Artificial Tears: A Systematic Review

Abstract

Artificial tears are the mainstay of dry eye disease management, but also have a role in corneal abrasion and wound healing, pain and inflammation management, conjunctivitis, keratitis, contact lens rewetting and removal, and foreign body removal. A systematic review of randomized controlled trials (PROSPERO registration CRD42022369619) comparing the efficacy of artificial tears in patients with dry eye to inform prescribing choices using Web of Science, PubMed and Medline databases identified 64 relevant articles. There is good evidence that artificial tears improve symptoms of dry eye disease within a month of regular use, applied about four times a day, but signs generally take several months to improve. Not all patients with dry eye disease benefit from artificial tears, so if there is no benefit over a month, alternative management should be considered. Combination formulations are more effective than single active ingredient artificial tears. Artificial tears containing polyethylene glycol are more effective than those containing carboxymethylcellulose/carmellose sodium and hydroxypropyl methylcellulose. Those classified as having evaporative dry eye disease, benefit from artificial tears with liposomes, especially of higher concentration. The data available is limited by the definition of dry eye disease applied in published studies being variable, as well as the disease severity examined and compliance with artificial tears being rarely quantified.

Keywords:

• artificial tears
• dry eye
• comfort
• contact lenses

Artificial tear drops are most commonly associated with the management of dry eye disease (DED). Artificial tears are typically included in first-line management options for dry eye, as they are easy to use, accessible in a wide range of formulations, and have a low risk-profile.¹ Most artificial tear preparations have been found to be effective in reducing the symptoms and signs of DED, however the Tear Film and Ocular Surface Society (TFOS) dry eye workshop in 2017 (DEWS II) concluded there had been relatively few high quality randomized controlled trials comparing different formulations with each other.^1,2 Furthermore, few clinical trials have compared the efficacy of different artificial tear products, and attempted to correlate this with patient characteristics, in order to aid management decisions for an individual.^3,4 The issue with this is that both practitioners and patients are faced with a bewildering array of different products with varying ingredients, and little or no clear way of knowing which is the most effective. Practitioners will often be asked “which is the best drop for dry eye”, but with no scientific evidence to base their answer on. In addition, other aspects that influence practitioner and patient choices are:

• formulation

  • percentage concentration⁵

  • molecular weight⁵

  • preservative used⁶

• storage bottle design.^7–10

Patients may therefore face a trial-and-error approach to product selection, incurring mounting costs and frustration in the process. This will be felt even more keenly by patients who are highly price sensitive, since over-the-counter products are no longer easily available via National Health Service (NHS) subsidised prescriptions¹¹ in the UK. A recent study¹² on the reported experience of dry eye management across four continents identified that on average, DED still caused a moderate impact on an individual’s quality of life (median impact 3/10); less than half of the individuals in any country had undergone a consultation with an eye or health-care practitioner about their dry eye; about half had tried dry eye treatment, with artificial tears being the most common treatment, followed by warm compresses, and both therapies were rated as reasonably effective (median 5–7/10).

Formulation

The majority of artificial tear products are aqueous-based and contain viscosity-enhancing agents, such as carbomer 940, carboxymethyl cellulose (CMC), dextran, hyaluronic acid, sodium hyaluronate (which has a smaller molecular size), hydroxypropyl guar (HP-guar), hydroxypropyl methylcellulose (HPMC hypromellose), polyvinyl alcohol, polyvinylpyrrolidone and polyethylene glycol, which aid lubrication and increase on-eye retention time.¹ Other ingredients may include osmotic agents, osmoprotectants, antioxidants, preservatives and inactives such as pH buffers, excipients and electrolytes.¹ Aqueous-based artificial tears target principally the muco-aqueous phase of the tear film, but have been shown to improve dry eye symptoms related to all subtypes of DED.²

In recent years, there has been an increase in the popularity and availability of lipid-based drops, which target the superficial tear lipid layer^13,14 as the emphasis on meibomian gland dysfunction and its role in evaporative dry eye continues to increase.¹ It has been demonstrated in randomised controlled trials that lipid-based drops are more effective at managing DED classified as evaporative.^3,4 These can take the form of nano-emulsion drops or liposomal sprays, which are applied to the closed eye and may be easier for those who struggle to instil drops, for example those with reduced manual dexterity or hand tremor. A completely water-free drop comprised of 100% lipid (perfluorohexyloctane) is available now, with the added benefit of being preservative-free.¹⁵

