The lack of correlation between symptoms and signs in patients with meibomian gland dysfunction: a secondary analysis of the multicenter, randomized controlled trial

Meibomian gland dysfunction (MGD) is a chronic ocular surface disease with symptoms that severely affect the quality of life of patients. The number of patients with meibomian gland dysfunction (MGD) has been increasing under the influence of many factors such as population aging, lifestyle changes, and widespread use of video terminals. MGD is globally widespread, with prevalence ranging from 3.5% to 69.3% and with a particularly high prevalence in Asian populations [1]. The main function of the meibomian glands is to secrete meibum, which constitutes the lipid layer of the tear film to prevent tear evaporation [2]. When the terminal duct of the meibomian gland is obstructed or the quality and quantity of meibum change, the stability of the tear film is destroyed, resulting in eye irritation or inflammatory reactions and, in severe cases, causing ocular surface damage [3].

Symptoms of eye discomfort in MGD patients include dryness, grittiness, soreness, burning sensation, blurred vision, ocular fatigue, and increased secretions. MGD is the main cause of evaporative dry eye, and while symptoms of dry eye and MGD often overlap, studies have reported that approximately 50% of MGD patients were asymptomatic [4]. In addition, some patients do not have pathological manifestations such as inflammation, or any obvious clinical signs, in which case, MGD can only be diagnosed by evaluating meibomian gland secretion [5]. This phenomenon suggests that MGD is nonobvious in some patients. However, few studies have reported whether symptoms and signs are related in MGD patients who have symptoms and signs. In this study, we will use the baseline data of patients before intense pulsed light (IPL) treatment to explore the relationship between symptoms and signs in MGD patients and provide new insights for better treatment and follow-up.

This multicenter study was conducted at four sites in China and was approved by the ethics committee. Subjects were enrolled from November 2017 to April 2018.

A total of 122 patients, with an average age of 42.2 years, were enrolled in this study, including 27 males with an average age of 38.6 years and 95 females with an average age of 43.2 years. The median SPEED score was 14.00 (11.75–18.00), median TBUT was 2.77(2.20–3.99) seconds, mean CFSS was 1.25 (0.00–2.00), and median MGYSS was 4.50(2.50–7.5). The results of the subjects after the intervention were published [9]. Here we performed statistical analysis on the baseline data.

As shown in Table 1, using Spearman correlation analyses, there was no correlation between SPEED scores and TBUT (r = 0.017, p = 0.852), CFSS (r = -0.171, p = 0.060), or MGYSS (r = -0.144, p = 0.112) in MGD patients. In addition, SPEED scores in MGD patients were not correlated with lid margin hyperemia (r = 0.064, p = 0.485), lid margin thickening (r = 0.091, p = 0.318), lid margin rounding (r = 0.048, p = 0.598), hyperkeratinization of the lid margin (r = -0.051, p = 0.579) or telangiectasia around meibomian gland orifices (r = 0.009, p = 0.920).

SPEED standard patient evaluation of eye dryness, TBUT tear break-up time, CFSS corneal fluorescein staining score, MGYSS meibomian gland yielding secretion score.

In order to find the influencing factors that cause the patient's symptoms and signs to be unrelated, we performed a subgroup analysis of the data. The median age of the 122 patients was 39.5 years. Taking 40 years old as the threshold, the patients were divided into 2 groups for comparison. As shown in Table 2, there was no significant difference in SPEED, TBUT, CFSS and MGYSS between the two groups (p > 0.05). However, more patients in the ≥ 40-year-old group had lid margin thickening, rounding, hyperkeratinization and telangiectasia than patients in the < 40-year-old group (p < 0.05). In male and female patients, there was no significant difference in the severity of symptoms and signs between the two group (p > 0.05). Of the 95 women enrolled, 71 were premenopausal and 24 were postmenopausal. Comparison of the symptoms and signs between premenopausal women and postmenopausal women was conducted, it revealed that more postmenopausal women had lid margin hyperemia, thickening, rounding, hyperkeratinization, and telangiectasia around meibomian gland orifices than premenopausal women (p < 0.05). According to geographical area, the 122 patients were divided into 2 groups. As shown in Table 2, although there was no difference in SPEED, TBUT, CFSS and MGYSS between the two groups (p > 0.05), more patients living in the north had lid margin hyperemia, thickening, rounding, hyperkeratinization, and telangiectasia around meibomian gland orifices than patients living in the south (p < 0.05).

