Effect of suture material on postoperative astigmatism

Howard V. Gimbel, M.D., F.R.C.S.(C), Marsha G. Raanan, M.S., Maryclare DeLuca

ABSTRACT
Two hundred patients were enrolled in a randomized, prospective
clinical trial comparing the use of 10-0 nylon, 10-0 polypropylene
(ProIene-), 11.0 poly~ster (Mersilene-), and 10-0 polyethylene
(Noviifile ) sutUre materials on the amount and decay curves of sur-
. gicaUY.'indllced astigmatism following intraocular lens (IOL) sur-
gery~ Patients with Mersilene and nylon sutures had the highest
ll"'Olltl~ of induced with-the-rule (WTR) cylinder (significantly
more than Prorene} at one day after surgery. However, the WTR
cylinder decayed rapidly for nylon during the first three months hut
more slowly for Mersilene because of its lack ofstretchahility. The
Prolene group had the lowest level ofinduced WTR cylinder at one
day, bi1tapinst~the~rule (ATR) drift occurred, .(eaving cases with
ATRastigmatism by a year. The nylon group had the second highest
alllsunt ofinduced WTR cylinder at one day, which had decayed to
'ATBcYliDderhyJ.1ve monthS. Between one and two years postop-
eratively, the nylon group experienced asignificant ATR shift. The
,'aIllonnt:ofearlyinducedWTR cylinder seemed to he related to the
knoHying technique and tissue gripping characteristics, whereas
.t heshapeofthe deCl,lY curve was related to the material character-
istics of the'suture.

K~y' W~l'ds:induced astigmati~m, intraocular lens surgery, nylon,

, polyester; polyethylene, polypropylene, suture, with-
the-rule astigmatism

For approximately ten years, we have used 10-0
polypropylene (Prolene ill) sutures routinely at the
Gimbel Eye Centre. Prolene does not hydrolyze
and break as nylon has a tendency to do. The ma-
jority of surgeons, however, use 10-0 nylon su-
tures. Polyester (Mersilene®) is also used and, more
recently, the synthetic substance polyethylene
(N ovafil®). All suture materials have differing char-
acteristics and elasticity. Nylon is generally the
most stretchable suture, Mersilene does not stretch
much, while Prolene and Novafil have similar han-
dling characteristics.
We conducted a prospective, four-group study
comparing the use of 10-0 Prolene, 11-0 Mersi-
lene, 10-0 Novafil, and 10-0 nylon sutures follow-
ing phacoemulsification with insertion of a
posterior chamber intraocular lens (IOL) to deter-
mine if there were significant group differences in
surgically induced cylinder and visual rehabilita-
tion time.
MATERIALS AND METHODS
Two hundred patients were enrolled in a pro-
spective, four-group study comparing the use of
10-0 Prolene, 11-0 Mersilene, 10-0 Novafil, and
10-0 nylon sutures following phacoemulsification
with insertion of a posterior chamber IOL. Patients

From the Gimbel Eye Centre, Calgary, Alberta, Canada, and the Department of Ophthalmology, University of Illinois at Chicago .
Analysis of the data was performed by the Center for Clinical Research, University of Illinois at Chicago.
Reprint requests to Howard Gimbel, M.D. , Gimbel Eye Centre, Suite 450, 4935-40th Avenue N. W., Calgary, Alberta, Canada
T3A 2Nl.

