Sei sulla pagina 1di 9

1040-5488/02/7902-0103/0 VOL. 79, NO. 2, PP.

103111
OPTOMETRY AND VISION SCIENCE
Copyright 2002 American Academy of Optometry

ORIGINAL ARTICLE

Oxygen Permeability (Dk) of Thirty-Seven Rigid


Contact Lens Materials
WILLIAM J. BENJAMIN, OD, MS, PhD, FAAO and QUIDO A. CAPPELLI, BCE, PE
University of Alabama at Birmingham, School of Optometry, Birmingham, Alabama

ABSTRACT: Purpose. Oxygen permeability (Dk) was determined for 37 available rigid contact lens materials in a
masked fashion. The results were compared with those of an earlier study that included different lots of 14 test materials
assessed in the current study. Methods. Six lenses of different thicknesses in each test and reference material were
obtained. Test materials were arranged in sets of six to eight materials per set. Each set of materials, with inclusion of
at least two reference materials for the purpose of simultaneous calibration, was measured to obtain preliminary
amperages. Four preliminary measures were performed per thickness, resulting in 24 per material, in a schedule
designed to spread the potential effects of machine drift and other factors. The mean preliminary amperages were used
to derive corrected Dk values according to American National Standards Institute (ANSI) Z80.20-1998, and the values
were linearly calibrated using the measured and established Dk values of the reference materials. Results. The
resistance (t/Dk) vs. thickness (t) plots for the 37 test and seven reference materials were approximated linearly. In 54
of 57 linear regressions, the coefficients of determination (R2) were >0.96, and in 48 instances were >0.98. Fourteen
Dk values from the current study and an earlier study were linearly correlated (R2 0.9846), with a slope close to unity
(1.056) and intercept close to zero (0.292). Ten of the current values fell within 10% of their corresponding earlier
values. Only three current Dk values fell outside of the ANSI Z80.20-1998 tolerance for Dk (20%). Two of these Dk
values met the product tolerance when an obvious outlying point was graphically identified and omitted from the linear
resistance (t/Dk) vs. thickness (t) regression. Conclusion. Omission of a single outlying point from a linear resistance
vs. thickness regression can help provide a more valid Dk value. The ANSI Z80.20-1998 tolerance of 20% on Dk and
the measurement reproducibility of 10% were achieved for the overwhelming majority of rigid contact lens materials
up to at least 160 Dk units. The corrected, calibrated Dk values for the 37 test materials, in ANSI units, ranged from
13.8 to 175.1, having an overall mean of 43.2, median of 31.9, and standard error of 6.57 (N 37). (Optom Vis Sci
2002;79:103111)

Key Words: oxygen permeability, contact lens, edge effect, boundary-layer effect

W
ith the introduction of contact lenses made of cellulose ment was the subject of confusion. Dr. Irving Fatt, by virtue of his
acetate butyrate in the early 1970s, a novel feature for earlier attempts to measure the transmittance of oxygen through
rigid contact lens materials, oxygen permeability, be- contact lenses using a polarographic method, became the leading
came the subject of increasing scrutiny. A rigid contact lens mate- authority on the oxygen permeability (Dk) of contact lens materi-
rial capable of passing more significant amounts of oxygen was als during that early period. Over time, the technique was refined.
made commercially available in 1978. Soon thereafter, several new Corrections for the boundary-layer effect and the edge effect
rigid gas-permeable (RGP) materials entered the contact lens mar- were described in 1982 and 1987, respectively.3, 4 The oxygen
ket and the oxygen era for contact lenses was created. With these permeability values of a set of reference materials were made avail-
new materials, many corneal traumas during rigid contact lens able in 1997 and 1998 for calibration of Dk measurements.57
wear became more subtle than overt, the partial result of chronic However, many Dk values for contact lens materials were pub-
hypoxia rather than acute anoxia.1 Thereafter, a primary index of lished or advertised without correction for either or both effects
the acceptability of a transparent polymer for contact lens wear and/or were not adequately calibrated.8, 9
became its oxygen permeability value.2 In 1982 the Contact Lens Manufacturers Association (CLMA)
Although the concept of a gas dissolving in a polymer and trans- developed an initial standardized method for the measurement and
porting from one surface to another was appreciated, its measure- reporting of oxygen permeability (Dk) and transmissibility (Dk/t).