In the subgroup analyses, we found that age, menopause, and region may affect the patient's lid margin appearance, but not the SPEED score. Therefore, we used multiple linear regression analysis in all subjects to control for possible effects of age, sex, and region. The results found that after controlling for these three factors, there was no correlation between symptoms and signs (p > 0.05). In addition, multiple linear regression analysis was also used in female subjects to adjust for possible effects of age, menopause, and region. Statistical results showed that symptoms and signs were still not correlated after controlling for these three factors (p > 0.05) (Table 3).

In order to further explore the reasons why symptoms and signs were not related, multiple linear regression analyses and binary logistic regression analyses were used to further analyze the effect of these influencing factors on symptoms and signs in all subjects. As shown in Table 4, age may be an influencing factor of lid margin thickening (R = 0.068, p < 0.0001), lid margin rounding (R = 0.099, p < 0.0001), hyperkeratinization of the lid margin (R = 0.068, p < 0.0001) and telangiectasia around meibomian gland orifices (R = 0.038, p = 0.013). Sex may affect TBUT (R = -0.614, p = 0.049) and lid margin rounding (R = -1.851, p = 0.013). In addition, region may be an influencing factor of lid margin hyperemia (R = -2.671, p < 0.0001), lid margin thickening (R = -2.118, p < 0.0001), lid margin rounding (R = -2.848, p < 0.0001) and hyperkeratinization of the lid margin (R = -2.378, p = 0.003). The results in female subjects are shown in Table 5. Age may affect lid margin thickening (R = 0.070, p = 0.023), lid margin rounding (R = 0.135, p = 0.001), hyperkeratinization of the lid margin (R = 0.089, p = 0.010). Menopause had significant effect on SPEED score (R = -4.112, p = 0.025), but not on signs. Furthermore, region may affect lid margin hyperemia (R = -3.125, p < 0.0001), lid margin thickening (R = -2.086, p = 0.003), hyperkeratinization of the lid margin (R = -2.800, p = 0.010).

Clinically, there is a group of patients whose meibomian glands are obstructed and secrete viscous or even toothpaste-like secretions under pressure, but they have no symptoms or have mild symptoms; there is a group of patients with obvious symptoms of MGD and often seek medical attention, but no abnormalities are detected during the clinical examination, and these patients have to blink frequently and forcefully to alleviate the symptoms. These patients were classified as having nonobvious obstructive MGD (NOMGD) [5]. It is unclear whether this symptom-sign dissociation occurs in MGD patients who have symptoms and signs.

Our results showed that the correlation between symptoms and signs in obvious obstructed MGD patients was poor, which suggests that MGD patients may have severe symptoms, but this does not mean that their signs are also obvious, and vice versa. Postmenopausal patients are older than non-menopausal women, and older patients are more likely to have abnormal appearance of MGD. However, in our results, menopausal patients were less symptomatic, and there is a separation of symptoms and signs. We speculate that factors such as differences in patients' own sensitivity to pain and relative corneal numbness with aging or disease progression may affect patients' symptoms. In addition, although the meibomian glands may have abnormal changes such as obstruction, the residual but active meibomian glands can maintain high levels of meibomian expression [10], which may make the patient's symptoms less obvious. From this point of view, the diagnosis of MGD and the assessment of the severity of the disease should be combined with a clinical examination of various aspects and needs a comprehensive evaluation.

At present, there are few reports on correlation analyses of symptoms and signs in MGD patients. Nishant et al. [11]analyzed the correlation between symptoms and signs of 53 MGD patients, and found that although the symptom scores (Ocular Surface Disease Index (OSDI)) and the signs scores were statistically different, but not clinically significant. MGD is the main cause of evaporative dry eye, and it has been reported that the correlations between symptoms and signs in dry eye patients is also poor. Nichols et al. [12] scored the severity of MGD according to MG gland orifice obstruction and secretions, and found that the severity of MGD in patients with dry eye was not correlated with symptom frequency. They also found that there was a poor relationship between symptoms and other clinical tests, including tear meniscus height, TBUT, CFSS, phenol red thread test, Schirmer test, and rose bengal staining. Yan et al. [13] analyzed correlations between the severity of various irritation symptoms and Schirmer test scores, TBUT, CFSS, and tear film images in 126 patients with dry eye, but they found no correlation between them. In fact, dry eye has many influencing factors, and the repeatability of examination findings is poor, which may lead to poor correlation between symptoms and signs.