42 J CATARACT REFRACT SURG-VOL 18, JANUARY 1992

were randomized to receive one of the four suture
types. Patients were eligible for enrollment if they
were scheduled to have uncomplicated primary
cataract/posterior chamber IOL surgery using pha-
coemulsification as the method of extraction. Pa-
tients also had to have less than 2.0 diopters (D) of
primary keratometric cylinder preoperatively and
had to agree to return to the Gimbel Eye Centre for
a three-month postoperative visit. Upon enroll-
ment, patients were randomized to receive one of
the suture materials. The randomization schedule
was generated using PROD AS (Professional Data-
base Analysis System) software. l
The wound was made superiorly centered about
the 110 degree axis to facilitate right-handed pha-
coemulsification. A small, limbus-based conjuncti-
val flap was formed. Very light cautery of episcleral
vessels, using a Codman Mentor CMC II cautery
set at 8 and Mentor bipolar eraser-type tip, was
done. Using a diamond knife, a reverse 0.75 mm
deep and 6.0 mm or 7.0 mm long groove was made
1.5 mm from the conjunctival reflection. The blade
orientation was then changed and, from the depths
of the groove, the incision was carried forward to
leave a thin shelf of corneal scleral tissue over
Schlemm's canal and the trabecular meshwork. A
central 3 mm entry into the anterior chamber was
made for phacoemulsification and then enlarged
with the diamond knife to 6.5 mm or 7.5 mm for
either 6.0 mm or 7.0 mm diameter posterior cham-
ber IOL implantation.
Continuous shoelace closure was used in all
cases. Suture adjustment was made after air was
used to fill the anterior chamber but not overinflate
it to cause the wound to gape. Because the nylon
suture readily released upon attempting to close
using a single throw on the first pass of the knot, a
double throw was used. For all other sutures, a
single throw on the first pass was used.
Patients were generally examined at one day (0
to 3 days), two to three weeks (12 to 27 days), two
months (6.0 to 11.4 weeks), three months (11.6 to
17.0 weeks), five to six months (4.5 to 7.5 months),
one year (10.0 to 15.75 months), and two years
postoperatively (18 to 27 months). At these visits,
keratometry measurements and refractions were
taken. All measurements were performed at the
Gimbel Eye Centre. Four cases were lost to follow-
up prior to the three-month visit and were
replaced.
The technicians performing keratometry mea-
surements were unaware of the randomly assigned
suture type group both prior to and following sur-
gery. Keratometric readings were made using the
Haag-Streit keratometer with the eyepiece ad-
justed for the operator.
To quantify surgically induced cylinder fully, we
used a vector analysis procedure that calculates the

Table 1. Patient characteristics.

Characteristics Prolene Mersilene Novafil Nylon

Total number of patients 52 48 50 50
Sex
N (%) male 16 (31 %) 14 (29%) 19 (38%) 17 (34%)
N (%) female 36 (69%) 34 (71 %) 31 (62%) 33 (66%)
Age
Mean (SD) 71.7 (11.2) 70.7 (12.3) 72.3 (10.4) 71.1 (11.1)
Range 47-94 38-93 44-87 42-90
Follow-up percentage
1 day 100% 100% 100% 100%
2-3 weeks 98% 96% 96% 94%
2 months 100% 90% 96% 98%
3 months 88% 77% 80% 88%
5-6 months 63% 56% 78% 76%
1 year 77% 90% 90% 88%
2 years 52% 58% 50% 54%
Follow-up
Mean (SD) 16.8 (7.3) 18.0 (6.9) 17.1 (6.4) 17.5 (6.6)
Range 2.5-26.75 2.0-25.5 3.0-25.25 0.75-26.5

J CATARACT REFRACT SURG-VOL 18, JANUARY 1992 43

power and axis of the surgically induced cylinder
based upon the equations developed by Naylor.2
Postoperative cylinder and surgically induced cyl-
inder were both analyzed with a one-way analysis
of variance using a Bonferroni adjustment for mul-
tiple comparisons.
Surgically induced astigmatism was also broken
down for each case into a with-the-rule (WTR)
component and an against-the-rule (ATR) compo-
nent. This method, developed by Cravy,3 allows
examination of WTR cylinder decay. The mean
changes in WTR cylinder within successive post-
operative intervals were also analyzed using a one-
way analysis of variance with a Bonferroni
adjustment for multiple comparisons.
Differences in sex of the patients in suture type
groups were analyzed using the chi-square
method. Age differences between suture types
were analyzed with a one-way analysis of variance.
All statistical analyses were performed using
PRODAS. 1
RESULTS
Characteristics of the patients who received each
suture type are presented in Table 1. Average age
for the whole population was 72 years; approxi-
mately two thirds of the cases were female and one
third was male. When analyzed separately, the su-
ture type groups did not reveal significant differ-
ences in age or sex.
Table 1 also shows the follow-up rates at the
major postoperative periods. Follow-up was 94%
to 98% at two to three weeks. Throughout the first
year, at least 77% of patients returned for visits
except at five to six months. Scheduling problems
caused follow-up inconsistencies for the five to six
month visit. Follow-up at two years decreased to

Table 2. Preoperative and postoperative cylinder.

