Q 2005 AS

Published by

J CATARACT REFRACT SURG - VOL 31, NOVEMBER 2005

Comparison of Colvard pupillometer

and infrared digital photography

for measurement of the dark-adapted

pupil diameter

Jay C. Bradley, MD, Justin E. Anderson, MD, Ke Tom Xu, PhD, Sandra M. Brown, MD

PURPOSE: To investigate the accuracy of pupil diameter measurement using the Colvard pupillometerand to determine the learning curve for inexperienced examiners.

SETTING: Texas Tech University Health Sciences Center, Lubbock, Texas, USA.

METHODS: In this population study, subjects with normal pupillary behavior were tested by 1 of 2 in-vestigators (examiner A, examiner B). After 5 minutes of dark adaptation at 1 lux, digital infrared pupilphotography of the right eye was performed, followed by measurement of the horizontal pupil diam-eter and vertical pupil diameter with the Colvard pupillometer. The photographs were digitally ana-lyzed to determine the horizontal and vertical pupil diameters. During phase I of the study,examiners were masked to the results of infrared pupil photography; during phase II, they reviewedthe infrared pupil photography results after each testing session. Bland-Altman plots were createdto detect measurement bias; results were graphed by subject test sequence to assess learning. Atest difference of less than G0.5 mm was considered clinically acceptable.

RESULTS: Fifty-nine subjects were tested in phase I, of whom 39 had adequate infrared pupil photog-raphy for analysis; 40 were tested in phase II, of whom 34 were included. Themean age of the analyzedsubjects was 27 years (range 18 to 44 years). For all subjects, the infrared pupil photography medianhorizontal pupil diameter was 7.09 mmG 0.75 (SD) (range 5.44 to 8.79 mm); the median vertical pupildiameter was 7.22 G 0.79 mm (range 5.45 to 9.10 mm). Examiner A initially had a negative bias (Col-vard pupillometer value less than infrared pupil photography value) for both horizontal and verticalpupil diameter measurements, which resolved during phase I after 23 subjects were tested; 18 ofthe final 19 subjects tested (11 phase I, 8 phase II) showed a test difference of less than 0.5 mm forall readings. The pupil diameter did not affect the bias. Examiner B had a strong positive bias that per-sisted throughout the study. Testing 26 subjects in 5 sessions during phase II did not improve theaccuracy. During the final testing session, 3 of 8 subjects had a test difference of 0.5 mm or more inat least 1 dimension. The pupil diameter did not affect the bias.

CONCLUSION: The Colvard pupillometer is susceptible to user errors causing unidirectional bias andseems to have a steep and variable learning curve.

J Cataract Refract Surg 2005; 31:2129–2132 Q 2005 ASCRS and ESCRS

Because of its relatively low cost, portability, and appar-

ently simple function, the Colvard pupillometer (Oasis

Medical Inc.) is probably the most frequently used

pupil-measurement device in the field of refractive sur-

gery. The goal of preoperative pupil diameter evaluation

is not simply to measure the pupil in the dark, but rather

to determine the patient’s physiologic dark-adapted

pupil diameter; that is, the largest diameter the pupil

CRS and ESCRS

Elsevier Inc.

is likely to achieve in a very-low-light environment un-

der binocular, ‘‘free viewing’’ conditions at distance fixa-

tion. Unless the examiner attends carefully to full dark

adaptation, assures patient alertness, instructs the patient

to maintain distance fixation to prevent accommodation,

and minimizes parallax, no measurement taken with the

Colvard pupillometer (or any other method) will be

accurate.

0886-3350/05/$-see front matterdoi:10.1016/j.jcrs.2005.04.041

2129

MEASUREMENT OF THE DARK-ADAPTED PUPIL DIAMETER

Table 1. Infrared pupil photograph measurements of all subjects.