Preservatives

Multidose eye drops, including artificial and medicated topical ocular drops, commonly contain preservatives to maintain sterility and prolong shelf life, however, these are also known to produce toxicity. Benzalkonium chloride, commonly found in multidose drops, can produce toxic, proinflammatory and detergent effects, which may actually lead to or exacerbate DED.¹⁶ For this reason, there has been a move towards preservative-free and unit dose formulations, due to the risk of toxic and allergic reactions, especially when frequent instillation is required. Newer preparations may contain less damaging preservatives such as polyquaternium, or “vanishing” preservatives such as sodium perborate and purite, or feature specially designed bottles, which prevent the entry of microorganisms.¹⁷ Preservative-free formulations are recommended for all types of dry eye, however this is even more important for severe dry eye or sensitive individuals, and more details can be found in the TFOS DEWS II iatrogenic report.⁶

Ideal Properties

It is important that artificial tear drops behave in a similar way to natural tears. One aspect of this is the physical property of rheology, which refers to the way fluids and soft solids flow. The viscosity of human tears is high between blinks, but reduces during each blink cycle in order to protect the ocular surface from damage due to fluid turbulence.¹ Hence, they do not display Newtonian behaviours and are referred to as having non-Newtonian properties. Hyaluronic acid has been the subject of a significant amount of research and has been shown to exhibit these non-Newtonian shear-thinning properties,¹⁸ making it more like the tear film and hence suitable for use in artificial tears.¹⁹ Hyaluronic acid, a common constituent of artificial tears, is a naturally occurring glycosaminoglycan, which is found in and around body cells and tissues, for example in synovial fluid, and vitreous and aqueous humour.²⁰ Its use in ophthalmology was pioneered by Andre Balazs in the late 1960s,²¹ with Polack and McNiece²² being the first to report its use in dry eye. Hyaluronic acid is water soluble and is capable of binding large quantities of water, compared to its own weight, but its physical properties vary depending upon its molecular weight.²³ There is evidence to suggest that high molecular weight hyaluronic acid (HMWHA) is clinically superior in the treatment of DED compared to its low molecular weight counterpart.²⁴ Furthermore, HMWHA has been found to be protective against corneal cell apoptosis due to benzalkonium chloride toxicity, ultraviolet light radiation and chemical burns,^25–27 as well as being anti-inflammatory and having a role in reducing pain sensation.^24,28

Artificial Tears for Dry Eye Disease

There have been several systematic reviews^2,29–31 conducted over the past decade, concluding that artificial tears are a safe and effective way of treating DED. A meta-analysis concluded that the effectiveness of sodium hyaluronate did not differ based on its preparation³⁰ and another³² suggested that CMC appeared to be better than hyaluronic acid in treating DED, but the results were not statistically significant. Two recent reviews^5,33 both identified that while hyaluronic acid was effective in reducing the symptoms of DED, the ideal drop frequency and formulation (both concentration and molecular weight) for different ages and severities were yet to be investigated.

To date, there has been no review of studies which compared different artificial tears to identify whether certain formulations are more effective. Hence, with the objective to better understand the evidence for the effect of different artificial tears in managing dry eye, a search was made of the Web of Sciences databases (Clarivate Analytics, Philadelphia, USA) which includes the Science Citation Index Expanded covering over 9200 of the world’s most impactful journals from 1900 to the present day along with PubMed (including MEDLINE) from its inception. The systematic review was prospectively registered on PROSPERO (CRD42022369619) and was conducted in the format prescribed by PRISMA (2020).³⁴ A search for “artificial tear*” AND “randomi?ed” identified 481 unique results which were screened independently by two researchers (DB and DS) and verified by a third (JSW). Studies were eligible to be accepted if they were in full paper form (not abstracts or book chapters), compared two or more artificial tears against each other (not just with a placebo) and involved randomisation to avoid bias. This resulted in 64 papers being accepted (Figure 1) and the full text scrutinized for the key factors, which were tabulated in a spreadsheet and are summarised in Table 1. The study design, artificial tears compared, number and age profile of participants completing the trial, duration of use and dosing, tests conducted which showed a significant difference/did not differentiate between the products or change from baseline and general comments (dyes used for ocular surface staining, adverse events when reported and subanalyses) were extracted. Missing information is highlighted in the table and risk of bias analysis performed with the Cochrane Tool reported.³⁵ No data synthesis was attempted due to heterogeneity particularly in drop duration.