Optometry and Vision Science, Vol. 79, No. 2, February 2002


104 Oxygen Permeability of Contact LensesBenjamin and Cappelli

This polarographic method was submitted in 1986 to a working radius of 7.80 mm and overall diameter of 9.5 mm. Upon com-
group of the International Standards Organization (ISO) having pletion of the lenses for each set of materials, the sets were for-
jurisdiction over contact lens matters to begin development of a warded to the University of Alabama at Birmingham Eye Physiol-
standard polarographic Dk measurement method at the interna- ogy & Ocular Prosthetics Laboratory for testing.
tional level. The polarographic method was deemed applicable to Five sets of rigid contact lenses were delivered from the CLMA
rigid (nonhydrogel) and conventional hydrogel materials. In 1988, during 1999, and a sixth set was delivered in January of 2000. Each
the ISO working group also began drafting a coulometric method set of lenses was composed of 8, 7, or 4 different rigid contact lens
for use with highly permeable nonhydrogel materials. The Amer- materials. There were 12 contact lenses supplied per material: sets
ican National Standards Institute (ANSI) Z80 Committee incor- A and B contained 96 lenses; sets C, D, and E contained 84 lenses;
porated the ISO drafts into a contact lens standard developing and set F contained 48 lenses. To set F were added the lenses of
within the United States in 1992. The polarographic and coulo- three materials from earlier sets for which remeasurement was re-
metric methods became part of ANSI Z80.20, issued in 1998.10 quested by the CLMA Technical Affairs Committee. At testing,
The polarographic method had become an ISO standard (ISO therefore, set F consisted of 84 lenses from seven designated mate-
9913-111) in 1996, and the coulometric method was published as rials. Each set of lenses was analyzed separately, in the sequence A,
an ISO standard in 2000 (ISO 9913-212). Use of a standard meth- B, C, D, E, and F between June 23, 1999, and January 21, 2000.
odology confirmed that RGP contact lenses were capable of trans- Each storage case containing lenses to be tested was delivered to
porting more oxygen to the cornea than conventional hydrogel the Eye Physiology & Ocular Prosthetics Laboratory with an iden-
contact lenses, which were limited by the permeability of water to tification code established by the CLMA that masked the identities
oxygen. Of recent development are highly permeable silicone-hy- of the materials from the investigator and operator of the Dk
drogel (unconventional) materials for which a standard method of apparatus. The identities of the materials were not known by the
measurement is under consideration. investigator until after the calibrated Dk values had been estab-
It became recognized that with respect to oxygen permeability, lished and the investigator was notified as to the identity of the
the marketing of contact lenses had been allowed to take prece- materials by the CLMA. These codes and the names of the test
dence over the science surrounding them.1, 9 In 1998, the Board of materials have been supplied in Tables 1 through 6 to assist the
Directors of the CLMA became concerned with incidents of seem- reader in reviewing the data and results.
ingly inaccurate Dk values reported by manufacturers seeking to The target (nominal) thicknesses of the lenses of each material
promote materials through the use of this characteristic. The were 0.11, 0.14, 0.18, 0.23, 0.29, and 0.40 mm, and two lenses
CLMA Board was additionally concerned with advertisements of were supplied by the CLMA at each nominal thickness per mate-
Dk values based on a gas-to-gas method, which without calibration rial. Lenses were intended to be of uniform thickness, such that the
yielded higher results.8 At that time, the CLMA Board formed an thicknesses were nearly constant across the central areas of the
ad hoc committee to review the situation and formulate recom- lenses. The central thickness of each lens was measured indepen-
mendations based on its findings. After receiving the ad hoc review dently four times with a dial gauge capable of individual readings
at its October 1998 meeting, the Board authorized a grant to the 0.0025 mm, and the mean was taken as the true thickness of the
Eye Physiology & Ocular Prosthetics Laboratory at the University lens. The lens of each pair that was closest in terms of thickness to
of Alabama at Birmingham. The purpose was to derive the Dk the nominal value was selected for determination of oxygen per-
values of the rigid contact lens materials that were being marketed meability (Dk) unless the thickness across the central surface ap-
in the United States using the ANSI Z80.20/ISO 9913-1 polaro- peared to be nonuniform. In addition, the thicknesses of lenses
graphic method. A Technical Affairs Committee was chosen to made of each of seven reference materials were measured and se-
administer the study for the CLMA. The results of this investiga- lected in the same manner. Therefore, the resulting Dk values in
tion are those in this report. this report were derived from measurements of six lenses per test
and reference material in those nominal thicknesses noted above
(four recordings per lens, or 24 measurements per material). The
METHODS reference contact lenses were assessed polarographically for oxygen
transport in the same sessions and in the same manner as were the
After compiling a listing of the available RGP materials, the test contact lenses.
rigid material manufacturers throughout the U.S. were notified of The test and reference lenses of a similar thickness (the same
the proposed study and were asked to supply the names of com- nominal thickness) within each set were evaluated in a single ses-
mercial sources for the materials. Buttons of rigid materials were sion, and four sessions were completed for every thickness series.
gathered from the various sources and forwarded to the CLMA The measurements were performed according to a schedule de-
Technical Affairs Committee for coding. The buttons were placed signed to spread the potential effects of machine drift and alter-
in sealed bags, assigned a coded number for each material, and ations of ambient conditions within each set. Preliminary amper-
forwarded in sets, according to expected Dk, to a single contact ages were obtained with a polarographic sensor having a 4-mm-
lens finishing laboratory (Abba Optical, Stone Mountain, Geor- diameter cathode with a 7.80-mm radius of curvature (Rehder
gia). The finishing laboratory then produced all of the test contact Development Company, Castro Valley, California) connected
lenses to be used for the study from the buttons that were supplied electronically to an O2 Permeometer Model 201T (Createch Inc.,
under code. For each material, two lenses were made in each of six Albany, California). The preliminary amperages were achieved at
different thicknesses from a single lot of material buttons. The 35C in a water-saturated humid atmosphere containing 20.9%
lenses were in a monocurve back-surface design having a base curve oxygen. They were the basis for computation of preliminary oxy-