Interval Prolene Mersilene Novafil Nylon

Preoperative N = 52 N = 48 N = 50 N = 50
Mean (SD) 0.85 (0.56) 0.87 (0.54) 0.93 (0.56) 0.85 (0.42)
Range 0.25 to 2.5 0.25 to 2.5 0.25 to 2.75 0.25 to 2.25
1 day* N = 52 N = 48 N = 50 N = 50
Mean (SD) 1.42 (1.00) 2.23 (1.54) 1.65 (1.08) 2.03 (1.69)
Range 0.0 to 4.25 0.13 to 7.12 0.12 to 5.0 0.13 to 8.5
2-3 weeks N = 51 N = 46 N = 48 N = 47
Mean (SD) 1.20 (0.83) 1.58 (1.12) 1.32 (0.65) 1.34 (0.90)
Range 0.25 to 4.5 0.13 to 5.13 0.0 to 2.75 0.0 to 4.5
2 monthst N = 51 N = 41 N = 47 N = 49
Mean (SD) 1.01 (0.59) 1.68 (1.30) 1.14 (0.52) 1.19 (0.76)
Range 0.13 to 2.63 0.25 to 6.62 0.12 to 2.63 0.25 to 3.88
3 months N = 46 N = 37 N = 38 N = 44
Mean (SD) 0.90 (0.47) 1.24 (0.91) 1.09 (0.59) 1.12 (0.79)
Range 0.0 to 2.0 0.12 to 4.75 0.25 to 2.25 0.0 to 3.75
5-6 months N = 31 N = 25 N = 39 N = 38
Mean (SD) 1.01 (0.57) 0.95 (0.61) 1.18 (0.58) 1.01 (0.61)
Range 0.12 to 2.37 0.0 to 2.75 0.13 to 2.5 0.12 to 2.75
1 year N = 38 N = 43 N = 45 N = 43
Mean (SD) 1.04 (0.62) 1.03 (0.73) 1.07 (0.66) 0.95(0.62)
Range 0.13 to 2.88 0.12 to 3.13 0.0 to 3.38 0.0 to 2.5
2 year N = 26 N = 26 N = 25 N = 27
Mean (SD) 1.23 (0.73) 0.875 (0.50) 1.16 (0.70) 1.19 (0.81)
Range 0.12 to 2.88 0.25 to 2.00 0.25 to 3.50 0.25 to 3.25
Significant differences:
* Prolene and Mersilene (P < .01)
t Mersilene and each group (P < .01)

