KeratoScoop's guest columnist: Frank J. Holly, Ph.D.

The Schirmer test (also called Schirmer I test) was invented by the German ophthalmologist Herr Doktor O. Schirmer early in the twentieth century. This simple, inexpensive and rapid test is designed to determine in a semi-quantitative fashion the ability of a person to lacrimate subsequent to a minor ocular irritation. This test which gained popularity later was the first of three tests developed by Schirmer. The Schirmer I test relied on the irritation caused solely by placing a 5 mm rounded end of a filter paper strip into the lower fornix .

The three Schirmer tests differed only in the source of irritation. The second Schirmer test involved the irritation of the nasal cavity with a camel hair brush. The third Schirmer test required the patient to stare at the sun. Since the latter caused retinal damage it soon went out of vogue. However, in the last few years, the camel brush technique appears to have come back in use sporadically.

Since the Schirmer I test is the only one that has gained widespread popularity over the years, in the United States it is commonly referred to simply as the Schirmer test.

The test consists of placing one end of a 35 mm long and 5 mm wide strip of unbonded porous paper (such as that used for filtration in analytical chemistry) in the lower fornix usually at one-third of the palpebral distance from the temporal canthus (near the temple). Even though the end of the strip is rounded to avoid excessive irritation, the strip is inserted dry, and the presence of such foreign body surrounded by the conjunctival mucosa causes minor trigeminal nerve irritation. In conscious human beings, the procedure also invokes a certain degree of apprehension, which undoubtedly enhances the elicited reflex tearing and is present even in anesthetized eyes

Despite the numerous articles published on this test since the appearance of the classical paper of Schirmer in 1903, prior to the eighties there is no evidence of attention ever having been paid to the basic factors affecting the wetting rate of the paper strip. Whether the wetting rate of the strip reflected the secretion rate, or depended on the magnitude of the wetted length (i.e. time) of the paper strip were apparently not even considered. During most of the twentieth century, only the paper type constituting the strip was the subject of concern. See Table I for the various types of papers used by different investigators.

Many of the more thorough investigators tested the various papers in vitro by immersing the tip of the strip in a beaker-full of water. Thus, the experiments were conducted under unlimited supply condition. Under such conditions, the wetting rate of the paper strip is the highest and depends solely on the characteristics of the paper and the properties of the fluid absorbed.

It is not surprising then that under such conditions the wetting rate of the paper strip widely varied with paper type and even depended on the direction the paper strip was cut from the sheet of paper. The test paper was finally standardized in 1961 by using Black Ribbon No. 589 which was actually quite similar in absorptivity to Whatman No. 41 filter paper.

It was recognized by Halberg and Berens that the wetting rate of the paper strip must depend on the amount of liquid that saturates a unit length of the paper. The higher this specific wetting volume, the slower the wetting rate of the paper for a given rate of supply of fluid (see later).

For almost eighty years, none of the investigators asked much less answered the basic questions related to the tear flow in the Schirmer strip. Without analyzing the nature and limitation of fluid flow in the strip, one is unable to answer even the most basic question: How, if at all, the wetted length of the Schirmer strip relates to the tear secretion rate and/or the tear volume in the eye?

Nature of Flow: It can be reasonably assumed hat the mechanism of fluid transport in the strip occurs by capillary absorption by a porous solid, the unbonded paper strip. The author and co-workers demonstrated that the wetting curve of the strip (wetted length plotted versus time when unlimited supply of fluid is available) can be quantitatively described by representing the paper strip with a straight (horizontal) capillary tube having a diameter of approximately one micrometer. In science, the rate of flow in such a capillary is known to be proportional to the radius of the capillary, the horizontal component of the surface tension at the meniscus edge, inversely proportional to the viscosity of the fluid and to the distance of the advancing meniscus from the fluid source (see figure 1 below).

Figure 1. Schematic diagram of the kinetic model for porous paper strip wetting in the absence of evaporation

Fluid absorption rate by the strip — unlimited supply: It can be calculated as well as experimentally determined (Holly et al) that the tear uptake rate by the strip starts at very high values and then rapidly decreases (inversely with the square root of time). This happens when the edge of the strip is immersed in a large volume of fluid (unlimited supply). This is actually the maximum rate at which the paper strip can absorb tears or any other fluid.

If the tear secretion rate is higher than this given (but time-dependent ) value (secretion rate > maximum absorption rate), then the test is useless as the result then only reflect the characteristics of the porous paper for a given fluid and has nothing to do with the supply rate of the fluid to the strip, which in the eye is the tear secretion rate.

When this maximum uptake rate by the strip is exceeded by the tear secretion rate, the Schirmer test readings become meaningless because they only reflect the properties of the paper and not the magnitude of the tear secretion rate

Fluid absorption rate by the strip — limited tear supply: If the secretion rate is smaller than the maximum absorption rate, then the wetting rate of the strip may reflect the secretion rate (this has to be so otherwise the whole test is meaningless) provided that the following statements are true.