Optometry and Vision Science, Vol. 79, No. 2, February 2002


Oxygen Permeability of Contact LensesBenjamin and Cappelli 105
TABLE 1.
Permeabilities and confidence limits for lens set A.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

AA1 Boston II Test 16.3 13.521.0


AB2 Paraperm O2 Test 15.7 13.918.3
AC3 OP-2 Test 17.5 15.520.3
AB4 Optacryl 60 Test 13.8 12.116.4
AA5 Boston ES Test 27.3 22.934.1
AA6 Boston IV Test 20.8 18.523.8
AD7 Polycon II Test 14.1 11.718.5
AE8 SGP 1 Test 14.9 13.417.0
Silafocon A (Polycon II) Reference 10.0 8.712.2
Paflufocon C (Fluoroperm 30) Reference 26.0 23.429.4
a
Each of the oxygen permeability (Dk) values reported for the eight coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as occurred for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

TABLE 2.
Permeabilities and confidence limits for lens set B.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

BA1 Boston RXD Test 31.9 29.035.4


BF2 Accucon Test 19.6 17.622.0
BB3 Paragon Thin Test 23.0 21.924.2
BG4 Fluorex 300 Test 23.9 22.126.0
BB5 Fluoroperm 30 Test 30.3 28.831.9
BC6 OP-3 Test 27.1 24.230.9
BH7 O-Perm 30 Test 31.9 30.333.6
BE8 SGP 2 Test 31.9 30.133.8
Silafocon A (Polycon II) Reference 10.0 9.310.8
Itafluorofocon A (Equalens I) Reference 51.2 49.053.7
a
Each of the oxygen permeability (Dk) values reported for the eight coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as followed for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

gen transmissibility (Dk/t) values according to ANSI Z80.20- basis of the range of Dk values expected by the CLMA within the
1998,10 equivalent to ISO 9913-1:1996,11 as well as for subse- set and were identified for inclusion in measurements of the set.
quent corrections for the boundary-layer effect and the edge One reference material was to have been of a lower Dk value than
effect.10, 11 Oxygen permeability (Dk) of each of the test materials the test materials in the set, and the second reference material
was then calibrated using a linear equation derived from the deter- higher than the test materials in the set. Test materials having a
mined and established standard Dk values of the reference resultant Dk falling between the two reference materials were cal-
materials. ibrated using a linear equation derived from the measured (uncali-
With the exception of lens set E, lenses of two reference mate- brated) Dk and the established standard Dk6, 7 of the two reference
rials were measured with each set for the purpose of calibration. materials. The large range of potential Dk values represented in
The reference materials for each set were determined by the chair- lens set E necessitated the use of four reference materials. The
man of the CLMA Technical Affairs Committee (QAC) on the calibration of set E, therefore, was performed using a linear equa-

Optometry and Vision Science, Vol. 79, No. 2, February 2002


106 Oxygen Permeability of Contact LensesBenjamin and Cappelli

TABLE 3.
Permeabilities and confidence limits for lens set C.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

CE1 SGP 3 Test 33.5 31.835.4


CI2 Transaire Test 19.7 19.120.5
CJ3 Alberta 45 Test 28.0 23.834.0
CA4 Equalens I Test 67.9 62.773.9
CA5 Boston 7 Test 68.3 64.572.5
CG6 Fluorex 500 Test 29.6 28.331.0
CB7 Paraperm EW Test 49.5 46.652.8
Paflufocon C (Fluoroperm 30) Reference 26.0 24.427.8
Tolofocon A (Menicon EX) Reference 62.4 57.268.5
a
Each of the oxygen permeability (Dk) values reported for the seven coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as followed for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

