Evaluating Moisture Migration in Schirmer Test Strips: Exploring Brand-Specific Variations and Introducing Calibration and Conversion Methods

Supplemental Digital Content is Available in the Text.

Abstract

Purpose:

Schirmer test results are widely used for ocular surface disease assessment, but Schirmer strips are not standardized. We compare the characteristics and tear volume with millimeter moisture migration in different brands of Schirmer strips and introduce methods for volume-based, brand-independent calibration.

Methods:

Physical parameters of Haag-Streit, EagleVision, TearFlo, Contacare, and MIPL/A6 Schirmer strip brands were compared. Schirmer strip millimeter moisture migration distances were assessed 5 minutes after application of incremental microliter volumes of human tears. Linear regression analysis of data points from each Schirmer strip brand was performed, and the root-mean-square deviation of data points to the best-fit linear regression was calculated. Calibration correction was performed by converting migration distance to the corresponding tear volume. A reference table and calibration method formulas were created.

Results:

Schirmer strips differed in design, shape, and manufacturing precision. Strip width, weight, and length were different between the 5 brands (P < 0.05). A wide range of Schirmer strip moisture migration values for identical tear volumes was observed among brands. Statistical measurement resulted in a root-mean-square deviation of 2.9 mm for all data points from all brands. Millimeter to volume and weight to volume-based calibration correction methods resulted in a 2.2- and 3.1-fold measurement error reduction, respectively.

Conclusions:

Our findings highlight the lack of standardization among different brands of Schirmer strips, raising concerns about potential sources of unintentional measurement errors. We propose volume-based Schirmer strip calibration methods and conversion of millimeter to microliter results to achieve brand-independent results and improve Schirmer test accuracy.

For over a century, the Schirmer test has been a widely used clinical tool for assessing tear production and volume.^1,2 This procedure evaluates the basal and reflex tearing capability of the eye by a placing filter paper strip over the temporal one-third of the lower lid margin of each eye with, or without (Schirmer I) anesthesia. The extent of moisture migration on the strip's external portion is then recorded in millimeters (mm).³

A score of ≤5 mm on the Schirmer I test indicates aqueous tear deficiency and is, for example, used as an objective diagnostic criterion for Sjögren syndrome by the American College of Rheumatology and the European League Against Rheumatism.^4–7 Similarly, a new onset of ocular sicca documented by a low Schirmer I test ≤5 mm is considered an organ-specific manifestation of chronic ocular graft versus host disease (oGVHD) after allogeneic hematopoietic transplantation as established by the National Institutes of Health Consensus Group.⁸ Furthermore, the International Chronic oGVHD Consensus Group also includes the Schirmer I test result as a component of their severity scale.⁹ The Schirmer I test is also established for the diagnosis of dry eye disease and is frequently used as an efficacy end point for therapeutics in interventional clinical trials.^10–15

Despite the common use of the Schirmer test in ocular surface evaluation, there is a lack of standardization of commercially available Schirmer strips which can affect results, diagnosis, and treatment. While clinicians expect consistent moisture migration on a Schirmer strip with the same volume of tears, variations in strip characteristics between manufacturers suggest otherwise in practice.¹⁶ Previously, García-Porta described differences in morphology, quality, and fluid uptake of phosphate-buffered saline among different brands of Schirmer strip and proposed that these disparities could affect wetting length.¹⁶ Similarly, Lewin reported variations in wetting length among different veterinary Schirmer strip brands after immersing the tip in water for 1 minute and the application of the Schirmer test in dogs for the same duration.¹⁷ Although efforts have been made to explore capillary dynamics of strips for quantifying basal tear production rates, practical applications have been limited.^18,19

Our study presents two tear volume–based methods for clinicians to calibrate Schirmer strip and obtain brand independent, more accurate Schirmer test results. We believe that our calibrated migration method is easy to establish in clinics, has the potential to standardize tear volume measurements in Schirmer strips, and improve accuracy in clinical assessments.