Table 1 Randomised Clinical Trials That Have Used Artificial Tears for the Treatment of Dry Eye Disease

Paper	Design	Comparators	Participants Completing	Age (Years)	Duration (Dosing)	Tests Showing Significant Difference	Tests Not Differentiating Products	Tests Showing Significant Difference	Tests Showing no Change	General Comments
						Cross Comparator		Compared to Baseline
Amran et al 2014^Citation69	Randomized, open-label, multi-center study	Cationic Emulsion - Cationorm PVA-Povidone - Refresh	N = 44 N = 35	61.3 ± 15.4 61.9 ± 12.5	4 weeks (4x/day)	Symptoms, TBUT, eyelid erythema, Conjunctival staining with Cationic	Schirmer’s Corneal staining			Sub-analysis with MGD participants CLINICAL TRIAL NOT REGISTERED
Aragona et al 2020^Citation48	Randomized, double-masked, multi-center study	CMC + HA - Optive Fusion UD CMC - Refresh Optive Sensitive/Optive UD	N = 180 N = 184	59.4 ± 13.8 57.5 ± 13.7	90 days (2x/day)	Lower ocular pain/discomfort CMC-HA	OSDI TBUT Ocular surface staining Schirmer’s II	OSDI Symptoms (VAS) TBUT Ocular surface staining	Schirmer’s II	10% minor AE CLINICAL TRIAL NOT REGISTERED
Baeyens et al 2012^Citation74	Randomized, double-masked, multi-center study	SH 0.18% - Vismed Carbomer 0.3% NaCL	N = 100 N = 96 N = 96	59.3 ± 15.0 (across groups)	84 days	Symptoms & Corneal staining with SH vs saline	Symptoms, corneal and conjunctival staining, Schirmer’s, TBUT SH vs carbomer	Symptoms Corneal staining	Conjunctival staining, Schirmer’s, TBUT	CLINICAL TRIAL NOT REGISTERED
Barabino et al 2014^Citation104	Randomized, double-masked, multi-center study	CMC 0.5% / glycerin 0.9% - Optive HA 0.2% / tamarind seed polysaccharide 0.2% - Xiloial	N = 25 N = 23	57.1 ± 17.4 52.2 ± 14.9	3 months (4x/day)	Symptoms with HA+TS	TBUT Ocular surface staining Schirmer’s	OSDI TBUT Ocular surface staining	Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Baudouin et al 2012^Citation57	Randomized, investigator-masked, multi-center study	CMC 0.5% and Osmoprotectant – Optive SH 0.18% - Vismed Multi	N = 37 N = 29	58.1 ± 14.2 55.4 ± 13.4	3 months	None	Osmolarity, Schirmer’s-I, OSDI, staining	Osmolarity, Schirmer’s-I, OSDI, staining	None	Clinical Trial NCT00987727 - only symptom primary and secondary outcomes and day 35 data missing Cochrane Risk of Bias R?C?M-O?I-S-B?
Benelli et al 2010^Citation62	Randomized, investigator-masked, single-center study	CMC 0.5% - Cellufresh PEG 400 2.5% - Blink Intensive HP‐guar 0.18%/PEG 400/PG - Systane	N = 20 N = 20 N = 20	Not stated	30 days (up to 4x/day)	Osmolarity with PEG400	VA Aberrometry Staining TBUT Schirmer’s	Aberrometry	Osmolarity VA Staining TBUT Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Brignole et al 2005^Citation40	Randomized, masked-observer, single-center study	CMC 1% - Celluvisc SH 0.18% - Vismed	N = 11 N = 10	69 ± 2 57 ± 2	2 months (3x/day)	CD44, comfort (only at day 7), keratitis recovery with SH	All other inflammatory markers, cornea and conjunctival staining, TBUT, corneal topography, tear meniscus height	Symptoms and ocular surface staining	None	Moderate dry eye and keratitis patients CLINICAL TRIAL NOT REGISTERED