TABLE 4.
Permeabilities and confidence limits for lens set D.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

DK1 Novalens (??) Test 17.8/33.6d NA


DB2 Paragon HDS Test 39.3 37.341.5
DA3 Boston EO Test 57.6 53.762.1
DB4 Fluoroperm 60 Test 42.7 41.444.0
DC5 OP-6 Test 57.0 51.963.2
DD6 Fluorocon Test 44.0 40.448.3
DH7 Operm 60 Test 56.5 51.163.2
Paflufocon C (Fluoroperm 30) Reference 26.0 24.328.0
Siflufocon B (Quantum 2) Reference 92.9 84.0104.2
a
Each of the oxygen permeability (Dk) values reported for the seven coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as followed for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.
d
Apparently, the six test lenses labeled DK1 were actually three lenses of each of two different materials. A different thickness
series of this material was delivered and tested under the code DK1A in lens set F. See text and Fig. 3 for details.

tion established by the least squares method involving the mea- nominal thickness were determined, was 10 for lens sets A and B;
sured and standard Dk values of four reference materials. There nine for lens sets C, D, and F; and 11 for lens set E. Six nominal
were several cases in which the test material resulted in a lower Dk thicknesses per material were represented in each set of lenses.
than had been expected by the CLMA, such that the test Dk was Because every lens was tested four times, in four different sessions
less than that of the lowest reference material measured with a set of according to the schedules accomplished, there were 240 prelimi-
lenses. In these instances, the calibration was performed using a nary amperages performed for lens sets A and B; 216 for lens sets C,
linear equation derived from the measured and standard Dk values D, and F; and 264 for lens set E. Six dark current readings were
of the lowest reference material and zero (the dark current). also performed per lens set. There were a total of 1428 preliminary
The number of lenses measured in a single session, during which amperages determined overall, with nearly double that number of
the preliminary amperages of test and reference lenses of the same central lens thickness measurements. This masked study consti-

Optometry and Vision Science, Vol. 79, No. 2, February 2002


Oxygen Permeability of Contact LensesBenjamin and Cappelli 107
TABLE 5.
Permeabilities and confidence limits for lens set E.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

EJ1 Alberta S Test 24.6 23.026.4


EG2 Fluorex 700 Test 41.3 40.342.3
EA3 Equalens II Test 114.6 103.8127.9
EB4 Fluoroperm 92 Test 64.0 60.268.5
EL5 Menicon SF-P Test 138.5 126.0153.8
EB6 Fluoroperm 151 Test 99.3 91.6108.3
EL7 Menicon Z Test 175.1 154.9201.6
Itafluorofocon A (Equalens I) Reference 51.2 49.553.0
Siflufocon B (Quantum 2) Reference 92.9 88.098.4
Melafocon A (Menicon SF-P) Reference 133.6 117.2155.7
Tisilfocon A (Menicon Z) Reference 161.1 142.1186.2
a
Each of the oxygen permeability (Dk) values reported for the seven coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as followed for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

TABLE 6.
Permeabilities and confidence limits for lens set F.

Corrected, Calibrateda
95% Confidence
Code Lens Name Test/Reference Oxygen Permeability
Limitsc
(Dk, in ANSI Units)b

CA4 Equalens I Test 55.0 49.861.2


CA4A Equalens I Test 51.2 47.655.4
CA5 Boston 7 Test 57.0 51.364.1
CA5A Boston 7 Test 55.8 51.361.3
DK1A NovaLens Test 18.7 17.320.4
EA3 Equalens II Test 93.0 82.5106.7
EA3A Equalens II Test 93.1 85.9101.7
Paflufocon C (Fluoroperm 30) Reference 26.0 25.027.1
Siflufocon B (Quantum 2) Reference 92.9 81.8107.6
a
Each of the oxygen permeability (Dk) values reported for the seven coded test materials in this table were the result of 24
independent evaluations of oxygen transmissibility (Dk/t). These values were corrected for the boundary-layer and edge effects, then
calibrated according to the reference materials noted. The reference materials were measured in the same manner and in the same
sessions using the same schedule as followed for the masked test materials.
b
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
c
95% Confidence limits were derived from 2 standard errors (N 6) surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

