**4. Discussion**

The meniscometry readings in this population of garmen<sup>t</sup> factory workers in Bangladesh ranged between 1–16 mm, and significantly lower SM readings were found in males, and in participants with higher OSDI, lower TBUT, lower Schirmer's readings, and higher MGD grades.

In the first study performed in a clinic in Japan (*n* = 90), significantly lower SM readings were observed among patients with DED. In addition, the Schirmer's readings, TBUT, mean tear film lipid layer interferometry grade, and vital staining scores were also observed to be lower among patients with DED. With regards to the SM readings, a positive linear correlation exists with the Schirmer's readings as well as the TBUT [16]. In the second study performed in a clinic in Japan (*n* = 175), significantly lower SM readings were observed among patients with DED, with a positive correlation with graticule scale tear meniscus height (TMH) [15].

There were no previous studies of SM in a community setting or study based on the Bangladeshi participants.

In our study (Table 2), lower Schirmer and higher OSDI were associated with abnormal meniscometry values after adjusting for the other clinical variables. This is not surprising since Schirmer's I test, apart from a measure of reflex tear secretion, may also be related to the tear volume in the lower meniscus. With a stricter threshold of SM (Table 3), only lower TBUT and higher OSDI were associated with abnormal SM. This may be related to the pre-existing lower meniscal volume to related to tear stability.

This study involved a large sample in the community within a single occupational group. The use of a uniform protocol by a single ophthalmologist provided an accurate and comprehensive analysis and with a higher confidence of estimates. Since the study did not employ a meibomian gland evaluator, the number of meibomian glands yielding liquid meibum was not documented. The number of subjects with SM less than 3 mm was limited. Another possible limitation is that this study involved only one occupational group. We are not certain why the distribution of SM is bimodal.
