in the MMF group in further two patients (one
Table 3 Efficacy and safety data
CsA patients (n)
MMF patients (n)
Stat. test
1. Graft failure (n)
During/after immunosuppression
2. Reasons for graft failure (n)
Immune reaction
0/4
0/2
1
2
1
1
1
0
Glaucoma
Graft infection
3. Immune reactions (n)
During immunosuppression
Mild
2
2
1
Severe
1a
After immunosuppression
Mild
1
4
Severe
1
1
4. Endothelial cell loss (loss of cells/day × mm2)
5. Side effects
–2.03 (–6.8 to 0)
–1.26 (–4.6 to +0.5)
t-test (0.25)
χ2 test (0.12)
Hepatotoxicity
3b
3
2
Arterial hypertension
0
Gingiva problems
3b
3
0
Neurovegetative disorders
Hodgkin’s lymphoma (1 month after keratoplasty)
Recurrence of acoustic neurinoma (7 months after keratoplasty)
Exacerbation of atopic dermatitis
6. Premature withdrawal of drug
0
0
1b
1
0
0
1b
2
2
a Two severe immune reactions in the same patient under CsA
b One patient out of each group with premature withdrawal of the drug

371
reactions were mild and could be reversed, as could one
of the two severe immune reactions. This shift from se-
vere to milder endothelial immune reactions has already
been demonstrated as the main mechanism by which
CsA improves graft survival [14] and now seems to be
true as well for MMF. Probably, in those patients with
mild immune reactions only some incomplete immune
tolerance has been induced. The first graft failure in the
MMF group was noted 2 years postoperatively. One may
conclude that, at least in some eyes, pharmacologically
induced immune tolerance decreases after that period of
time. Here, the question arises of whether long-term ad-
ministration of systemic immunosuppressives is justified
to maintain immune tolerance and improve graft sur-
vival. In light of the similar efficacy of the 2 drugs tested
in this study and the higher risk of severe side effects in
the CsA group, MMF seems to be more likely to be used
for extension of the treatment period. Furthermore, in
contrast to CsA, blood level-adapted dosing of MMF
seems necessary only in special situations (e.g. treatment
failure, severe side effects), resulting in lower costs for
drug monitoring. Finally, the broad therapeutic range
makes MMF more likely to be administered in patients
with suboptimal compliance who fail to visit the
ophthalmologist or general practitioner on a regular ba-
sis.
Side effects were observed in only 6 of 29 patients in
the MMF group, in contrast to 11 of 27 patients in the
CsA group. The main side effects in the MMF group
were neurovegetative disorders, e.g., tremor and vertigo.
These side effects were well tolerated by most of the pa-
tients and vanished completely after cessation of immu-
nosuppressive therapy. Hodgkin’s lymphoma was detect-
ed 1 month after keratoplasty in one patient. Given the
short period between keratoplasty and diagnosis of the
lymphoma, occurrence of the latter may have been coin-
cidental. Probably, the same is true for the case of recur-
rence of acoustic neurinoma 7 months after keratoplasty.
In summary, no statistically significant difference in
efficacy between systemic MMF and systemic CsA ad-
ministered for 6 months after high-risk penetrating kera-
toplasty could be shown. Systemic MMF was proven to
be at least as safe as CsA. Pharmacologically induced
tolerance, however, decreases in the long run. Therefore,
long-term administration of MMF after high-risk pene-
trating keratoplasty should be investigated.
Discussion
The avascularity of the recipient cornea, reduced expres-
sion of MHC class I molecules, the absence of MHC
class II molecules, the secretion of immunomodulatory
molecules by the cornea itself, corneal expression of
complement-inhibiting molecules and the immune privi-
lege (anterior chamber associated immune deviation,
ACAID) of the anterior chamber are responsible for the
excellent prognosis of corneal grafts in “normal”-risk
keratoplasty patients [21]. Without systemic immuno-
suppression more than 90% of HLA-untyped grafts re-
main clear in the long run in such a situation [6, 18, 20,
25]. In high-risk keratoplasty, i.e. in patients with deep
host cornea vascularization in three or four quadrants, a
history of previous keratoplasty, position of the graft
close to the limbus or transplantation of a highly immu-
nogenic graft (limbokeratoplasty), up to 75% of high-
risk grafts fail due to immune reactions [7, 8, 22].
The benefit of systemic CsA in reducing immune re-
actions is unequivocal in organ transplantation [10]. The
effectiveness of systemic CsA in corneal high-risk trans-
plantation has been proven in several experimental [3,9]
and clinical [7, 8,14] studies. However, severe side ef-
fects are responsible for premature withdrawal of the
drug in about 10% of the patients [14]. Therefore, there
is a need for immunomodulatory alternatives. In this
study, MMF is the first immunosuppressant since the in-
troduction of systemic CsA into corneal surgery 15 years
ago being clinically investigated in a prospective ran-
domized trial after penetrating keratoplasty.
Concerning efficacy, i.e. clear graft survival, inci-
dence of immune reactions and endothelial cell loss, no
statistically significant difference was found between
systemic MMF and systemic CsA administered for 6
months after high-risk penetrating keratoplasty. This
may indicate that systemic MMF is at least as effective
as systemic CsA. However, the respective error of the
second kind (beta) was only 0.42 for the log rank test of
Fig. 2 (ratio of grafts without immune reactions), i.e.,
with a probability of 0.58 (power of the test) a real dif-
ference between the treatment effects would have been
detected using the sample sizes in this study. According
to our findings, to arrive at the conclusion that the 2
treatment procedures are statistically equivalent, the
power of the test would have to be increased to 0.95, i.e.,
250 patients per group would have to be treated with a
non-significant difference in outcome. This high number
indicates that there is a clinically negligible difference
between the two treatment effects. There is currently no
statistical procedure to test the equivalence of two sur-
vival curves.