Dry eye is mainly characterized by an imbalance in the homeostasis of the tear film, which is determined by the lacrimal functional unit (LFU) [14]. Failure of any part of the unit may cause an imbalance in tear film homeostasis. Lin et al. [15] detected the secretion of aqueous tears in 459 symptomatic dry eye patients by the Schirmer test, and found that 38.8% of MGD patients had a Schirmer test ≤ 5 mm, suggesting that MGD patients may suffer from insufficient aqueous secretion, which indicates that some of these patients may have mixed dry eye. Since the Schirmer test or tear meniscus height examination was not performed in our study, the possibility of mixed dry eye cannot be ruled out, which may be one of the reasons for the poor correlation between symptoms and signs. In addition, Song et al. [16] conducted the partial correlation analysis between the OSDI score and the sign score of 176 dry eye patients, and found that the symptoms of the patients were positively correlated with the abnormal appearance of the eyelid margin. Inclusion criteria of this study did not include the severity of MGD, but our inclusion criteria included a meibomian gland function score of ≤ 12, which means our recruited patients had more severe meibomian gland dysfunction. This may be the reason for the different results of our study.

Since morphological changes of the eyelid margin are objective indicators of MGD, we examined the eyelid margin [17]. An abnormal eyelid margin is the external manifestation of MGD, and its internal causes are gland orifice obstruction, duct dilation, and abnormal secretion of meibum [18, 19]. In this study, we found that MGD patients in the older group were more likely to have abnormal lid margins. Age has been identified as an important factor affecting MGD [1]. In aging human and animal tissues, researchers found that the proliferation of acinar basal cells decreased, the glands were atrophied and infiltrated by inflammatory cells, and the expression of PPAR-γ was reduced or nuclear translocated [20, 21]. The abnormal changes in the eyelid margin may be related to aging. We also found that postmenopausal women were more prone to eyelid margin abnormalities. Because postmenopausal women are relatively older, age may be a factor that causes eyelid margin abnormalities. In addition, meibomian gland function has a certain relationship with sex hormones, especially after menopause. The rapid decline in sex hormone levels can induce or exacerbate MGD [22, 23]. Finally, we also noticed that abnormalities in eyelid margins were more common in MGD patients living in northern China. It is worth noting that our research was conducted in winter, the average temperature in Beijing in winter was -3 °C-0°C, and the relative humidity was 42%-43%. However, Hangzhou had an average temperature of 5 °C to 7 °C with a relative humidity of 68%-77%, and Wenzhou had an average temperature of 9–11 °C with a relative humidity of 67%-74%. The climate of Beijing was cold and dry, but the climate in Zhejiang was relatively warm and humid. Some studies have shown that a dry environment is a risk factor for MGD. Suhalim et al. [24] established a mouse model in a desiccating environment, which resulted in abnormal meibomian gland function, including an increase in the proliferation of acinar basal cells, altered lipid production, and abnormal cell differentiation. Environmental dryness or humidity may be an important reason for the difference in eyelid margins between patients living in the north and south. Although morphological changes of the eyelid margin can reflect the condition of the meibomian gland [25], patients with MGD may not have eyelid margin abnormalities, and a comprehensive assessment is needed to evaluate the presence and severity of MGD.

Our study has some limitations. Firstly, although patients undergoing eye surgery or eyelid surgery within 6 months were included in the exclusion criteria of our study. However, studies have shown that cataract surgery may cause dry eye and the condition may persist for more than a year [26, 27]. Of our subjects, 3 had cataract surgery before 6 months, which may have affected the outcome. Additionally, our inclusion criteria in our study included the MGYSS of ≤ 12, which may have caused some selectivity bias. This may have limited our search for general symptom-sign associations. Furthermore, our small sample size may be a limitation. These limitations need to be addressed in future studies.

Taken together, we found that the correlation between symptoms and signs of obvious MGD in patients is poor and that age, menopause, and a dry environment may aggravate the disease. We believe that the severity of the disease needs to be comprehensively evaluated. More evidence is needed to clarify the relationship between symptoms and signs for an accurate diagnosis of the disease.