MATERIALS AND METHODS

Schirmer Strips

Sterile Schirmer strips from single batches of 5 different manufacturers were used: TearFlo (HUB Pharmaceuticals LLC, Scottsdale, AZ; manufactured by Madhu Instruments, New Delhi, India), Haag-Streit (Clement Clarke, Harlow, UK; for consistency, only flat bottom strips were used for all tear migration experiments), EagleVision Colorbar (Corza Medical, Parsippany, NJ), Contacare Ophthalmics and Diagnostics (Gujarat, India), and MIPL/A6 Schirmer test strips with blue mark (Madhu Instruments, New Delhi, India). Strip geometry was measured with digital calipers (Model: EC799A-8/200, Starrett, Ethol, MA) and their weight determined with a scientific scale (Mettler, Toledo XSE105 Dual Range).

Tear Migration Experiments

Human tears (M.T.M., P.S.) were collected with a glass capillary (Drummond Scientific, Broomall, PA), then transferred to an Eppendorf tube, and frozen until used. All tear measurements were repeated 3 times. Volumes for moisture migration experiments were 2 μL, 3 μL, 4 μL, 5 μL, 6 μL, 7 μL, 8 μL, 9 μL, 10 μL, 11 μL, 12 μL, 13 μL, 14 μL, 15 μL, 17 μL, 20 μL, and then in increments of 2 μL until the strip was completely saturated.

Each dry Schirmer strip was weighed on the plexiglass surface of a cell culture dish lid (Cell star 628 160; Greiner Bio-One GmbH, Frickenhausen, Germany) inside the calibrated scale, and its dry weight was determined. Next, a pipette was used to carefully drop a specified volume of human tears on the surface at the top margin of the Schirmer strip, and complete absorption of the volume was observed. The strip was then immediately weighed and again after 1 and 5 minutes. The most advanced point of moisture migration of the strip was read in mm on the migration scale printed on the Schirmer strip and documented after exactly 5 minutes. For the Haag-Streit strips, which do not have a printed scale, the moisture migration was measured with the digital calipers from the apex of the notch downward, as specified by the manufacturer. All experiments were performed in a medical facility at a room temperature of 72°F.

Statistical Analysis

All data analysis was performed using MATLAB (Mathworks, Natick, MA), calculating mean and SD of the geometric parameters. Distribution analysis was performed with the Kruskal–Wallis test and the 2-sample Kolmogorov–Smirnov test. Linear regressions (linear least-squares fits with offsets) of data points from each brand of strips, or jointly from all brands, respectively, were performed, followed by calculation of the root-mean-square deviation (rmsd) of data points to the best-fit regression. For tear volume measurements by weight, the dry weight of each strip was subtracted from its weight 5 minutes after tear application. A density of 1.000 mg/mL was assumed for tears. Best-fit straight lines through data points of measured volume after 5 minutes versus applied volume exhibited slopes smaller than 1.0, indicating evaporation. A consensus rate of evaporation was calculated by globally analyzing data after 5 minutes from all brands jointly.

RESULTS

Schirmer Strips Exhibit Different Geometric Parameters Between Brands

We examined Schirmer strips from 5 different brands (Fig. 1A) and observed that they are not standardized but differ in design of shape, width, weight, overall length as well as distance from the top of the strip to the first marking, millimeter scales, and in manufacturing precision. Dry Schirmer strip thickness was not significantly different between brands (Table 1). Strip weight was significantly different among strips except between the MIPL and Tearflo brands, which are manufactured by the same company. We also observed that Schirmer strips with the greatest weight and width (Haag-Streit) exhibited the lowest migration in mms for each volume.

FIGURE 1.