44 J CATARACT REFRACT SURG-VOL 18, JANUARY 1992

50% to 58%. Each of the four groups had a mean
follow-up of between 16.8 and 18.0 months.
Preoperative and postoperative keratometric
cylinder values are summarized in Table 2 and pre-
sented graphically in Figure 1. There were no sig-
nificant differences among groups in distribution of
wound sizes and mean preoperative cylinder. Be-
cause 6.0 mm and 7.0 mm diameter optic IOLs
were used in this study, 6.5 mm or 7.5 mm inci-
sions were made. The overall distribution of inci-
sion sizes for all suture material groups was 45%
6.5 mm incisions and 55% 7.5 mm incisions. The
distribution within each suture material group was
similar. Net postoperative and surgically induced
cylinder was not significantly different (all P values
~ .3) between 6.5 mm and 7.5 mm incision sizes,
overall or within any suture material group. There-
fore no stratified analyses are presented.
For the first three months after surgery, the Mer-
silene group consistently had the highest postop-
erative mean cylinder, while Prolene consistently
had the lowest. The order of decreasing mean post-
operative cylinder was Mersilene, nylon, Novafil,
and Prolene.
The largest difference between the Mersilene
and Prolene groups occurred at the one-day visit.
At one day, the 0.8 dioptric difference between
Mersilene (2.2 D) and Prolene (1.4 D) was statis-
tically significant (P < .01). At two months, cylin-
der in all groups except the Mersilene group
regressed substantially, such that Mersilene had
significantly (P< .01) more postoperative cylinder
than each of the other suture type groups.
After three months, the suture type groups were
more comparable through one year. There were no
M 2.5
E
A ~ PROLENE -+- MERSILENE -+- NOVAFIL
N
2
p
o
S
T 1.5
o Fig. 1.
P tive cylinders over time for
c each suture material group.
y
L
I 0.5
N
~ I
J''"'1'''S--r
TIME
R 2'3-'S.--------1'Y-R---···----·.
PREOP D W MM M
J CATARACT REFRACT SURG-VOL 18, JANUARY 1992
statistically significant differences among groups at
subsequent times. At one year, cases in all groups
still generally had about a diopter of postoperative
cylinder, on the average, which was similar to pre-
operative levels. By two years, postoperative ke-
ratometric cylinder means were slightly higher
than the one-year means (approximately 0.1 to
0.2 D higher) for all groups except the Mersilene
group, in which the two-year mean was slightly
lower (by 0.14 D) than the one-year mean.
The means of the surgically induced cylinder
values (as computed by vector analysis) at each
postoperative period are presented in Table 3 and
Figure 2. During the first three months, Mersilene
had the highest amount of induced cylinder and
Prolene had the lowest. At one day, the differences
in induced cylinder between nylon and Prolene
and between Mersilene and Prolene were statisti-
cally significant (P < .01). Initially, the induced
cylinder regressed after one day for all groups,
although Mersilene maintained significantly
higher induced cylinder through two months. After
two months postoperatively, there were no statis-
tically significant differences among groups. How-
ever, while the Mersilene and Novafil groups
maintained stable levels of induced cylinder (ap-
proximately 0.9 D) at one and two years, both the
nylon and Prolene groups had two-year mean in-
duced cylinders which were higher than at one
year. The Prolene group mean increased by a third
of a diopter (from 0.88 D to 1.23 D) and the nylon
mean increased by approximately one half diopter
(from 0.76 D to l.27 D).
Figure 3 shows the WTR cylinder decay (as cal-
culated by the Cravy method) for each of the four
-a- NYLON
(Gimbel) Line graphs of mean
preoperative and postopera-
,
2 YR
45
Table 3. Surgically induced cylinder at the major periods.

Postoperative
Period Prolene Mersilene Novafil Nylon
1 day· N = 52 N= 48 N = 50 N = 50
Mean (SO) 1.09 (0.72) 1.85 (1.58) 1.45 (0.93) 1.74 (1.44)
2-3 weekst N = 51 N = 46 N = 48 N = 47
Mean (SO) 0.85 (0.81) 1.32 (1.05) 1.04 (0.68) 1.05 (0.70)
2 monthst N = 51 N = 41 N = 47 N = 49
Mean (SO) 0.81 (0.53) 1.27 (1.23) 0.88 (0.57) 0.90 (0.72)
3 months N = 46 N = 37 N = 38 N = 44
Mean (SO) 0.80 (0.44) 1.06 (1.02) 0.97 (0.49) 0.91 (0.53)
5-6 months N = 31 N = 25 N = 39 N = 38
Mean (SO) 0.89 (0.50) 0.80 (0.81) 0.94 (0.65) 0.79 (0.49)
1 year N = 38 N = 43 N = 45 N = 43
Mean (SO) 0.88 (0.52) 1.10 (0.87) 0.86 (0.57) 0.76 (0.48)
2 year N = 26 N = 26 N = 25 N = 27
Mean (SO) 1.23 (0.67) 0.97 (0.58) 0.92 (0.64) 1.27 (0.65)
Significant differences:
• Prolene and Mersilene, Prolene and nylon P < .01
t Prolene and Mersilene P = .05
t Prolene and Mersilene P = .03