tuted the largest and most thorough polarographic analysis of ox- rounding the slope of the resistance (t/Dk) vs. thickness (t) regres-
ygen permeability yet performed for clinically available rigid con- sion line computed using the edge-corrected resistances of the six
tact lens materials. lenses measured for each of the test materials. The lower confi-
dence limits for the test and reference materials ranged from 2.4 to
17.0% of the Dk value, averaging 8.4% (SD 3.50, N 57) with
RESULTS a median of 8.1%. The higher confidence limits ranged from 2.4 to
The corrected, calibrated oxygen permeability (Dk) values are 31.2% of the Dk value, averaging 10.9% (SD 6.17, N 57)
shown in Tables 1 through 6 corresponding to lens sets A to F. The with a median of 9.8%. The slope of each edge-corrected linear
Dk values are in ANSI units of 1011 (cm2/s)(ml O2)/(mlmm resistance vs. thickness equation was 1/Dk, the inversion of which
Hg). Confidence limits at 95% for each Dk value are also listed. was the corresponding Dk that, after calibration, was reported in
These limits were derived from 2 standard errors (N 6) sur- Tables 1 through 6. The intercept was a measure of the resistance

Optometry and Vision Science, Vol. 79, No. 2, February 2002


108 Oxygen Permeability of Contact LensesBenjamin and Cappelli

of the boundary layer imposed between the oxygen sensor and the
back of the test contact lens. The slope and the intercept were
independent of test lens thickness. The mean resistance (t/Dk) of
the boundary layer encountered in this investigation was 0.00386
109 (mlmmHg)/(cm/s)(ml O2); (SD 0.00166, N 57).
Seven representative resistance (t/Dk) vs. thickness (t) plots are
shown in Fig. 1 across the range of Dk values from 10.0 (reference
Polycon II) to 175.1 (Material EL7, or test Menicon Z). The plots
of t/Dk vs. t for the materials approached linear relationships, and
the plots in Fig. 1 are demonstrative. Forty-eight of the linear
coefficients of determination (R2) were 0.98, and 54 were
0.96. The lowest of the 57 R2 values was 0.9285 (Material AD7,
or test Polycon II). Materials AA1 (Boston II, R2 0.9373) and
AA5 (Boston ES, R2 0.9550) were the only remaining materials
with computed R2 values 0.96. The plots for the three materials
having the highest Dk values (Menicon Z test and reference mate-
rials, and test Menicon SF-P), in Fig. 2, revealed that linear regres- FIGURE 2.
sion was also descriptive up to the highest Dk values found in this Resistance (t/Dk) vs. thickness (t) plots for the three materials in this study
that resulted in the highest Dk values. Note, again, that the plots were
study (R2 0.98 in all three cases). Any deviations from linear
closely approximated by linear relationships (R2 0.98).
relationships could have been due to measurement error, thickness
nonuniformity, and/or polymer variability between the six test
lenses of each material. of Dk, although the lenses were to have been derived from a single
The 2 Dk values for Material DK1 in Table 4 regarding Lens Set lot of buttons. Because the identity of the material(s) composing
D revealed an extreme example of polymer variability. The re- the DK1 lenses was in question, a second group of 12 lenses of this
sistance (t/Dk) vs. thickness (t) plot for Material DK1 is shown in material was obtained by the CLMA and delivered under the code
Fig. 3. The six lenses selected for measurement were apparently DK1A. These lenses became part of lens set F, and the results for
split between two different materials varying significantly in terms Material DK1A (NovaLens) are reported in Table 6. It is the Dk
value in Table 6 (18.7 Dk units) that should be accepted for the
NovaLens material. The confidence limits for this Dk value (17.3
to 20.4) encompassed one of the estimates in Table 4 (17.8) based
on three of the original DK1 test lenses. The resistance vs. thick-
ness plot for Material DK1A in Fig. 3 revealed a marked similarity
to that of three of the original DK1 test lenses.
Of further note are the Dk values for Materials CA4 (Equalens
I), CA5 (Boston 7), and EA3 (Equalens II) in Tables 3 and 5

FIGURE 3.
FIGURE 1. Resistance (t/Dk) vs. thickness (t) plots for the six test lenses measured
Representative resistance (t/Dk) vs. thickness (t) plots for seven of the 57 according to the code DK1 and the six lenses that were coded DK1A.
test and reference materials assessed in this study, ranging from Dk 10.0 Three of the DK1 lenses were apparently made of the Novalens material
to Dk 175.1. Note that the plots were closely approximated by linear (DK1A). The other three lenses were made of an unidentified polymer
relationships (R2 0.96). (labeled DK1(?) on the graph).