A, Photograph of 5 different brands of Schirmer strips used to compare moisture migration. Parallel alignment of dry strips shows differences in shape, length, width, and millimeter scales between manufacturers. (1) Haag-Streit, (2) Contacare Ophthalmics and diagnostics, (3) TearFlo, (4) MIPL/A6, and (5) EagleVision/Corza Medical. B, Photograph of the different brands 5 minutes after application of 5 μL of human tears at the tip of the strips. The red lines show the wide range of moisture migration lines observed for the same volume of tears applied. All moisture migration values were read from imprinted scale except for Haag-Streit which was measured with digital calipers: (1) Haag-Streit (unmarked strip) = 5 mm, (2) Contacare = 9 mm, (3) MIPL = 9 mm, (4) TearFlo = 8 mm, and (5) EagleVision = 10.5 mm.

TABLE 1.

Physical Characteristics (Mean and SD) of Five Different Brands of Commercially Available Schirmer Strips

	Haag-Streit	Contacare	MIPL	TearFlo	EagleVision	P *
Length (mm) N = 5
Total	45.88 ± 0.13	40.08 ± 0.03	40.92 ± 0.1	40.95 ± 0.04	39.77 ± 0.13	0.0002
Top to first marking or to notch apex in unmarked strip	6.63 ± 0.1	4.90 ± 0.1	5.10 ± 0.2	4.79 ± 0.2	5.99 ± 0.1	0.0005
Width (mm) N = 5
Strip head	6.07 ± 0.05	5.04 ± 0.06	5.0 ± 0.03	5.06 ± 0.08	4.97 ± 0.06	0.004
Strip body	5.94 ± 0.01	5.07 ± 0.07	5.0 ± 0.03	5.06 ± 0.07	4.97 ± 0.06	0.004
Dry weight (mg) N = 60	22.29 ± 0.54	16.44 ± 0.77	17.66 ± 0.85	17.47 ± 0.74	16.68 ± 0.47	<0.00001
Thickness (mm) N = 5
Strip head	0.25 ± 0.02	0.23 ± 0.02	0.22 ± 0.01	0.21 ± 0.03	0.22 ± 0.01	0.1
Strip body	0.22 ± 0.01	0.23 ± 0.02	0.22 ± 0.01	0.20 ± 0.03	0.23 ± 0.01	0.3
Ruler marking	No	Yes	Yes	Yes	Yes	—
Ink	No	No	Yes	No	Yes	—
Sterile	Yes	Yes	Yes	Yes	Yes	—

^*Kruskal–Wallis test.

Moisture Migration of Tears Differs Depending on Schirmer Strip Manufacturer

To evaluate the impact of these differences on tear migration, we applied human tears in volumes ranging from 2 μl up to 30 μL and measured the migration distance result of the wetted front after 5 minutes. In initial experiments, differences in the migration behavior of tears and buffered salt solution on Schirmer strips were observed. Saline solution had a higher moisture migration (average 6.6%; range: 0.08%–19.6%) compared with human tears. Therefore, all experiments were performed with human tears. Gradual application of tear volume in minute increments (e.g. for a 10 μL volume, applying 2.5 μL at 0, 1, 2, 3 minutes) versus the complete volume at 0 minutes did not significantly change total strip moisture migration; therefore, the application was performed at once as described in the Methods.

As shown in Figures 1B, 2 a wide range of migration values were observed for each applied tear volume. A statistical measurement of the overall best-fit linear regression of all data points from all brands (dashed black line in Fig. 2) led to an rsmd of approximately 2.9 mm (equivalent to volume variations of 2.0 μL). Substantial variations in tear migration among different brands can be discerned (different colors for each brand in Fig. 2). This is highlighted in the distinct slopes of the separate best-fit regression lines of data points from each brand (Fig. 2, colored solid lines). As an example, for the range of values observed for a given tear volume, the application of 8 μL of tears yielded migration values spanning from 6.6 mm to 14 mm among various strip brands.

FIGURE 2.