This liquid portion could be named the specific tear volume of a strip and it actually has been determined by several investigators under unlimited supply conditions. Halberg and Behrens found that their paper strip: (No. 589 Black Ribbon filter paper) absorbed 0.5 microliters of fluid per mm length for their 5 mm wide paper strip. Prause et al found 0.65 microliter/mm during their experiments for the filter paper strip they used. Holly and co-workers determined the value of specific wetting volume for the porous paper they used (Whatman No. 41, also Black Ribbon 589) as 0.63 microliter per mm. They also found that this average specific wetting volume was reasonably constant under various uptake rates and during several time intervals.

Evaporation from the surface of the paper strip: Evaporative loss can be considerable. The specific evaporation rate from one mm segment of wetted filter paper was found to be 0.043 microliter/(mm.min) at 30% humidity and 26 degree of Celsius. Turbulent ambient air conditions could increase this value considerably. Such an evaporative loss which increases with increasing wetted length will eventually reach the absorption rate of the tear by the strip. At this point a steady state is reached where the wetted length of the strip remains the same no matter how long the measurement is continued. Needless to say, such a state of affairs, yields an artifactually low tear secretion rate.

It follows that the tear evaporation during the test has to be prevented in order to make the measurements meaningful. This can be readily achieved by enclosing the paper strip in a transparent polyethylene sheath.

Fortunately in most patients, the tear secretion rate is considerable lower than the maximum uptake rate by the paper strip, so that the Schirmer test is conducted under the condition of limited fluid supply. Under this conditions, the strip wetting rate is linearly proportional to the tear secretion rate. This proportionality constant is the specific strip wetting volume, which was found to be 0.63 microliter/mm for Whatman No. 41 filter paper strips

As an aside, this is not true for the so-called thread tests. The absorption capacity of the threads are so low that a normal lacrimation rate overwhelms the absorption capacity of the thread in a few seconds after which the wetting rate will be determined by the threadÕs properties and not by the tear secretion rate.

Holly et al demonstrated by using a Harvard pump that constant rate flow delivered to the end of the paper trip results in a linear increase of the wetting length with time as long as the rate does not exceed the maximum supply rate even at the end of the experiment (up to ten minutes). The slope of the wetting length increase with time (the wetting rate) was found to be equal to the ratio of the fluid supply rate divided by the specific wetting volume. This was only true however, when the evaporation was prevented by a plastic sheath surrounding the strip. Thus the relation between the fluid uptake (or tear secretion rate) is then governed by the simple equation:

Where F is the rate of flow in microliters/min, V is the specific volume of wetting in microliters/mm, l is the length of wetting and t is the time. The rate of wetting is written in the differential form to allow for possible variation in the rate of secretion, i.e. the uptake.

The reproducibility of the measurements could also be enhanced by having the paper strip undergo lipid extraction by organic solvents.

Once the basic principles were recognized and the relation between the supply rate and the wetting rate was found, actual wetting curves obtained with human volunteers involving specially modified Schirmer strips could be analyzed and the time dependence of the tear secretion rate during the lacrimation response induced by the insertion of the Schirmer strip could be obtained. Several hundred wetting curves were analyzed and interesting, generally valid things were found. (see Figure 2)

In every case examined, the tear secretion rate varied during the measurement but in a predictable, mathematically describable manner. The tear secretion rate started at a high value, e.g. ten microliters per minute (initial tear secretion rate), then it exponentially decreased to a lower constant value, e.g. one microliter per minute. In a mathematical form:

Where F_i is the initial tear secretion rate, F_f is the final tear secretion rate, k is the tear secretion decay coefficient, and t is the time (see Figure 3).

Graphically the initial tear secretion rate could be obtained from the initial slope of the wetting curve, the final slope of the wetting curve yielded the final tear secretion rate, while the tear decay coefficient could be calculated from the intercept of the initial and final slopes (see Figure 2 below).

Figure 2. Length of wetting of paper strip as a function of time by using the average inetic parameters obtained in unanesthetized patients

It is not practical to determine and plot the complete wetting curves for 10 or even five minutes in clinical practice. However, it is plausible to make three or four readings at various times using modified Schirmer strips (lipid-extracted and plastic-coated) and then by the use of nomograms or computerized algorithms obtain the three kinetic parameters.

Effect of Topical Anesthesia: The frequency distributions of the three kinetic parameters were obtained in a heterogeneous population with and without the use of topical anesthesia. In all cases, highly skewed distribution was obtained. The only difference caused by topical anesthesia was that the extremely high values for these three parameters were absent. In other words, the frequency distribution curves were truncated in the anesthetized eyes.