Brodwall et al 1997^Citation105	Randomized, investigator-masked, single-center study	Polyacrylic acid 0.2% - Visco Tears PVA 1.4%	N = 38 N = 41	60.2 61.8	4 weeks (Drops/day a study variable; avg 3–5)	Symptoms (16/27 study days), hyperaemia, Rose Bengal staining, compliance with polyacrylic acid	TBUT Schirmer’s	Symptoms & signs (unspecified)	TBUT Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Bron et al 1998^Citation106	Randomized, double masked, multi-center study	Carbomer 940 0.2% - Lacrinorm/GelTears, Laboratoire Chauvin Carbomer 940 0.2% - Viscotears/Vidisic/Lacrigel	N = 92 N = 87	58.6 ± 16.2 64.0 ± 14.0	4 weeks (4x/day)	None	Symptoms TBUT, Schirmer’s, corneal and conjunctival staining	Symptoms TBUT, Schirmer’s, fluorescein/ lissamine green staining	None	AEs in n=21 Lacrinorm group and 17 Viscotears group CLINICAL TRIAL NOT REGISTERED
Calvao-Santos et al 2011^Citation41	Randomized, open-label, single-center study	Tears Again [lipidic] Opticol [aqueous] Optive [mucin] No treatment	N = 7 N = 6 N = 7 N = 7	24 to 53 years	30 days (not stated)	None	OSDI TBUT Schirmer’s	Symptoms, Schirmer’s for tears again	TBUT	Patients with digital eye strain. Compared drops primarily acting on one tear layer CLINICAL TRIAL NOT REGISTERED
Chiambaretta et al 2017^Citation50	Randomized, investigator-masked, multi-center study	HA-trehalose HA	N = 52 N = 49	60.0 ± 12.2 58.5 ± 13.4	84 days (3–6x/day; average 4)	Symptoms with HA-trehalose	Cornea & conjunctival staining	OSDI [Schirmer’s, TBUT, staining, hyperaemia, no statistics presented]	None	AEs: 3 with HA-trehalose vs 24 with HA Clinical Trial NCT02023268 - only staining as primary outcome and day 35 data missing Cochrane Risk of Bias R?C-M-O?I?S-B?
Christensen et al 2004^Citation61	Randomized, double-masked, multi-center study	PEG 400 0.4% / PG/HP‐guar 0.3% - Systane CMC 0.5% - Refresh Tears	N = 42 N = 45	58.5 59.5	6 weeks (4x/day)	Lissamine green staining, dryness, refreshed and FB symptoms with 0.5% PEG	Fluorescein staining, use ratings, ocular signs or symptom frequency	Corneal & conjunctival staining only with 0.4% PEG	Conjunctival staining with CMC	CLINICAL TRIAL NOT REGISTERED
Cohen et al 2014^Citation63	Randomized, double-masked, multi-center study	CMC 1% - Refresh LiquiGel PEG 400 0.4%/ PG/HP‐Guar 0.3% - Systane Gel	N = 70 N = 67	57.5±16.6 56.5±15.0	6 weeks (4x/day)	Corneal staining with PEG	Conjunctival staining, TBUT, symptoms	Corneal staining	Lissamine green staining, TBUT, symptoms	CLINICAL TRIAL NOT REGISTERED
Comez et al 2013^Citation107	Randomized, patient-masked, 2 group contralateral, single-center study	PG 0.3% and PEG 0.4% - Systane SH 15% - Eyestil HPMC - Tears Naturale CMC 0.5% - Refresh Tears	N = 17 N = 13	47.4±14.5 46.3±15.5	12 weeks (5x/day)	None	OSDI, Osmolarity, Schirmer’s, TBUT	OSDI, osmolarity, Schirmer’s, TBUT	None	~30% drop-out CLINICAL TRIAL NOT REGISTERED
Craig et al 2021^Citation4	Randomized, double-masked, multi-center study	Aminomethylpropanol, HP-guar - Systane Ultra Dimyristoyl phosphatidylglycerol, HP-guar, mineral oil, polyoxl 40 stearate - Systane Complete	N = 49 N = 50	43 ± 17 45 ± 16	6 months (4x/day +)	Lipid thickness	Symptoms, TMH, lipid, osmolarity, hyperaemia, expressibility, blinking	Symptoms (OSDI, DEQ-5, SANDE) NIBUT, LWE Cornea & conjunctival staining	TMH, osmolarity, hyperaemia, expressibility, blinking	Symptoms improved @1+ month, LWE @ 2+ months, lipid @ 3+ months staining @ 4+ months. 1 in 3 had no benefit in signs or symptoms. Those with lipid layer grade ≤ 3 benefit more from lipid-based drop Clinical Trial ACTRN12619000390189 - additional questionnaire, acuity and lid data presented Cochrane Risk of Bias R+C+M+O+I+S-B+