Measured moisture migration for different tear volumes. Shown are replicate data points (triangles, circles, and crosses) for each experiment performed in triplicate for the different brand Schirmer strips, indicated by different colors. For each brand, the best-fit linear regression is shown as colored solid line, with slopes and offsets for the different brands listed in Table 2, along with the rmsd of the data points to the straight line. The overall consensus regression of all data points is shown as dotted black line, has a slope of 1.43 mm/μL with an offset of −0.38 mm, and fits with an overall rmsd of 2.89 mm.

Importantly, we observed that within the same brand, the migration data exhibited a markedly improved linear regression, with rmsd values ranging between 0.89 and 1.47 mm (Table 2). This provided an opportunity to mitigate brand dependency, by using the best-fit linear relationships for each brand as calibration curves between the observed migration and the tear volume. By applying this conversion from migration distance to the corresponding tear volume, it became possible to compare the estimated tear volumes relative to the known applied volumes. This calibration correction eliminated the bias from different brands, and errors are reduced 2.2-fold (with rmsd of 0.89 μL; Fig. 3). A conversion table of best-fit mm migration distances to calibrated tear volumes for each brand is provided in Table 3.

TABLE 2.

Best-Fit Parameters of the Linear Regression as Shown in Figure 2

	Haag-Streit	Contacare	MIPL	TearFlo	EagleVision
Slope a (mm/μL)	1.16	1.38	1.44	1.43	1.51
Offset b (mm)	−1.11	0.14	−0.12	−0.38	0.88
rmsd (mm)	1.47	1.12	1.23	1.09	0.89

Given a measured moisture migration m, the corresponding tear volume, v, can be calculated as, v = (m–b)/a.

FIGURE 3.

Tear volumes determined based on moisture migration and calibration for different brands (symbols) based on the parameters from Table 2. The black solid line depicts the expected identity of pipetted volumes and measured volumes. The rmsd of all data points to this line is 0.89 μL.

TABLE 3.

Conversion Table for Millimeters of Moisture Migration of Tears to Microliter Volume in Different Brands of Schirmer Strips

Moisture Migration (mm)	Corresponding Tear Volume (μL)
	Haag-Streit	ContaCare	MIPL	TearFlo	EagleVision
1	1.8	0.6	0.8	1	0.1
2	2.7	1.3	1.5	1.7	0.7
3	3.5	2.1	2.2	2.4	1.4
4	4.4	2.8	2.9	3.1	2.1
5	5.3	3.5	3.6	3.8	2.7
6	6.1	4.2	4.3	4.5	3.4
7	7	5	5	5.2	4.1
8	7.8	5.7	5.7	5.9	4.7
9	8.7	6.4	6.4	6.6	5.4
10	9.6	7.1	7.1	7.3	6
11	10.4	7.8	7.7	8	6.7
12	11.3	8.6	8.4	8.7	7.4
13	12.2	9.3	9.1	9.4	8
14	13	10	9.8	10.1	8.7
15	13.9	10.7	10.5	10.8	9.3
16	14.7	11.5	11.2	11.5	10
17	15.6	12.2	11.9	12.2	10.7
18	16.5	12.9	12.6	12.9	11.3
19	17.3	13.6	13.3	13.6	12
20	18.2	14.4	14	14.3	12.7
21	19	15.1	14.7	15	13.3
22	19.9	15.8	15.4	15.7	14
23	20.8	16.5	16.1	16.4	14.6
24	21.6	17.2	16.8	17.1	15.3
25	22.5	18	17.5	17.8	16
26	23.3	18.7	18.2	18.5	16.6
27	24.2	19.4	18.9	19.2	17.3
28	25.1	20.1	19.6	19.9	18
29	25.9	20.9	20.3	20.6	18.6
30	26.8	21.6	21	21.3	19.3
31	27.6	22.3	21.7	22	19.9
32	28.5	23	22.4	22.7	20.6
33	29.4	23.8	23.1	23.4	21.3
34	30.2	24.5	23.8	24.1	21.9
35	31.1	25.2	24.5	24.8	22.6

Tear volumes are calculated based on the linear regressions for data points of each brand using the parameters in Table 2.