Examiner A Examiner B All Subjects

HPD VPD HPD VPD HPD VPD

Number of subjects 29 27 44 42 73 69Mean (mm) 6.88 7.02 7.16 7.32 7.05 7.20Median (mm) 6.93 7.14 7.21 7.27 7.09 7.22Range (mm) 5.44–8.47 5.45–8.60 5.70–9.10 5.58–9.53 5.44–9.10 5.45–9.53SD (mm) 0.69 0.71 0.78 0.82 0.75 0.79

HPD Z horizontal pupil diameter; VPD Z vertical pupil diameter

Values refer to pupil diameters measured after a minimum of 5 minutes of dark-adaptation at a constant environmental illumination of 1 lux.

In this study, we sought to determine how often and by

how much the results of Colvard pupillometer testing

would differ from infrared pupil photography. We consid-

ered a deviation of more than 0.5 mm in either direction

to be clinically unacceptable if the measurement was tobe used for refractive surgery.

SUBJECTS AND METHODS

This study was approved by the institutional review board,and all subjects provided informed consent. Individuals with nohistory of eye disease or injury affecting pupil diameter were re-cruited from department staff, acquaintances of the investigators,and medical students. Refractive error was not determined. Thesubjects were nonrandomly recruited and tested by 1 of 2 investi-gators (examiner A [J.C.B.], a first-year ophthalmology resident;examiner B [J.E.A.], an intern who will be an ophthalmology res-ident), neither of whom had previous experience with the Colvardpupillometer; their only instruction was in the user’s guide pro-vided by Oasis Medical.

Participants donned wraparound ‘‘cataract’’ sunglasses fora minimum of 5 minutes; ambient illumination was not controlledduring this period. The subject was then formally dark-adapted at1 lux for 5 minutes. After adaptation, multiple still digital infraredphotographs of the right pupilwere takenusing experimental tech-niques previously described.1,2 The subject was instructed to gazeat the far end of the room ‘‘as though looking into the distance.’’ Af-ter photography, the horizontal and vertical pupil diameters of theright eye were measured using the Colvard pupillometer; the sub-ject was instructed to avoid looking at the examiner or the pupill-ometer and to continue to gaze at a far wall. The left eye was notoccluded. The examiners were careful to minimize parallax.

The clearest infrared photograph was selected for analysis.Using digital image software, the horizontal and vertical pupil

Accepted for publication April 11, 2005.

From the Cabarrus Eye Center (Brown), Concord, North Carolina,and the Departments of Ophthalmology and Visual Sciences(Bradley, Anderson) and Family and Community Medicine (Xu),Texas Tech University Health Sciences Center, Lubbock, Texas,USA.

No author has a financial or proprietary interest in any material ormethod mentioned.

Reprint requests to Sandra M. Brown, MD, 201 LePhillip Court NE,Concord, North Carolina 28025, USA. E-mail: sbrownmd@carolina.rr.com.

J CATARACT REFRACT SURG -2130

diameters were calculated in millimeters using a 5.0 mm segmentof the photographed ruler as the length standard. Each photo-graph was independently measured by 2 investigators (S.M.B.,J.C.B.) and the results averaged.

During phase I (masked), the results of infrared pupil pho-tography were not compared with those of the Colvard pupillom-eter until all testing was completed; this prevented biasedmodification of the examiners’ Colvard pupillometer technique.During phase II (unmasked), examiners compared infrared pupilphotography with Colvard pupillometer values at the conclusionof each day of testing to encourage learning.

RESULTS

Fifty-nine subjects were recruited for phase I. Exam-

iner A tested a total of 33 subjects, of whom21were included

in the final analysis; examiner B tested a total of 26 subjects,

of whom 18 were included. Exclusions were due to inade-

quate infrared photographs caused by the shallow depth offocus of the camera or upper lid shadowing that prevented

vertical pupil diameter measurement. Forty subjects were

recruited for phase II. Examiner A tested a total of 11 sub-

jects, of whom 8 were included; examiner B tested 29 sub-

jects, of whom 26 were included.

The mean age of all subjects was 27 years (range 18 to

44 years); 84% percent were in their third decade. The re-

sults of infrared pupil photography of all subjects are givenin Table 1. Themean infrared pupil photography horizontal-

to-vertical pupil diameter ratio of all subjects was 0.98.

