The efficacy of varenicline solution nasal spray in the treatment of DED was evaluated in three randomized, double-masked, placebo (vehicle)-controlled trials: the short-term (4-week/28-day) ONSET-1 [25] and -2 [26] studies, which were conducted at multiple centres in the USA (Sect. 3.1; Fig. 1); and the longer-term (12-week/84-day) MYSTIC study [27], which was conducted at a single centre in Mexico (Sect. 3.2). This section focuses on findings for patients who received varenicline solution at the approved dosage of 0.03 mg in each nostril twice daily (Sect. 5) or vehicle in these studies.

Trial design of the randomized, double-masked, vehicle-controlled, multicentre, phase IIb ONSET-1 and phase III ONSET-2 trials in adults with dry eye disease [25, 26], with further information available in Table 1. Efficacy results for the intranasal varenicline solution dosage approved in the USA are reported in the animated figure (available online). BID twice daily, BL baseline, DED dry eye disease, LSM least squares mean, NA not reported, OR odds ratio, pts patients, STS Schirmer test score

Supplementary file2 (MP4 6190 KB)

Eligible patients were aged ≥ 22 years, had received a diagnosis of DED and had used (or expressed the desire to use) an artificial tear substitute for symptoms within 6 months of the study start [25,26,27]. The ocular inclusion criteria included: baseline Schirmer test score (STS) performed with topical anaesthesia (a measure of basal tear film production) of ≤ 10 mm/5 min and STS ≥ 7 mm greater in the same eye upon intranasal mechanical stimulation with a cotton swab; corneal fluorescein staining score of ≥ 2 in at least one corneal region or a sum of ≥ 4 for all corneal regions; OSDI score of ≥ 23, with up to three responses of “not applicable” (ONSET-1 and -2 only); and best-corrected visual acuity < 0.7 logarithm of the minimum angle of resolution [25,26,27]. If both eyes qualified, then the eye with the greatest increase in STS upon mechanical stimulation or, if no difference, the eye with the lower basal STS, was selected as the study eye [25, 26]. If neither measure differed, the right (as opposed to left) eye was chosen as the study eye [25, 26]. Of note, eligibility in the ONSET studies was not limited by the baseline severity of DED symptoms, as assessed in terms of the eye dryness score (EDS). All three studies excluded patients who used contact lenses within 7 days of screening or anticipated using lenses during the study period [25,26,27].

Participants in the phase IIb ONSET-1 study (n = 182) were randomized to receive varenicline solution 0.006 mg (n = 47), 0.03 mg (n = 48) or 0.06 mg (n = 44), or vehicle (n = 43) [25], while those in the phase III ONSET-2 study (n = 758) were randomized to receive varenicline solution 0.03 mg (n = 260) or 0.06 mg (n = 246), or vehicle (n = 252) [26] (Fig. 1). These doses of varenicline solution (and vehicle) were delivered with each actuation of an intranasal spray device; patients administered one 0.05 mL spray in each nostril twice daily for 4 weeks [25, 26]. Randomization was not stratified by baseline factors in ONSET-1 [25], but was stratified according to preprocedure anaesthetized STS (≤ 5 mm vs > 5 mm), preprocedure EDS (< 60 vs ≥ 60) and study site in ONSET-2 [26]. Use of artificial tears was permitted during the studies [20, 26].

The primary efficacy endpoint in ONSET-1 was the change in the anaesthetized STS in the study eye from baseline to day 28 [25], whereas that in ONSET-2 was the proportion of patients achieving a ≥ 10 mm improvement in anaesthetized STS in the study eye from baseline to week 4 [26]. Secondary endpoints included the change from baseline in EDS, as assessed in a controlled adverse environment (CAE) chamber (at day 21 in ONSET-1 and day 28 in ONSET-2) and in the clinic (at day 28 in both ONSET-1 and -2) [25, 26].

Baseline demographics and ocular assessments were generally similar between the varenicline solution 0.03 mg and vehicle groups in both trials [25, 26]. Across the four treatment groups, the mean age of patients ranged from ≈ 58–67 years; the majority were female (71–80%) and White (81–93%). Baseline anaesthetized STS and EDS score ranged from 4.5–5.1 and 58.1–65.2, respectively (Table 1) [25, 26].

Four weeks of treatment with varenicline solution 0.03 mg resulted in statistically significant and clinically meaningful improvements in signs and symptoms of DED (Table 1) [25, 26]. In terms of basal tear film production, the improvements in anesthetized STS from baseline to day 28 and the proportions of patients achieving a ≥ 10 mm improvement in anaesthetized STS at week 4 were significantly (p < 0.001) greater with varenicline solution than with vehicle in both studies (Table 1) [25, 26].