M ~ PROLENE + MERSILENE +- NOVAFIL -B- NYLON

E
A
N
I
N
o
U
Fig. 2. (Gimbel) Line graphs of mean C
surgically induced cylinder E
over time for each suture ma-
o
terial group. C
y
L
1 0.5
N
o
E
R
TIME
O~-'-.--.-----------r-------------------r---
1 3 2 3 5 1 2
OW M M M Y Y

groups. Mersilene and nylon had the largest
amounts of induced WTR cylinder at one day. Sig-
nificant decay occurred during the first three
months and continued at a slower rate through one
year. After one year, the Mersilene group had no
significant decay in either direction. Between two
months and one year, mean ATR shift for this group
was -1.07 D, with 43% of cases having> 1.0 D of
ATR shift. However, between one and two years,
mean shift was +0.09 D (WTR direction) and only
8% of cases had ATR cylinder axis shifts> 1 D
(Figure 4).

46 J CATARACT REFRACT SURG-VOL 18, JANUARY 1992

 1.5
- PROLENE -+- MERSILENE --* NOVAFIL -B- NYLON

0.5
WTR
Fig. 3. (Gimbel) Line graph of mean

ATR 0~:===5=~~
I "
Cravy-calculated induced as-
tigmatism depicting WTR cyl-
-0.5 inder decay over time for each
suture material group.

-1

TIME
-1.5~r-.-.---.-------------.----------------------.
1 3 2 3 5 1 VR 2 VR
o W M M M

The pattern ofWTR cylinder decay was different
in the nylon group. Mean ATR shift between two
months and one year was -0.47 D, with 21 % of
cases shifting in the ATR direction by more than
one diopter. Between one and two years, the mean
ATR shift was -0.6 D and 32 % of nylon cases had
ATR shifts greater than 1.0 D, showing that, unlike
Mersilene, much of the overall ATR drift in the
nylon group occurred after one year.
The Prolene group also had a significant mean
decrease in WTR cylinder (0.8 D) during the first
three months (Figure 3), but this leveled off be-
tween three and six months. Since so little
« 0.5 D) WTR was induced at one day, by six
months these cases had mean ATR cylinder of over
half a diopter. After six months, the decay stabi-
WTR ATR
Fig. 4. (Gimbel) Bar charts of percent of cases with> 1 D of
ATR cylinder shifts between one year and two years
for each suture material group.
lized. Between one and two years, 16% of cases had
ATR drift> 1.0 D (Figure 4).
Although 1.45 D of total induced cylinder had
been induced in the Novafil group at one day
(Figure 2), it was almost evenly split between WTR
and ATR, resulting in less than 0.5 D of mean WTR
cylinder for this group at one day (Figure 3). By
two months, the proportions ofWTR and ATR cyl-
inder for these cases were similar. Through one
year, there was no large preponderance of either
WTR or ATR cylinder shifts in this group; however,
between one and two years, 17% of cases did have
> 1.0 D of ATR shift.
To assess whether data indicating late shifts in
axis from one to two years were affected by patient
attrition between one and two years (see Table 1),
the data were analyzed by the Cravy method using
only cases that were examined at one day, two
months, three months, one year, and two years.
The pattern of late ATR drift was the same
(Figure 5).
Table 4 shows the percentage of cases with un-
aided visual acuities of 20/40 or better preopera-
tively and at the various postoperative visits.
The unaided visual results were comparable,
with no statistically significant differences
among suture types at any period. The percent-
age of cases with 20/40 or better acuity at one
day after surgery was 40% to 62%. By two
months, 74% to 81 % had acuities of 20/40 or bet-
ter. Twenty-nine percent to 42% had acuities of
20/25 or better. This level of visual acuity was
maintained through one year. At two years, 68% to
81 % of the cases that returned had unaided visual