Dausch et al 2006^Citation76	Randomized, investigator-masked, cross-over, multi-center study	Liposomes - Tears Again Carbomer triglycerides - Liposic	N = 74 with deficient lipid layer	n=1 <25 years n=9 25–45 years n=16 46–60 years n=49 >60 years	6 weeks (3x/day)	Symptoms, LIPCOF, TBUT, Schirmer’s, lid margin inflammation with Tears Again		Symptoms, LIPCOF, TBUT, Schirmer’s, lid margin inflammation	-	Photo sequence of phospholipid liposomes sprayed on eyelid reaching ocular surface CLINICAL TRIAL NOT REGISTERED
Davitt et al 2010^Citation64	Randomized, double-masked, single-center study	PEG 400/PG/HP‐guar CMC 0.5% - Optive	N = 52 N = 53	33 x 18–64 years, 19 x ≥65 years 41x 18–64 years, 12 x ≥65 years	6 weeks (4x/day)	Cornea & conjunctival staining with PEG 400/PG/HP‐guar group	Symptoms, TBUT	Symptoms	TBUT	CLINICAL TRIAL NOT REGISTERED
Diaz-Llopis et al 2019^Citation108	Randomized, investigator-masked, multi-center study	Seawater spray - Quinton CMC 0.5% -Viscofresh	N = 60 N = 60	68.1 ± 6.3 66.8 ± 8.4	12 weeks (5x/day)	OSDI, IL-1 β and IL-6 with seawater spray	Cornea & conjunctival staining, Schirmer I, osmolarity, TBUT, TMH	OSDI, Cornea & conjunctival staining	Schirmer I, osmolarity, TBUT, TMH	CLINICAL TRIAL NOT REGISTERED
Downie et al 2020^Citation109	Randomized, double-masked, multi-center study	CMC, glycerin, flaxseed oil and castor oil and osmoprotectants (levocarnitine, Erythritol & trehalose) (OM3) Refresh Optive Advanced	N = 120 N = 122	54.3 ± 17.3 52.8 ± 16.7	90 days (2x/day +)	Combined corneal / conjunctival staining with OM3	OSDI TBUT	OSDI TBUT Combined corneal / conjunctival staining	None	AEs (OM3 0% vs ROA 4.1%) Clinical Trial NCT02553772 Cochrane Risk of Bias R+C+M+O+I+S+B?
Dumbleton et al 2009^Citation110	Randomized, double-masked, single-center study	PEG 400 0.25% - Blink gel tears CMC 1% - Refresh Liquigel	N = 56 N = 54	46.3 ± 19.3 47.2 ± 19.1	30 days (3x/day)	Symptoms with PEG	Phenol red test, TMH, NIBUT, hyperaemia, corneal and conjunctival staining		Hyperaemia, corneal and conjunctival staining	No notable AE’s CLINICAL TRIAL NOT REGISTERED
Essa et al 2018^Citation3	Randomized, investigator- masked, crossover, single-center study	SH 0.4% - Clinitas Soothe SH 0.15% - Hyabak Phospholipid liposomes -Tears Again CMC - TheraTears	N = 50 (for all treatments)	60.8 ± 14.2	4 weeks (drops/day a study variable; average 2–3)	None	OSDI NIBUT FBUT TMH Phenol Red LIPCOF Ocular surface staining Lipid layer grading Osmolarity (baseline visit only)	OSDI LIPCOF Conjunctival staining	NIBUT FBUT TMH Phenol Red Lipid layer grading	Artificial tears performed similarly. However, osmolarity balanced preferred in those with low baseline tear volume and lipisomal spray for those with lipid layer deficiency. Clinical Trial NCT02420834 Cochrane Risk of Bias R?C?M-O?I+S+B?
Fogt et al 2019^Citation111	Randomized, observer- masked, crossover, non-dispensing, single-center study	Omega 3 - Refresh Optive MEGA-3 Refresh Optive	N = 19 with thin lipid	46.5 ± 8.7	60 minutes (Single application)	Lipid layer thickness (overall), Symptoms with MEGA-3	None	Lipid layer thickness Symptoms	Symptoms Schirmer’s	Clinical Trial NCT03380624 - 15 min data missing Cochrane Risk of Bias R?C?M-O?I+S-B?