Results are displayed graphically for examiner A (Figure 1)

and examiner B (Figure 2).

The Colvard pupillometer reticule extends to 13.0 mm

horizontally and vertically (Figure 3). Both examiners not-

ed that the reticule was suboptimal due to central crowding

from the graduations on the horizontal axis that obscuredthe pupil center and lack of 0.5 mm graduations on the ver-

tical axis.

DISCUSSION

In our study, 2 inexperienced but educated examinerswho understood accommodation and parallax tested a

cohort of cooperative subjects with the Colvard pupillom-

eter in an optimized experimental environment. When

VOL 31, NOVEMBER 2005

MEASUREMENT OF THE DARK-ADAPTED PUPIL DIAMETER

compared with infrared pupil photography, examiner A

showed a significant learning effect (Figure 1) and consis-

tently recorded Colvard pupillometer measurements withinG0.5 mm of the infrared pupil photography values after

testing approximately 20 subjects. Between subjects 22

and 32, there was 1 clinically significant outlier with Col-

vard pupillometer measurements almost 1.0 mm smaller

than infrared pupil photography values; the final 10 sub-

jects fell within the clinically acceptable range with no out-

liers. The Colvard pupillometer measurements of examiner

B were erratic (Figure 2), with most being larger than theinfrared pupil photography values; no improvement was

seen during phase II.

One previous study3 compared the Colvard pupillom-

eter with infrared videography and flash digital photogra-

phy. We have criticized the results4 because the reported

mean dark-adapted pupil diameter values (video 6.28 mm,

photograph 6.24 mm) seemed too small for a study co-

hort with a mean age of 36 years. This study also used 2 ex-aminers who found mean horizontal pupil diameter

Colvard pupillometer values of 6.21 mm and 5.84 mm, re-

spectively, after dark adaptation at less than 0.63 lux for

Bland Altman Graph

-1.50

-1.00

-0.50

0.00

0.50

1.00

5.50 6.00 6.50 7.00 7.50 8.00 8.50Mean of CP and IRPP (mm)

CP

m

in

us

IR

PP

(m

m)

Learning Effect

-1.50

-1.00

-0.50

0.00

0.50

12 17 22 27 32 37 42Subjects by Test Sequence

CP

m

in

us

IR

PP

(m

m)

HPD masked HPD unmasked VPD masked VPD unmasked

Figure 1. Results for examiner A. After testing approximately 22 subjects

while masked to the infrared photography results, this examiner achieved

a consistent clinical test agreement ofG0.5 mm. Additional practice with

feedback on the infrared pupil photography results at the end of each

testing session did not improve his performance. The Bland-Altman graph

shows that bias was not a function of pupil diameter. (CP Z Colvard pu-

pillometry; IRPP Z infrared pupil photography; HPD Z horizontal pupil

diameter; VPD Z vertical pupil diameter; masked Z no knowledge of

the result of infrared pupil photography testing; unmaskedZ knowledge

of the result of infrared pupil photography testing).

J CATARACT REFRACT SURG -

an unspecified length of time. These mean values are toosmall5,6 for this population and could not represent mea-

surement of the true dark-adapted pupil diameter.

Based on the results of our experiment, several con-

clusions can be reached regarding the use of the Colvard

pupillometer for measurement of the dark-adapted pupil

diameter under typical clinical circumstances. Foremost,

the Colvard pupillometer is only as good as the user,

and the inherent design of the device does not act to pre-vent the introduction of strong unidirectional bias. Thus,

when Colvard pupillometer measurement is delegated to

support staff or junior residents, it is important that they

be properly trained and their accuracy assessed. Second,

it is possible to obtain accurate measurements with the

Colvard pupillometer. However, measurement errors

greater than 0.5 mm can occur due to user error resulting

from (1) failure to properly dark-adapt the patient, (2)failure to recognize accommodative miosis or decreased

alertness, (3) failure to avoid parallax, (4) failure to read

the reticule precisely, or (5) inaccurate recollection of

the 4 reticule readings and/or incorrect subtraction. Third,

because even experienced and generally accurate

Bland Altman Graph

-1.00

-0.50

0.00

0.50

1.00

1.50

5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50Mean of CP and IRPP (mm)