J CATARACT REFRACT SURG-VOL 18, JANUARY 1992 47

acuities of20/40 or better and 32% to 43%,20/25
or better.
DISCUSSION
Commonly, when analyzing induced cylinder,
the means and ranges of dioptric power changes
from the preoperative to the postoperative mea-
surements are compared without regard to any cyl-
inder axis changes. Thus, surgically induced
cylinder change will be accurately described only
for cases in which the postoperative axis is exactly
the same as the preoperative cylinder. For exam-
ple, a preoperative cylinder of + 1.00 X 90, con-
verted to a postoperative cylinder of + l.00 X 180,
would appear to have no change in cylinder when,
in fact, a +2.00 D change in cylinder occurred at
axis 180. Computing induced cylinder is more
complex if the postoperative axis is neither the
same as nor 90 degrees away from the preoperative
value . Naylor's equations are based upon a trigo-
nometric solution for the resultant of two obliquely
crossed cylinders (vector analysis programs). In-
stead of solving for the resultant cylinder, however,
the equations find the second component (surgi-
cally induced cylinder power and axis) from a
knowledge of the first component (preoperative
cylinder power and axis) and the resultant (post-
operative cylinder power and axis). Vector analy-
sis is gaining more widespread use and has been
used to evaluate induced cylinder following cata-
ract extraction 4 ,5 and following astigmatic
keratotomy. 6 , 7
Many factors determine the amount of surgically
induced astigmatism, the decay rate of that in-
duced astigmatism, and the possible drift toward
ATR astigmatism. In this study, one factor (the way
the incision is made) was constant for all groups.
The type of closure was also standard for all groups.
However, the effect of suture type on the judgment
of suture tension using a continuous suture can
influence the amount of induced cylinder from
tight or loose sutures. In this study, the surgeon's
greatest experience and, therefore, probably best
judgment on suture tension was with the Prolene
material.
Even though most surgeons use nylon, one of the
authors (H.V.C.) does not because of its biodegrad-
ability; it can break and extend through conjunc-
tiva, irritating the eye. Also, reports by Byes
("Nylon Biodegradation and Late Post-op Astigma-
tism Increase-the Mersilene Advantage," pre-
sented at the Welsh Cataract Congress, November
1988) and Cravy 8 indicate that use of nylon suture,
which eventually dissolves causing the wound to
relax late, may lead to ATR drift in astigmatism
years after surgery. We, in fact, saw a dramatic ATR
drift between one and two years postoperatively
with the use of nylon.
Prolene does not hydrolyze and biodegrade and
is more wiry than nylon or Novafil. Prolene slides
quite easily through the tissue so we did not see
much early induced WTR cylinder, as with Mersi-
lene. However, with the tension not maintained,
the looser wound led to an early drift toward ATR
astigmatism, which stabilized after six months.
Mersilene is very wiry, more so than Prolene.

I
1.5
N
D -+- MERSILENE -+- NOVAFIL -B- NYLON
U
C
Fig. 5 . (Gimbel) Line graph of mean E
Cravy-calculated induced as- D 0.5
tigmatism depicting WTR cyl-
inder decay over time for each A
suture material group. Cases S 0
were examined at the one day, T
two week , two month, three I
month, one year, and two year -0 .5
visits.
G
M
A - 1
T
I TIME
S -1.5
M 1 3 2 3 1 VR 2 VR
D W MM

48 J CATARACT REFRACT SURG-VOL 18, JANUARY 1992

Table 4. Percent of cases with postoperative VAse of 20/40 or better.

Postoperative
Period Prolene Mersilene Novafil Nylon
1 day
20/15-20/25 11 (21 %) 7 (16%) 8 (16%) 5 (10%)
20/30-20/40 17 (33%) 19 (42%) 12 (24%) 26 (52%)
2-3 weeks
20/15-20/25 11 (22%) 10 (23%) 17 (36%) 13 (28%)
20/30-20/40 25 (51 %) 18 (41%) 20 (43%) 21 (44%)
2 months
20/15-20/25 20 (38%) 18 (42%) 18 (38%) 14 (29%)
20/30-20/40 21 (41 %) 15 (34%) 17 (36%) 25 (52%)
3 months
20/15-20/25 15 (33%) 14 (38%) 15 (38%) 14 (33%)
20/30-20/40 19 (41 %) 16 (43%) 13 (32%) 19 (46%)
5-6 months
20/15-20/25 10 (30%) 12 (44%) 17 (46%) 14 (38%)
20/30-20/40 13 (39%) 7 (26%) 15 (40%) 15 (40%)
1 year
20/15-20/25 16 (41 %) 17 (46%) 11 (30%) 17 (44%)
20/30-20/40 17 (44%) 10 (27%) 18 (49%) 10 (26%)
2 years
20/1.5-20/25 9 (33%) 9 (35%) 10 (43%) 8 (32%)
20/30-20/40 12 (44%) 12 (46%) 6 (26%) 9 (36%)