This resulted in the following observation; the average value of the parameters were much lower in the anesthetized eyes. (see Figure 4 below). The difference between the mean values were considerably less, and the node values (most probable values) were not statistically different in the un-anesthetized and anesthetized eyes. .

Figure 3. Time dependence of tear secretion rate calculated by using average kinetic parameters obtained in unanesthetized eyes.

Influence of the exponentially decreasing secretion rates: Since the highest tear secretion rate occurs immediately after the insertion of the strip and since evaporation is greater toward the end of the measurement in an unmodified Schirmer test, most of the wetting will occur during the first minute of the measurement. When expressed as a relative fraction of the total wetting (LR), this value will only depend on two parameters; the ratio of the initial and final tear secretion rate and on the magnitude of the secretion decay coefficient. In most cases 40-50% of the total wetting was obtained during the first minute. In some extreme cases, 85% of the total wetting was achieved during the first minute.

False low readings in Sjšgren patients: We may digress here and mention the interesting finding of Prause et al. These Danish investigators found among patients suffering from Sjšgren's syndrome who are having severely limited lacrimating ability and correspondingly low tear volume, that the rounded end of the Schirmer strip in the fornix does not get fully saturated until at least 5 mm of wetting (outside the eye) is achieved. That is to say, until a considerable time is elapsed. In normal patients, the Schirmer strip end saturates with tears almost instantly due to the large, accessible tear volume in the tear meniscus. Due to this effect, the Schirmer test reading (in 5 minutes) yields artifactually low readings in Sjšgren's patients

Figure 4. Effect of topical anesthesia on the wetting curves of porous paper strip.

An important and well controlled study was done by the author and his co-workers using the modified Schirmer test (at that time referred to as the Schirmer-Holly test) to investigate the difference in lacrimation response between dry eye patients and normals. The measurements were made in a group of dry eye patients and compared to measurements obtained with age- and sex-matched normals. All the patients in the dry eye group possessed a history of chronic symptomology characteristic of dry eyes including burning, itching, dryness, mucus discharge chronic foreign body sensation, objective findings of less than 5 mm wetting in 5 minutes obtained by the classical Schirmer test, and the typical corneal and conjunctival staining pattern with vital stain Rose Bengal (a staining severity of 3+ or greater in each eye on three separate occasions). Every attempt was made to recruit subjects similar in age and sex distribution so that the presumable differences could be more accurately analyzed.

The determination of the wetting curve for ten minutes in the absence of evaporation using lipid-extracted porous strips (Black Ribbon #589) under topical anesthesia was conducted to eliminate extreme values and minimize the scatter of the data.

The results obtained were interesting if somewhat unexpected. There was no significant difference between the initial tear secretion rate between dry eye patients and normals! Conversely, quite a large difference was found between the final tear secretion rates. The dry eye patients exhibited about one half the normal final tear secretion rates. The secretion decay coefficient, which is a measure of how rapidly the tear secretion rate decreases to the final value was significantly higher in dry eye patients than in normals.

Since, as we have seen, the conventional Schirmer test mostly reflects the magnitude of the initial tear secretion rate, and, due to evaporation, is also the most inaccurate toward the end of the measurement where the final tear secretion rate dominates, the conventional Schirmer test is indeed a poor diagnostic tool to detect the lacrimation deficiency in sicca patients.

Even though the Schirmer test has been in clinical use for a century and has been studied by researchers during that time, it has taken eighty years before the researchers were able to analyze and quantitatively describe the kinetics of the capillary tear flow in porous non-bonded paper strips.

Researchers in the eighties recognized that the capillary flow in the strip is the controlling factor determining the maximum rate of wetting of the paper strip, if and only if, unlimited fluid supply is available at one end of the strip. (secretion rate > maximum absorption rate(secretion rate > maximum absorption rate This maximum wetting rate decreases with time. Thus the test is only meaningful if the actual tear secretion rate in the measured eye does not exceed this maximum wetting rate during the whole time interval of the measurement. The filter paper traditionally employed in manufacturing Schirmer strips has the absorption capacity to meet this requirement. Lacrimation tests utilizing cotton threads do not meet this requirement and can be used only to approximate the magnitude of the tear volume and not the secretion rate in the eye.

Lipid-extraction of the paper strips increases absorption capacity as well as the maximum wetting rate thus it is helpful to ensure the fulfillment of this requirement. Such a pretreatment also prevents false positive results when the inserted strip end is lipid contaminated and fails to absorb tears readily in the cul-de-sac provided that the paper strip end is not touched by the person administering the test.

To make the test results more meaningful, evaporation rate of the tears from the wetted portion of the Schirmer strip has to be prevented. This can be readily achieved by a plastic sheath made of thin, polyethylene film.