Optometry and Vision Science, Vol. 79, No. 2, February 2002


Oxygen Permeability of Contact LensesBenjamin and Cappelli 109

regarding Lens Sets C and E. These Dk values were thought to DISCUSSION


be high when reported to the manufacturer (Polymer Technology
Several aspects of the determination of oxygen permeability
Corporation, Boston, Massachusetts), relative to the Dk values
(Dk) in this study were accomplished to enhance the validity of the
that had been determined by that company, as communicated to
resultant Dk values. The Dk values were corrected for the bound-
the CLMA Technical Affairs Committee. The company requested
ary-layer effect and the edge effect according to ANSI Z80.20-
that the committee have these three materials remeasured and sup-
1998 and ISO 9913-1:1996,10, 11 and they were calibrated using a
plied another 12 buttons of each material from a different lot
set of reference materials whose Dk values had been previously
number to be additionally tested. The total of six materials were
established.6 8 The identities of the test materials were masked
then measured within lens set F. The results of the retest on mate-
until after the Dk values had been determined. Lenses of six dif-
rials CA4, CA5, and EA3, and for the additional three materials
ferent uniform thicknesses were assessed for each material com-
CA4A, CA5A, and EA3A are found in Table 6 regarding Lens Set
pared with four required in the ANSI/ISO standards and used in
F. The Dk values for CA4 and CA5 upon retesting remained
most previous Dk reports. Four preliminary amperages were de-
higher than predicted by the manufacturer. The values for CA4A
termined for each lens thickness and recorded according to a
and CA5A were similar to those of the retest. Hence, the Dk value schedule designed to spread the effects of machine drift and alter-
of the Equalens I test material (CA4 and CA4A) was established as ations of ambient conditions across each of the six sets of materials
the mean of three resultants, or 58.0 Dk units, and the Dk value of that were evaluated. The reference materials were assessed in the
the Boston 7 material (CA5 and CA5A) was established as the same sequences during the same sessions according to the schedule
mean of three resultants, or 60.4 Dk units. that was used for the test materials.
The original Dk value of material EA3 in Lens Set E (114.6 An earlier study involving one of the authors determined 16 Dk
Dk units) is difficult to explain and could be labeled an outlier. An values according to the polarographic ANSI and ISO standards.8
outlier from 43 test Dk determinations (discounting material This initial study involved masking, calibration, and the measure-
DK1) and 14 additional reference determinations in this study ment of seven contact lenses per material in the nominal uniform
might have been expected. The retest Dk value for EA3 was 93.0 thicknesses chosen for the current study and 0.09 mm. Table 7
Dk units and for the additional lot of EA3A, the Dk was 93.1 Dk compares the Dk values found in the current study to those found
units, values that were still higher than predicted by the manufac- for materials of the same name (although of different lot numbers)
turer. The mean of the three determinations was 100.2 Dk units. in that earlier investigation.8 The Dk values for that study were
The most logical course was to establish the Dk of the Equalens II corrected for the edge effect and boundary-layer effect, as in the
material (EA3 and EA3A) as the mean of the three resultants, or current study, and the permeability values had been calibrated
100.2 Dk units. There were no other requests for retesting, al- according to the reference materials used in the current study. The
though several of the Dk values in Tables 1 through 6 were signif- difference percentages showed that of the 14 pairs of correspond-
icantly lower than those advertised by their manufacturers. ing test materials that overlapped between the two studies, only

TABLE 7.
Comparison of corrected and calibrated Dk values (reference values in parentheses).
Corrected, Calibrated Corrected, Calibrated
Lens Name Oxygen Permeability Difference Oxygen Permeability
(Current Study) (Previous Study8)

Boston II 16.3 21.0% 13.2


Boston IV 20.8 2.4% 21.3
Boston 7 60.4 6.1% 56.8
Boston ES 27.3 15.8% 23.3
Equalens I 58.0 21.8% 46.4 (51.2)b
Fluorex 700 41.3 3.1% 42.6
Fluoroperm 30 30.3 3.7% 29.2 (26.0)b
Fluoroperm 60 42.7a 8.5% 46.5
Fluoroperm 90 64.0 9.4% 70.3
Fluoroperm 151 99.3 5.3% 104.7
Menicon SF-P 138.5 9.4% 126.1 (133.6)b
Menicon Z 175.1 8.7% 160.4 (161.1)b
Polycon II 14.1 38.0% 9.6 (10.0)b
SGP 3 33.5 2.9% 31.6
a
Fluorocon and Paragon HDS, made of similar materials to that of Fluoroperm 60, resulted in Dk values of 44.0 and 39.3 units,
respectively.
b
The reference lots of Equalens I, Fluoroperm 30, Menicon SF-P, Polycon II were established at 51.2, 26.0, 133.6, and 10.0 Dk units,
respectively, and the reference lot of Menicon Z was tentatively established at 161.1 units. Reference Dk values are in parentheses.