Fondi et al 2018^Citation112	Randomized, patient-masked, crossover single-center study	SH and trehalose - Thealoz Duo HA, trehalose and carbomer - Thealoz Duo Gel	N = 40 (for both treatment)	43.7 ± 12.3	1 week (actual 3.2 ± 2.6x/day HT & 1.9 ± 2.2x/day HTC-gel)	None	Corneal / conjunctival staining TBUT Sleep quality	Corneal / conjunctival staining TBUT Sleep quality	None	Clinical Trial NCT02980913 Cochrane Risk of Bias R?C?M-O?I?S+B?
Garcia-Lazaro et al 2011^Citation67	Randomized, investigator masked, cross‐over, single-center study	PEG 400 2.5% - Blink Intensive Tears HPMC 0.3% - Artific Tears	N = 20	57.5 ± 8.4	1 month (3x/day)	Tear meniscus volume with PEG	None	Tear meniscus volume	None	CLINICAL TRIAL NOT REGISTERED
Gensheimer et al 2012^Citation113	Randomized, double-masked, contralateral, non-dispensing, single-center study	Glycerin 1% with PLL-g-PEG - Eyeon PG 0.3% and PEG 0.4% - Systane	N = 16	44.5	120 mins (single application)	NIBUT, TBUT with glycerine	None	NIBUT with glycerine	TBUT	CLINICAL TRIAL NOT REGISTERED
Gokul et al 2018^Citation54	Randomized, double-masked, contralateral, non-dispensing, single-center study	Systane Balance Systane Ultra	N = 30	27 ± 9	30 mins (following 2.5 mins in adverse conditions)	Lipid thickness with liposomal Systane Balance	NIBUT	Lipid thickness, NIBUT	Glare acuity, temperature variation, TMH	CLINICAL TRIAL NOT REGISTERED
Grene et al 1992^Citation66	Randomized, double-masked, single-center study	CMC 1.0% - Celluvisc Lubricant HPMC 0.3% - Tears Naturale 2	N=28? N=28? severe	??	2 months (8x/day)	Symptoms, corneal erosions and impression cytology grades with CMC	Schirmer’s Corneal & conjunctival staining Lid & conjunctival swelling	Corneal staining, Symptoms, impression cytology grade (CMC only)	Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Iester et al 2000^Citation70	Randomized, open-label?, multi-center, study	HPMC 0.3% HA 0.4%	N = 55 N = 58	56.4 ± 12.8 52.2 ± 10.6	2–3 months (6x/day)	Symptoms, Tear ferning Osmolarity, impression cytology With HA	TBUT Staining Schirmer’s I	TBUT, staining, Schirmer’s I, symptoms Impression cytology	-	Ferning, osmolarity and impression cytology only measured in ~33% of sample each CLINICAL TRIAL NOT REGISTERED
Jacobi et al 2012^Citation114	Randomized, open-label? Single-center study	HP-Guar - Systane UD Tamarindus indica seed polysaccharide 1% - VISINE INTENSIV	N=14 N=14	44 ± 8 overall	3 months (5x/day)	TBUT with HP-Guar	OSDI Schirmer’s II LIPCOF Corneal & conjunctival (rose Bengal)	TBUT LIPCOF OSDI with HP-Guar	Schirmer’s II LIPCOF Corneal & conjunctival (rose Bengal)	CLINICAL TRIAL NOT REGISTERED
Jerkins et al 2020^Citation115	Randomized, double-masked, multi-center study	Systane Balance Refresh Optive advanced	N = 117 N = 114	56.7 ± 14.7 55.6 ± 16.4	35 days (4x/day)	TBUT with Systane	Symptoms	Symptoms TBUT	None	2 lipid based drops Clinical Trial NCT02776670 - exploratory lid wiper epitheliopathy and questionnaire additionally reported Cochrane Risk of Bias R+C+M?O?I-S+B?
Johnson et al 2006^Citation60	Randomized, double-masked, contralateral, single-center study	SH 0.1% SH 0.3% NaCL 0.9%	N = 13 (for all treatments)	Range 21–34	6 hours (single application)	NIBUT (0.3% SH performed better than 0.1% SH)	Symptoms	Symptoms NIBUT	None	CLINICAL TRIAL NOT REGISTERED