A post hoc exploratory analysis conducted on the primary endpoint of ONSET-1 using a last observation carried forward (LOCF) approach to account for missing data yielded comparable results to the primary analysis [25]. The least squares mean (LSM) improvements in anesthetized STS from baseline were 11.7 (95% CI 9.24–14.26) mm in the varenicline solution group (n = 48) versus 3.2 (95% CI 0.62–5.80) mm in the vehicle group (n = 43) [p < 0.0001] [25]. Similarly, post hoc analyses performed on the primary endpoint of ONSET-2 showed that significantly (p ≤ 0.03) greater proportions of varenicline solution than vehicle recipients achieved a ≥ 10 mm improvement in anaesthetized STS, irrespective of the baseline severity of basal tear production [odds ratios (ORs) of 3.46 (95% CI 1.99–6.04) and 1.83 (95% CI 1.03–3.24) for anaesthetized STS ≤ 5 and > 5 mm] or DED symptoms [ORs of 3.37 (95% CI 1.81–6.29) and 2.05 (95% CI 1.22–3.46) for EDS < 60 and ≥ 60]) [26].

In ONSET-1, varenicline solution significantly improved EDS from baseline relative to vehicle, as assessed in the CAE chamber (Table 1) and in the clinic [LSM treatment difference (varenicline solution vs vehicle) of − 13.3 mm; p = 0.021] [25]. In comparison, varenicline solution did not significantly improve EDS from baseline in the CAE chamber compared with vehicle in ONSET-2 (Table 1), potentially because of limitations imposed by social distancing requirements during the coronavirus disease 2019 pandemic, which restricted use of the CAE chamber and collection of symptom data in ≈ 30% of the study population [26]. In terms of EDS assessed in the clinic, the improvement from baseline to day 28 with varenicline solution was significantly greater than with vehicle by week 2 [LSM treatment difference (varenicline solution vs vehicle) of − 3.7 mm; nominal p = 0.049]; this trend continued at week 4 [LSM treatment difference of − 4.4 mm; nominal p = 0.038] [26].

Varenicline solution was effective in increasing basal tear production, regardless of baseline severity of DED symptoms, according to an LOCF analysis of data pooled from ONSET-1 and -2 [28]. The mean improvement in anesthetized STS from baseline to week 4 was significantly (p ≤ 0.01) greater among varenicline solution 0.03 mg than vehicle recipients in both the baseline EDS < 40 subgroup (8.4 vs 4.1 mm) and the baseline EDS ≥ 40 subgroup (9.9 vs 3.9 mm). Similarly, significantly (p ≤ 0.03) more varenicline solution 0.03 mg than vehicle recipients achieved a ≥ 10 mm improvement in anaesthetized STS at week 4 in both the baseline EDS < 40 subgroup (30.9 vs 18.5%; OR of 3.02) and the baseline EDS ≥ 40 subgroup (38.8 vs 17.9%; OR of 3.17) [28].

Varenicline solution was administered bilaterally (i.e. to each nostril) in these studies and the pooled analysis also showed that basal tear production was increased in fellow eyes as well as in study eyes [28]. For study eyes, the LSM improvement in anesthetized STS from baseline to week 4 was 10.5 mm in varenicline solution 0.03 mg recipients versus 5.0 mm in vehicle recipients; the corresponding results for the fellow eyes were 8.8 versus 2.8 mm (values estimated from a graph). Similarly, more varenicline solution 0.03 mg than vehicle recipients achieved a ≥ 10 mm improvement in anaesthetized STS at week 4, based on outcomes in study eyes [OR of 2.86 (95% CI 1.98–4.14)] as well as in fellow eyes [OR of 3.53 (95% CI 2.34–5.31)] (values estimated from a graph) [28].

Eligible patients enrolled in the phase II MYSTIC study (n = 123) were randomized to receive varenicline solution 0.03 mg (n = 41) or 0.06 mg (n = 41), or vehicle (n = 41) [27]. As in the ONSET studies, these doses of varenicline solution (and vehicle) were delivered with each actuation of an intranasal spray device; patients administered one spray (0.05 mL) in each nostril twice daily for 12 weeks. The primary efficacy endpoint was the change in the anesthetized STS in the study eye from baseline to day 84 [27].

Baseline characteristics were generally similar with respect to mean age (51 years in the varenicline solution group vs 56 years in the vehicle group), female sex (78% vs 80%) and anaesthetized STS (5.5 vs 5.3; Table 1); all patients were Hispanics/Latinos [27].

Varenicline solution 0.03 mg significantly improved tear film production in patients with DED over a 12-week period (Table 1) [27]. Notably, tear production increased as early as 5 min after administration of varenicline solution and was maintained throughout the 84-day study, with no evidence of drop-off in efficacy or development of tolerance with long-term use [27]. Just over one-third of varenicline solution recipients compared with one-quarter of vehicle recipients experienced a ≥ 10 mm improvement in anaesthetized STS at week 12, although the between-group difference was not statistically significant (Table 1) [27].