Optometry and Vision Science, Vol. 79, No. 2, February 2002


110 Oxygen Permeability of Contact LensesBenjamin and Cappelli

FIGURE 4.
FIGURE 5.
Resistance (t/Dk) vs. thickness (t) plots for the reference Polycon II mate-
A plot of the Dk values of the 14 test materials shown in Table 7, those that
rial, test Polycon II material (AD7), test Boston II material (AA1), and test
overlapped between the current investigation of 37 test materials and an
Boston ES material (AA5). The three test materials were those having
earlier investigation involving 16 test materials. The regression line was
coefficients of determination (R2) 0.96 [0.9285, 0.9373, and 0.9550,
almost the same as that of a 1:1 relationship, having a slope near unity
respectively] when all six data points were included in the linear regres-
(1.056), intercept near zero (0.292), and R2 0.9846.
sions (solid lines). Note the deviation of the single data point at 0.40 mm
away from a straight line formed by the other five points (dashed lines) for
each of the three test materials compared with the linearity of the refer- and Boston II materials were those that resulted in the lowest linear
ence material (R2 0.9949).
coefficients of determination in this study (R2 0.9285 and
0.9383, respectively). The only other R2 value below 0.96 was that
three pairs fell outside the 20% tolerance specified in ANSI Z80.20 of the Boston ES material (R2 0.9550). The resistance (t/Dk) vs.
(Polycon II, Boston II, and Equalens I). The difference between thickness (t) plots for the reference Polycon II, test Polycon II
the Dk values of the Boston ES material was 15.8%. The differ- (AD7), Boston II (AA1), and Boston ES (AA5) materials measured
ences in Table 7 were 10% for the overwhelming majority (the in the current study are shown in Fig. 4. The reader can see that the
other 10) of the 14 overlapping test materials. data of the three test materials did not follow linear relationships
It is not yet known how Dk may sometimes vary between lots of (the solid lines in Fig. 4) as well as did those of the Polycon II
a contact lens material or even between buttons taken from a single reference material. Apparently, one way in which a high Dk value
lot. It was already established, above, that the mean Dk value can be obtained is for the data to be scattered away from a linear
associated with the current Equalens I test material (58.0) was oxygen resistance vs. thickness relationship, having an R2 0.96,
greater than of the reference material (51.2). The test Polycon II such that the slope of the regression line (1/Dk) is artificially low.

TABLE 8.
Deletion of single outlying points.

Corrected, Calibrated Oxygen Permeability


(Dk, in ANSI Units)a
Code Test Lens Name
Before Deletion After Deletion
(N 6) (N 5)

AA1 Boston II Dk 16.3 13.8


95% CIb 13.521.0 11.318.2
R2 0.9373 0.9425

AA5 Boston ES Dk 27.3 22.6


95% CIb 22.934.1 19.626.9
R2 0.9550 0.9768

AD7 Polycon II Dk 14.1 11.4


95% CIb 11.718.5 10.013.7
R2 0.9285 0.9696
a
American National Standards Institute (ANSI) Dk units are 1011 (cm2/s)(ml O2)/(ml mm Hg).
b
95% Confidence intervals were derived from 2 standard errors surrounding the edge-corrected slope of the linear resistance
(t/Dk) vs. thickness (t) equation used to correct for the boundary-layer effect for each material.