The improved Schirmer test is capable of determining the lacrimation rate and its time dependence. Due to the now known and well-defined type of the tear secretion rate decay (during lacrimation) it should be sufficient to make only two or three readings during the five minutes to obtain meaningful results of the initial and final tear secretion rates in patients. The final tear secretion rate is the most useful in diagnosing dry eye patients since this parameter showed the largest decrease in dry eye patients.

The lacrimation response in patients having various dry eye conditions should also be studied once such an improved Schirmer test becomes widely available. For instance, lacrimation rates in iatrogenic dry eyes have never been determined.

It may be concluded that the classical Schirmer test, as it is conducted, does not reflect the tear secretion kinetics during lacrimation response. The results are also biased against showing the relatively small difference between dry eye patients and normals since in the majority of dry eye patients the lacrimation ability is not seriously compromised.

An improved Schirmer test would prove to be a useful diagnostic tool. It would not increase the cost much and would only require small additional effort on the part of the test provider. It would yield diagnostically valuable data on the tear secretion rate.

The industry has not shown much interest to help develop and market such a ÒprestigeÓ item, possibly because of the lack of interest shown by the profession and also, because the profit margin for such an item might be quite low.

Schirmer, O. (1903): Studien zur Physiologie und Pathologie der Trþnen absonderung and Trþnen abfŸhr. Grafes Arch Ophthalmol. 56, 197-291.

Spector, S.A. (1931) Chronic keratoconjunctivitis, chronic pharyngitis, and chronic arthritis due to ovarian insufficiency, Klin. Med. 9, 876-883

Sjšgren, H. (1933) Zur Kenntnis der Keratoconjunctivitis sicca. Acta Ophthalmol. Suppl. 2, p.1.

De Roetth, A. Sr. (1941) On the hypofunction of the lacrimal gland. Am. J. Ophthalmol. 24, 20-25.

Eisner, G. (1961) (1961) Der Einfluss der Papierwahl auf di Resultate des SchirmerÕschen Testes. Ophthalmologica, 141, 314-319.

Halberg, G. and Berens, C. (1961) Standardized Schirmer tear test kit. Am. J. Ophthalmol. 51, 840-842.

Norn, M.S. (1965) Tear secretion in normal eyes. Acta Ophthalmologica (Copenhagen) 43, 567-573.

Van Bijsterveld, O.P. (1969) Diagnostic test in the sicca syndrome. Ach. Ophthamol (Chicago) 82, 10-14.

Scherz, W., Doane, M.G., and Dohlman, C.H. (1974) Tear volume in normal eyes and keratoconjunctivitis sicca. Arbrecht v. Graefes Arch. Klin. Exp. Ophthalmol. 192, 141-150.

Lamberts, D.W., Foster, S., and Perry, H. (1979) Schirmer test after topical anesthesia and the tear meniscus height in normal eyes. Arch. Ophthalmol. 97, 1082-1085.

Hype, T.J. (1980) Fluid uptake in Schirmer papers and its relevance to their employment in tear collection in lysozyme tests. Brit. J. Ophthalmol. (1980) 88, 48-49.

Jordan, A. and Baum, J. (1980) Basic tear flow: Does it exists? Ophthalmology (Rochester) 87(9), 920-930.

Clinch, T.E., Benedetto, D.A., Laibson, P.R. and Felbert, N.T. (1982) The kinetics of Schirmer strip wetting (abstract) Suppl. Invest. Ophthalmol. Vis, Sci. 22, 185.

Holly, F.J., D.W. Lamberts, and E.D. Esquivel, (1982) Kinetics of capillary tear flow in the Schirmer strip. Current Eye Res. 2(1), 57-70.

Holly, F.J., S.J. Laukaitis, and E.D. Esquivel, (1984) Kinetics of lacrimal secretion in normal human subjects. Current Eye Res. 3(7), 897-910.

Holly, F.J., W.E. Beebe, and E.D. Esquivel, Lacrimation Kinetics in Humans as Determined by a Novel Technique. Chapter in Precorneal Tear Film in Health, Disease and Contact Lens Wear

(edited by F.J. Holly, D.W. Lamberts, and D.L. MacKeen) Dry Eye Institute, Lubbock, TX 1986.

Beebe, W.E., E.D. Esquivel and F.J. Holly (1988) Comparison of lacrimation in dry eye patients and normals. Current Eye Res. 7, 419- .

Holly, F.J.: Lacrimation kinetics as determined by a Schirmer-type technique. Chapter in Lacrimal Gland, Tear Film, and Dry Eye Syndromes (ed. By D.A. Sullivan,) Plenum Press, New York, 1994.

Did you find this editorial useful? Write to the editor and let us know your opinion.

Back to Keratoscoop

Column archives