In addition, we conducted a study to explore a more direct method of tear volume measurement. This involved weighing each strip before and after tear application and using the weight difference to calculate the tear volume in each strip. This approach bypasses variation in moisture migration and migration reading errors and removes the largest errors from variability among individual strips. Nonetheless, evaporation poses a challenge, given that the Schirmer test is timed for 5 minutes. However, our observations indicate that evaporation rates remain consistent across all measured strips, independent of the brand (see Figures, Supplemental Digital Content 1, http://links.lww.com/ICO/B598, http://links.lww.com/ICO/B599, http://links.lww.com/ICO/B600, http://links.lww.com/ICO/B601, http://links.lww.com/ICO/B602). By factoring in the calculated consensus evaporation loss of 15.2%, the weight-based measurement of tear volume further improves the consistency between the known and measured tear volumes resulting in an overall rmsd of 0.64 μL and a 3.1-fold error reduction compared with the uncalibrated migration results (Fig. 4).

FIGURE 4.

Tear volumes determined based on net weight differences. Data shown are for different brands (symbols), accounting for consensus evaporation of 15.2% after 5 minutes. The black solid line depicts the expected identity of pipetted volumes and measured volumes. The rmsd of all data points to this line is 0.64 μL.

We propose a transition from reporting migration distances to reporting tear volume to overcome the design differences of the Schirmer strip brands as shown below.

Calibration Methods for Schirmer Strips

METHOD A1: Calibrated migration method applicable for the 5 brands listed in Table 3.

1. Conduct the Schirmer test and measure moisture migration at 5 minutes in mm.

2. Consulting calibration in Table 3, specifically the column corresponding to the strip brand used, look up the corresponding tear volume. (For example, with a moisture migration of 10 mm using EagleVision strips, the volume is equivalent to = 6 μL.)

Alternatively, use the formula v = (m – b)/a to convert the observed moisture migration mm, m, to the tear volume, v, based on the slopes a and offset b provided in Table 2 for the different brands. (For example, with a moisture migration of 10 mm using EagleVision strips, the volume is (10–0.88)/1.51 = 6.04 μL.)

METHOD A2: Calibrated migration method for any brand of Schirmer strips.

1. Collect calibration data by pipetting a range of tear volumes from 2 to 30 μL (in increments of 1 μL up to 15 μL and then in 2 μL increments) followed by measurement of the moisture migration at 5 minutes. Measure migration for all volumes in triplicate.

2. Perform a linear regression analysis of all data points i of the millimeter migration, m_i, vs. pipetted volumes, v_i, in the form, m_i = a×v_i + b, to determine the best-fit slopes and offset values a and b, respectively.

3. For Schirmer tests, measure the moisture migration as usual. Then, having determined calibration parameters a and b: given a measured mm migration m*, use the formula v* = (m*–b)/a to find the corresponding tear volume v*.

METHOD B: Calibrated weight method for any brand of strips.

This protocol requires a scientific scale with at least 0.1 mg accuracy, both for establishing the calibration and for Schirmer strip measurements.

1. Weigh a Schirmer strip and record dry weight in mg. Apply the first volume of 2 μL of tears, and weigh the strip again after 5 minutes. Calculate the net weight difference in mg, which corresponds to 2 μL of applied tear volume (assuming a density of 1 g/mL) allowing for evaporation in the measurement at the 5-minute time point.

2. Repeat step 1 for a range of tear volumes from 2 μl to 30 μL (in increments of 1 μL up to 15 μL and then in 2 μL increments), all measurements in triplicate.

3. Perform a linear regression analysis of all data points i of the measured volumes after 5 minutes, w_i, versus applied volumes, v_i, in the form, w_i = a_w × v_i + b_w, to determine the slopes and offset values a_w and b_w, respectively.

4. After the calibration experiment, perform tear volume measurements by weighing a strip before and after 5 minutes of regular Schirmer tests: given a measured net weight difference w*, use the formula, v* = (w* – b_w)/a_w, to find the corresponding tear volume v*.