Optometry and Vision Science, Vol. 79, No. 2, February 2002


Oxygen Permeability of Contact LensesBenjamin and Cappelli 111

One method of circumventing this problem would be to allow Hence, it is hoped, published and advertised Dk values of RGP
the deletion of an obvious single outlying point from determina- contact lens materials will become more reliable and clinically rel-
tion of a regression line. Using this statistical technique to omit the evant as the use of standardized Dk methodology involving error
outlying data point in Fig. 4 for the test Polycon II (AD7) material, correction and calibration becomes more widespread.
the deleted data point was associated with the nominal thickness of
0.40 mm. The recomputed coefficient of determination (R2) was
then 0.9696 for the dashed line associated with the test Polycon II ACKNOWLEDGMENTS
(AD7) material shown in Fig. 4. The corrected and calibrated We thank the Contact Lens Manufacturers Association, Southern Council of
oxygen permeability by this analysis became 11.4 Dk units, and the Optometrists, and American Optometric Association for support of this study.
95% confidence limits ranged from 10.0 to 13.7 Dk units (Table Received May 1, 2001; revision received September 24, 2001.
8). By omission of a single outlying point in the same fashion, the
R2 values for the resistance (t/Dk) vs. thickness (t) plots became REFERENCES
0.9425 and 0.9768 for the Boston II (AA1) and Boston ES (AA5)
materials, respectively, represented by the appropriate dashed lines 1. Benjamin WJ. Oxygen transport through contact lenses. In: Guillon
M, Ruben CM, eds. Contact Lens Practice. London: Chapman &
in Fig. 4. The oxygen permeability became 13.8 Dk units (11.3 to
Hall, 1994:3966.
18.2) and 22.6 Dk units (19.6 to 26.9), respectively. Hence, the 2. Benjamin WJ. EOP and Dk/L: the quest for hyper transmissibility.
Dk values of these three test materials better matched those found J Am Optom Assoc 1993;64:196200.
in the earlier study (Table 7) after the single outlying points were 3. Fatt I, Chaston J. Measurement of oxygen transmissibility and per-
deleted from the analysis (Table 8). meability of hydrogel lenses and materials. ICLC 1982;9:7688.
A linear regression between the 14 pairs of Dk values in Table 7, 4. Fatt I, Rasson JE, Melpolder JB. Measuring oxygen permeability of
shown in Fig. 5, revealed a very high correlation (R2 0.9846), gas permeable hard and hydrogel lenses and flat samples in air. ICLC
slope of nearly unity (slope 1.056), and intercept near zero 1987;14:389401.
(intercept 0.292). Hence, the reproducibility of edge-cor- 5. Fatt I, Ruben CM. Oxygen permeability of contact lens materials: a
rected, boundary layer-corrected, calibrated Dk values using the 1993 update. J BCLA 1994;17:115.
polarographic ANSI/ISO technique was demonstrated for the 6. Benjamin WJ, Ho A, Winterton LC, Nakada K. The oxygen perme-
ability of reference materials. Optom Vis Sci 1997;74(Suppl):95.
overwhelming majority of RGP materials up to at least 160 Dk
7. Benjamin WJ, Ho A, Winterton LC, Nakada K. Calibration of O2
units. The linearity of the resistance (t/Dk) vs. thickness (t) plots permeability: reference materials. Invest Ophthalmol Vis Sci 1998;
gave further testimonial to the validity of the polarographic mea- 39:S874.
surements up to a corrected, calibrated Dk value of at least 175 Dk 8. Benjamin WJ, Wilson GS. Calibration of 22 material Dk values from
units. 2 different sites. Invest Ophthalmol Vis Sci 1999;40:S906.
9. Benjamin WJ. Wiggle room and the transitional Dk statistic. ICLC
1998;25(7&8):11820.
CONCLUSION
10. American National Standards Institute. ANSI Z80.20, Standards for
Oxygen permeability (Dk) of 37 RGP contact lens materials Ophthalmic Optics. Contact Lenses: Standard Terminology, Toler-
available in the United States ranged from 13.8 to 175.1 Dk units ances, Measurements and Physicochemical Properties. Merrifield,
[ 1011 (cm2/s)(ml O2)/(mlmm Hg)]. The mean and median VA: The Optical Laboratories Association, 1998:9, 167, 34, 4357.
Dk values were 43.2 and 31.9 units, respectively, and the standard 11. International Standards Organization. ISO 9913-1:1996. Optics and
Optical Instruments: Contact Lenses. Part 1. Determination of Ox-
error was 6.57 (N 37). The 95% confidence limits around the
ygen Permeability and Transmissibility with the FATT method. Ge-
material Dk values were generally less than 10% when six con- neva, Switzerland: International Standards Organization, 1996:
tact lenses made of the same material in different thicknesses were 113.
tested using a polarographic method in conformance with ANSI 12. International Standards Organization. ISO 9913-2:2000. Optics and
Z80.20-1998 and ISO 9913-1:1996. Boundary-layer and edge Optical Instruments: Contact Lenses. Part 2. Determination of Ox-
effects were corrected, and calibrations were performed by similar ygen Permeability and Transmisibility by the Cuolometric Method.
testing of reference lenses. Although the effects of polymer vari- Geneva, Switzerland: International Standards Organization, 2000:
ability between lots and within lots of contact lens materials are yet 111.
unexplored, the product tolerance for Dk values established in
ANSI Z80.20-1998 (20%) was verified in 11 of 14 materials William J. Benjamin
that had previously been measured using the ANSI/ISO polaro- University of Alabama at Birmingham
graphic method. The Dk values of two of the remaining three School of Optometry/HBP Building
materials were brought into conformity with the product tolerance 1716 University Blvd.
when single outlying points were graphically identified and deleted Birmingham, Alabama 35294-0010
from the linear resistance (t/Dk) vs. thickness (t) regressions. e-mail: wbenjamin@icare.opt.uab.edu

Optometry and Vision Science, Vol. 79, No. 2, February 2002