When pulled tightly it induces a high level ofWTR
cylinder. It then holds tightly because of its lack of
stretchability, maintaining a higher level of WTR
cylinder over time than do other suture types. This
was the pattern we observed. Even though the de-
cay rate was slower, eventually the level of induced
cylinder came down, with levels comparable to the
other suture types by a year. The manufacturer had
recommended the use of 11-0 rather than 10-0
suture based on the belief that 11-0 Mersilene was
equivalent in strength to 10-0 nylon. This thinner
suture probably "cheese-wires," cutting through
tissue to release the tension.
One author (H.V.C.) had no previous clinical
experience with Novafil and included it in the
study for completeness. The rapid decay of in-
duced WTR astigmatism in the Novafil group may
relate to the stretchability of that material, which
quickly forgives an overtight adjustment of the su-
ture. Prolene and nylon seem to fall midway be-
tween Mersilene and N ovafil in this regard.
In retrospect, the two weaknesses of the com-
parability of groups in the study were the use of
11-0 rather than 10-0 Mersilene and the unavoid-
able differences in knot-tying technique. The
amount of early induced WTR cylinder is believed
to be related to the knot-tying technique (in turn,
due to the handling characteristics) and to tissue
gripping characteristics (ease of sliding through tis-
sue). The shape of the decay curve, however, is
related to the material characteristics of the suture.
This study, carried out over 21/2 years, resulted
in absolute levels of surgically induced cylinder,
which reflected the incision techniques used at the
time. The surgeon's (H.V.C.) incision technique is
currently very different. Nevertheless, the relative
differences among suture materials do apply to cur-
rent techniques.
In general, we are concerned with apposing the
external aspect of the wound with sutures and have
not been concerned with the internal aspect. Not
all patients develop ATR drift, even in groups in
which the means clearly shift in that direction. Per-
haps cases in which the internal aspect of the
wound is well apposed are the ones that do not
demonstrate the ATR drift. If this is true, there may
be aqueous access to the large portion of the inci-
sion in cases in which the internal corneal aspect of

J CATARACT REFRACT SURG-VOL 18, JANUARY 1992 49


Fig. 6. (Gimbel) Aqueous access to inner aspect of incision
may cause permanent gaping and, therefore, late
ATR.
the incision is not apposed. This could cause per-
manent gaping, wound slippage, and late ATR
drift. If this hypothesis is correct, then self-sealing
.50 J CATARACT REFRACT SURG-VOL 18. JANUARY
internal architecture may prevent late ATR drift
(Figure 6).
REFERENCES
l. Conceptual Software, Inc. Professional Database Analysis Sys-
tem (PRODAS) Reference Manual, Version 3.0, Houston,
1985
2. Naylor EJ. Astigmatic difference in refractive error. Br J Oph-
thalmol1968; 52:422-424
3. Cravy TV. Calculation of the change in corneal astigmatism
following cataract extraction. Ophthalmic Surg 1979; 10:
38-49
4. Jaffe NS, Clayman HM. The pathophysiology of corneal astig-
matism after cataract extraction. Trans Am Acad Ophthalmol
Otolaryngol1975; 79:615-630
5. Neumann A, McCarty G, Sanders D, Raanan M. Small incisions
to control astigmatism during cataract surgery. J Cataract Re-
fract Surg 1989; 15:78-84
6. Thornton SP, Sanders DR. Graded nonintersecting transverse
incisions for correction of idiopathic astigmatism. J Cataract
Refract Surg 1987; 13:27-31
7. Neumann A, McCarty G, Sanders D, Raanan M. Refractive
evaluation of astigmatic keratotomy procedures. J Cataract
Refract Surg 1989; 15:25-31
8. Cravy TV. Long-term corneal astigmatism related to selected
elastic, monofilament, nonabsorbable sutures. J Cataract Re-
fract Surg 1989; 15:61-69
1992