CP

m

in

us

IR

PP

(m

m)

Learning Effect

-1.00

-0.50

0.00

0.50

1.00

1.50

4 14 24 34 44 54Subject Test Sequence

CP

m

in

us IR

PP

(m

m)

HPD masked HPD unmasked VPD masked VPD unmasked

Figure 2. Results for examiner B. This examiner maintained a strong pos-

itive bias through both phases of the experiment. Additional practice with

feedback on the infrared pupil photography results at the end of each

testing session did not improve his performance. The Bland-Altman graph

shows that bias was not a function of pupil diameter. (CP Z Colvard pu-

pillometry; IRPP Z infrared pupil photography; HPD Z horizontal pupil

diameter; VPD Z vertical pupil diameter; masked Z no knowledge of

the result of infrared pupil photography testing; unmaskedZ knowledge

of the result of infrared pupil photography testing).

VOL 31, NOVEMBER 2005 2131

MEASUREMENT OF THE DARK-ADAPTED PUPIL DIAMETER

2132

Figure 3. Two alternate reticules for handheld pupillometers. Upper

Left: Photograph taken through the Colvard pupillometer showing

scale from zero to 13.0 mm with 0.5 mm graduations. Upper Right: In-

ner graduation mark is at 2.5 mm. Lower Left: Inner graduation mark

is at 2.0 mm and there is a pupil centering target (open circle). The retic-

ules match the scale of the external ruler. The reticules give readings of

7.0 mm horizontally and 7.5 mm vertically. (Reticule designs copyright

2004, Sandra M. Brown, MD)

examiners can be wrong (as evidenced by examiner A’s

subject 32), it is essential to be aware of population

mean dark-adapted pupil diameter values5,6 so an atypi-

cally small measurement (ie, 5.0 mm in a 36-year-old7

or 31-year-old8 patient) can be rechecked. Finally, because

the vertical pupil diameter is larger than the horizontal

pupil diameter in most patients,2 Oasis Medical should

give consideration to modifying the Colvard pupillometer

reticule (Figure 3) to allow easier performance of vertical

measurements.

REFERENCES

1. Brown SM, Khanani AM, Xu KT. Day to day variability of the dark-

adapted pupil diameter. J Cataract Refract Surg 2004; 30:639–644

J CATARACT REFRACT SURG -

2. Khanani AM, Brown SM, Archer SM. Horizontal versus vertical dark-

adapted pupil diameters in normal individuals. J Cataract Refract

Surg 2004; 30:2557–2558

3. Twa MD, Bailey MD, Hayes J, Bullimore M. Estimation of pupil size by

digital photography. J Cataract Refract Surg 2004; 30:381–389

4. Brown SM, Bradley JC, Khanani AM. Duration of pupillary constriction in

response to a photographic flash [letter]. J Cataract Refract Surg 2005;

31:455–456

5. BorgmannH. Grundlagen fur eine klinishe Pupillographie. II. Abhangig-

keit des Pupillendurchmessers in Dunkelheit vom Lebensalter. Al-

brecht von Graefes Arch Klin Exp Ophthalmol 1972; 184:300–308

6. Loewenfeld IE. Pupillary changes related to age. In: Thompson

HS, Daroff R, Frisen L, et al, eds, Topics In Neuro-Ophthalmology.

Baltimore, Williams & Wilkins, 1979; 124–150

7. Brown SM, Campbell CE. Systematic underablation in laser in situ ker-

atomileusis: ablation pattern identified by advanced topographical

analysis. J Cataract Refract Surg 2003; 29:1621–1625

8. Kohnen T. Consultation section: refractive surgery problem. J Cataract

Refract Surg 2004; 30:1392–1395

VOL 31, NOVEMBER 2005