DISCUSSION

In this study, we evaluated the correlation between volume and moisture migration of human tears in 5 different brands of Schirmer strips and discovered significant differences of migration for the same volume between manufacturers. The observed variability among different brands of Schirmer strips could lead to unintentional Schirmer test result errors and subsequently inadequate classification of the ocular criteria for Sjögren syndrome and oGVHD or other causes of ocular surface disease contingent on the strip used for testing.^6,8–10 To address the potential errors mentioned above, we propose calibration of Schirmer strips and converting Schirmer test results from mm to μL using the described calibration methods (METHODS A1, A2, B). For example, using METHOD A1 and Table 3, a 5-mm moisture migration result on a Schirmer test would be equivalent to 5.3 μL of tears on a Haag-Streit Schirmer strip and 2.7 μL on an EagleVision Schirmer strip. The documented μL volume results would be comparable, universal, and independent of strip brands.

Currently, when a clinician applies the Schirmer test, a consistent result among brands is expected. However, we have demonstrated that this is not the case. For example, we observed that the same tear volume of 5 μL resulted in migration values ranging from 4 to 10.5 mm depending on the Schirmer brand used (Figs. 1B, 2). These results imply that a patient with ocular surface symptoms and 5 μL of tears will only have a Schirmer I test result of ≤5 mm if a Haag-Streit strip is used for testing, but none of the others although the same amount of tears would be absorbed into each strip. In clinical practice, a measurement of 5 mm on the Schirmer test indicates the presence of dry eye disease and tear gland dysfunction, while a measurement of 10.5 mm is considered within the normal range.^20,21 Our observed wide range of variability is patient-independent and different from other known contributors of Schirmer test inconsistencies such as eye and lid position or intrapatient variability.^22–24 This brand-dependent variability of moisture migration among Schirmer strip brands and potential for clinical implications was previously observed.¹⁶ Importantly, if uncorrected, Schirmer test measurements will depend on the used brand rather than the clinical presentation itself and could still affect the overall ocular surface disease score and influence therapy decisions or clinical trial outcomes.^4,6,8,9,12 Therefore, mitigation of patient-independent factors should be considered such as our proposed methods of volume calibration of Schirmer strips, which reduced brand-related measurement errors 2.2-fold to 3.1-fold. We demonstrated that weight measurement of Schirmer strips before and after applying tears allows for the most accurate tear volume calculations (3.1-fold error reduction) by side-stepping altogether the error of migration measurements and differences in strip design (Fig. 4; METHOD B). However, in the clinical setting, weight measurements of Schirmer strips may not be practical because scientific scales are not readily available in eye clinics and traditionally the mm migration documentation is part of the examination evaluation. In those cases, we suggest converting mm migration into μL using the proposed METHOD A1 and using conversion Table 3 or creating a brand-specific calibration table as outlined in section METHOD A2. Because volume to mm migration differs substantially between Schirmer strip brands, a volume-based cut-off value for ≤5 mm migration in microliters may be more accurate. From averaging the calibrated moisture migration values of all strips used in this study, we suggest that 5 mm of moisture migration corresponds to 3.8 μL of tears.

A limitation of this study is the neglect of more detailed nonclinical physical models of moisture migration, including the impact of ambient temperature and humidity on evaporation, although all experiments were performed at room temperature in a medical facility to mimic the clinical settings in which Schirmer testing is usually performed.¹⁹ In view of the large measurement errors we describe in the article, we believe improvements from more refined migration models to be clinically insignificant.²⁵

In summary, we describe large measurement errors between Schirmer strip brands for same tear volumes, which can introduce errors in the clinical grading and assessment of diagnostic criteria of ocular surface disease. We propose that calibrations of the Schirmer strip and converting Schirmer test results from millimeters into microliters will allow more accurate, brand-independent, and reliable results for analysis.