Eight patients experienced immune reactions during
or after treatment with MMF. In these patients, the novel
immunosuppressant was insufficient to induce complete
immune tolerance. However, six of these eight immune

372
References
1. Allredge OC, Krachmer JH (1981)
Clinical types of corneal transplant re-
action. Arch Ophthalmol 99: 599–604
2. Böhringer D, Reinhard T, Godehardt E,
Sundmacher R (2001) Regressions-
analyse des idiopathischen Endothel-
zellverlustes nach perforierender
Keratoplastik. Klin Monatsbl
9. Kruse HA, Bechrakis NE, Heimann H,
Diamantstein T, Hoffmann F (1991)
Specific immunosuppressive therapy
by monoclonal antibodies on rejection
of orthoptic corneal allogenic grafts in
the rat. Invest Ophthalmol Vis Sci
32:1119
10. Land W (1987) Optimal use of
Sandimmun in organ transplantation.
Springer, Berlin Heidelberg New York,
pp 7–24
17. Reis A, Reinhard T, Voiculescu A,
Kutkuhn B, Spelsberg H, Althaus C,
Sundmacher R (1999) Mycophenolate
mofetil versus cyclosporin A in high-
risk keratoplasty patients: a prospec-
tive, randomised clinical trial.
Br J Ophthalmol 83:1268–1271
18. Schönherr U, Martus P, Händel A,
Naumann GOH (1996) Transplant-
atreaktion nach Keratoplastik wegen
Keratokonus. Ophthalmologe
Augenheilkd (in press)
3. Coupland SE, Klebe S, Karow AC,
Krause L, Bartlett RR, Hoffmann F
(1994) Leflunomide (HWA-486) in the
allogenic rat penetrating keratoplasty
model. Invest Ophthalmol Vis Sci
35:1878
4. European Mycophenolate Mofetil
Cooperative Study Group (1995)
Placebo controlled study of mycophe-
nolate mofetil combined with cyclo-
sporin A and corticosteroids for the
prevention of acute rejection. Lancet
345:1321–1325
5. European Mycophenolate Mofetil
Cooperative Study Group (1999)
Mycophenolate mofetil in renal trans-
plantation: 3 years results from the pla- 13. Pleyer U, Weidle EG, Lisch W, Steuhl
cebo-controlled trial. Transplantation
68:391–396
6. Girard LJ, Esnaolo N, Rao R (1993)
Allograft rejection after penetrating
keratoplasty for keratoconus.
Ophthalmic Surg 24:40–43
7. Hill JC (1989) The use of cyclosporine
in high risk keratoplasty. Am J
Ophthalmol 107:506–510
8. Hill JC (1994) Systemic cyclosporine
in high risk keratoplasty. Short- versus
long-term therapy. Ophthalmology
101:128–133
11. Langenbucher A, Nguyen NX, Kus
MM, Blüthner K, Küchle M, Seitz B
(2000) Regressionsverhalten des
Hornhautendothels nach nichtmech-
anischer perforierender Keratoplastik.
Klin Monatsbl Augenheilkd
93:227–231
19. Severin M (1986) Immunreaktionen
nach Keratoplastik. Klin Monatsbl
Augenheilkd 188:200–208
20. Sharif KW, Casey TA (1991) Penetrat-
ing keratoplasty for keratoconus:
complications and long-term success.
Br J Ophthalmol 75:142–146
21. Streilein JW, Yamada J, Dana MR,
Ksander BR (1999) Anterior chamber-
associated immune deviation, ocular
immune privilege, and orthotopic cor-
neal allografts. Transplant Proc
31:1472–1475
216:393–399
12. Morris R, Hoyt E, Murphy P (1990)
Mycophenolaic acid and morpholinoe-
thylester (RS-61443) is a new immuno-
suppressant that prevents and halts
heart allograft rejection by selective in-
hibition of T- and B-cell purine synthe-
sis. Transplant Proc 22:1659
22. Sundmacher R (1977) Immunreakti-
onen nach Keratoplastik. Klin
KP, Möhrle C, Richter U, Zierhut M,
Selbmann HK (1990) Klinische
Verlaufsformen immunologischer
Transplantatreaktionen nach perforie-
render Keratoplastik. Fortschr
Ophthalmol 87:14–19
Monatsbl Augenheilkd 171:705–722
23. Sundmacher R, Steffanson A,
Mackensen G (1983)
Verlaufsbeobachtungen nach
Keratoplastik. Klin Monatsbl
14. Reinhard T, Sundmacher R (1997)
Systemische Cyclosporin-A Prophyl-
axe nach Keratoplastiken mit erhöhtem
Risiko für Immunreaktionen als
einzigem erhöhten Risikofaktor.
Ophthalmologe 94:496–500
15. Reinhard T, Hutmacher M,
Sundmacher R, Godehardt E (1997)
Akute und chronische Immunreakti-
onen nach perforierender Keratoplastik
mit normalem Immunrisiko.
Augenheilkd 80:224–227
24. The Tricontinenetal Mycophenolate
Mofetil Renal Transplantation Study
group (1996) A blinded, randomised
clinical trial of mycophenolate mofetil
for the prevention of acute rejection
in cadaveric renal transplantation.
Transplantation 61:1029–1037
25. Troutman RC, Lawless MA (1987)
Penetrating keratoplasty for keratoco-
nus. Cornea 6:298–305
Klin Monatsbl Augenheilkd
210:139–143
16. Reis A, Reinhard T, Godehardt E,
Braunstein S, Sundmacher R (1998)
Effect of mycophenolate mofetil,
cyclosporin A, and both in a murine
corneal graft rejection model.
Br J Ophthalmol 82:700–703