Subject recruitment

A prospective and diagnostic study recruited a total of 108 eyes from 63 patients with ocular disorders who visited the Outpatient Department of Beijing Tongren Hospital, Beijing, China between February and April 2022. The subjects were divided into two groups according to DED diagnostic criteria [8]. The experimental group included 63 eyes with DED while the control group included 45 eyes without DED.

The diagnosis of DED was made according to the consensus on DED in China (2020) and required at least one of the following symptoms: dryness, sandiness, burning, tiredness, discomfort, redness, and blurred vision with an Ocular Surface Disease Index (OSDI) questionnaire score ≥ 13 scores [8]. In the case of the above symptoms, the following signs were necessary for diagnosis: (1) fluorescein tear film breakup time (FBUT) ≤ 5 s or a non-anesthesia Schirmer Ι test value ≤ 5 mm/5 min; or (2) 5 s < FBUT ≤ 10 s or 5 mm/5 min < non-anesthesia Schirmer Ι test ≤ 10 mm/5 min, accompanied by a corneal fluorescein staining positive(≥ 5 dots) [8].

All included subjects signed informed consents forms in accordance with the tenets of the Declaration of Helsinki. The study was approved by the Institutional Review Board of Beijing Tongren Hospital, Beijing, China. The subjects over 18 years-of-age with the same ethnicity (Chinese) and were willing to participate were included in the study. Patients who had previously undergone corneal or ocular surgery, had worn contact lenses in the previous 24 h, had any history of Stevens Johnsons Syndrome or Sjögren syndrome, or other systemic illness or risk factors known to impact the tear film, or had undergone any treatment related to dry eyes within the previous two weeks, or those who had taken medication recently were excluded.

Ocular examinations

Each subject was required to carry out specific ocular examinations in the following order:

1. Ocular Surface Disease Index (OSDI) questionnaire

2. Non-invasive tear breakup time (NIBUT)

3. Tear meniscus height (TMH) assessment

4. Phenol red thread (PRT) test (in triplicate)

5. Fluorescein tear breakup time (FBUT)

6. Corneal fluorescein staining

7. Schirmer I test

Subjects rested for at least 10 min before the NIBUT, TMH assessment, FBUT and Corneal fluorescein staining. A 15-min rest was arranged before the PRT test and the Schirmer I test [9, 10]. This practice ensured that the tear film, tear volume and ocular condition recovered to its original status. All examiners and analyzers were blinded from the subjects clinical and demographic details, and all assessments were performed in a dimly lit room (temperature 20–25 °C, humidity 30–40%) between 8 am and 4 pm on a single day.

Dry eye symptoms questionnaire

Each subject was asked to complete the OSDI questionnaire (validated Chinese version [11]). This questionnaire featured three segments, including ocular symptoms, vision-related function, and environmental triggers, to evaluate subjective dry eye symptoms in a detailed manner [12]; the outcome ranges from 0 to 100 with higher scores referring to severer symptoms.

Non-invasive tear breakup time (NIBUT)

The NIBUT was tested with an Oculus Keratograph (K5M; Oculus Optikgeräte GmbH, Wetzlar, Germany). Subjects rested their chin on the chin rest with their foreheads pressed against the forehead band, watched the fixation pattern inside the device and blinked according to instruction. The tear breakup time was examined automatically and the first and average tear breakup time was reported [13].

Tear meniscus height (TMH) assessment

The Oculus Keratograph was also used for TMH assessment. Subjects placed their chin and forehead as discussed before and watched a fixed pattern without blinking. An image of the tear meniscus was captured by the tester, and the height of the tear meniscus was measured by the integrated ruler at the central point perpendicular to the lid margin relative to the pupil center. The TMH was measured by Keratograph 5 M for consecutive 3 times, and the mean value was recorded [13].

Phenol red thread (PRT) test

A yellow cotton thread soaked with phenol red (Tianjin Jingming New Technological Development Co., Ltd) was used for this test. The thread, 70 mm in length with a 3 mm fold at the beginning, was fixed in the strip with a scale (10 mm in width and 70 mm in length) and covered with plastic film. The examiner removed the front of the plastic film without touching the thread, inserted the folded portion into the middle and outer 1/3 of the inferior palpebral conjunctiva and started timing. The test was conducted without topical anesthesia. When the tears are absorbed, the thread turns from white to orange, and then to red. After 20 s, the thread was pulled out and the length of the thread stained red was measured. The length of the color-changing thread refers to the volume of tears [7].

The test was conducted three times for each subject. Two consecutive tests were performed by the same examiner with a 15-min interval to measure intra-observer repeatability. Then, another examiner carried out an individual test with another 15-min interval to test the inter-examiner reproducibility.

Fluorescein tear breakup time (FBUT)

Before conducting the test, an aseptic fluorescein strip moisturized with normal saline was dipped into the subject’s conjunctival sac in both eyes. The subject was then required to blink several times to ensure that the tear film and cornea were stained evenly. Then, the examiner observed the tear film through a cobalt-blue filter with a slit-lamp microscope with the subject’s eyes opened. The interval between the last complete blink and the appearance of the first corneal black spot in the stained tear film was recorded three times by a stopwatch and the mean value was calculated as the FBUT. Records that were less than or equal to 5 s were regarded as abnormal.

Corneal fluorescein staining

As the cornea has been stained during the FBUT test, the examiner could observe the cornea through a cobalt-blue filter using a slit-lamp microscope directly. The number of stained spots was counted and the appearance of stained spots was considered abnormal [14].

Schirmer I test

Standardized Schirmer I test strips were utilized to assess aqueous tear production. The strips were placed without anesthesia over the middle and outer 1/3 of the inferior palpebral conjunctiva in both eyes simultaneously and then left for 5 min with the eyes closed. The length of the strip that had become wet was read as the outcome and a reading less than or equal to 5 mm was considered abnormal [15].

Statistical analysis

As for a study design with 2 repeated measures, the uncertainty was set to be 20% in the repeatability and reproducibility result, which means the sample size for precision studies must be over 48 according to the formula [16]:

$$1.96\frac^-1\right)}}=0.2Sw$$

Sw: within-subject standard deviation; n: number of the subject; n’: number of repeated measurements.

As for the study designed to evaluate the diagnostic validation of the PRT test, the sensitivity was set to be 90% and the allowable error was set to be 10% according to the previous study [17, 18], which means the sample size for precision studies must be over 84 according to the formula:

$$n=\frac \sqrt+Z_\sqrt}}$$

π: sensitivity; σ: permissible fluctuation range of sensitivity; β: value of Class II error is 0.8, and Z (1-β) is 0.84; α: value of Class I error is 0.05, and Z (1-α) is 1.65

SPSS version 26.0 (SPSS, Inc., Chicago, IL, USA) was used to conduct statistical analysis. The Shapiro–Wilk test and histogram were used to test the raw data for normality. The independent-sample t test and Kruskal–Wallis H test was used to assess the differences between experimental group and control group. The intra-class correlation coefficient (ICC) and Bland–Altman analysis were carried out to assess the intra-observer repeatability and the inter-examiner reproducibility. The Spearman rank-order correlation coefficient was calculated to evaluate correlations among the PRT test and other ocular examinations. Receiver operating characteristic (ROC) curve analysis was conducted to assess the diagnostic accuracy of the PRT test. The standard deviation (SD) and coefficient of variance (CV) were computed to assess the fluctuation of certain ocular examination parameters. All P values were two-sided and were considered statistically significant at P < 0.05.