Johnson et al 2008^Citation56	Randomized, double-masked study, single-center study	Carbomer 934 0.3% - Lacryvisc SH 0.18% - Vismed	N = 33 N = 32	Median 36 Median 39 Range 21–64	1 month (drops/day a study variable; median 2.1–2.3)	Corneal & conjunctival staining with SH	Symptoms NIBUT TBUT	Symptoms Corneal & conjunctival staining	NIBUT TBUT	CLINICAL TRIAL NOT REGISTERED
Khaireddin and Schmidt, 2010^Citation116	Randomized, multi-center study	HA - Vismed light Phospholipid - Tears Again	N = 103 N=113 Evaporative	n=9 <25 years, n=26 25–45 years, n=42 46–60 years, n=139 >60 years	3 months 3x/day +	LIPCOF, lid inflammation NIBUT with Tears Again	Schirmer’s	LIPCOF, lid Inflammation NIBUT	Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Khanal et al 2007^Citation117	Randomized, investigator-masked, single-center study	Castor oil 0.1.25% HPMC 0.32% - Artelac Single Dose Unit	N = 27 N = 26	Unclear from text	1 month (3x/day)	Tear evaporation with HPMC	Schirmer’s, osmolarity	Tear evaporation; Lipid layer with caster oil	Schirmer’s, osmolality	CLINICAL TRIAL NOT REGISTERED
Labetoulle et al 2018^Citation118	Randomized, double-masked, multi-center study	HP-Guar - HA dual-polymer – Systane Hydration SH 0.15% - Hyabak	N = 50 N = 49	61.7 ± 12.3 56.7 ± 14.3	6 weeks (4x/day)	None	Symptoms, TBUT, ocular surface staining	Ocular surface staining	Symptoms, TBUT	Fluorescein dye only Clinical Trial NCT02470429 - exploratory end points additionally reported in n=30 Cochrane Risk of Bias R+C+M?O?I+S+B?
Lahia et al 2020^Citation119	Randomized, double-masked, single-center study	Sacha inchi microemulsion (SIME) HA 0.2%	N = 26 N = 26	53.3 ± 12.6 overall	1 month (3x/day)	Ocular protection index with SIME	Symptoms, Corneal & conjunctival staining, TBUT	Symptoms, osmolarity in hyperosmolar subgroup. Corneal and conjunctival (nasal) staining, TBUT & lid redness only with SIME	Osmolarity, Conjunctival temporal staining	Fluorescein dye only Clinical Trial NCT03569202 Cochrane Risk of Bias R+C+M+O+I+S+B+
Lee et al 2011^Citation58	Randomized, observer-masked, single-center study	CMC 0.5% - Refresh Plus SH 0.1% - Hynex	N = 33 N = 32	39 ± 14.6 37 ± 13.4	2 months (6x/day)	None	Corneal & conjunctival staining TBUT Symptoms	Cornea & conjunctival staining TBUT Symptoms	None	Fluorescein staining only CLINICAL TRIAL NOT REGISTERED
Lievens et al 2019^Citation52	Randomized, double-masked, multi-center study	CMC 1.0% and glycerin 0.9% CMC 1.0%	N = 94 N = 94	≥ 18 years of age	1 month (2x/day +)	Symptoms Corneal staining TBUT With CMC-GLY at day 7 only	Symptoms Corneal staining TBUT at all other time points	Symptoms, corneal staining, and TBUT	None	Clinical Trial NCT02280473 Cochrane Risk of Bias R+C+M+O+I+S+B+
Marner et al 1996^Citation71	Randomized, open-label, crossover, multi-center study	Carbomer gel - Lubrithal PVA 1.4% - Lacril/Liquifilm	N=54 (for all treatment)	64.3, range 38–89	2 weeks (drops/day a study variable (carbomer 3.9 vs PVA 4.6x)	Symptoms, TBUT. Instillation frequency with carbomer	Schirmer’s I Ocular surface staining, Corneal sensitivity	Schirmer’s I, TBUT, ocular surface staining, symptoms	None	Rose Bengal only used AEs 33% with carbomer, 8% with PVA CLINICAL TRIAL NOT REGISTERED
Mihaltz et al 2018^Citation43	Randomized, investigator-masked, single-center study	Carbomer, triglycerides - Artelac Lipids UD SH - Artelac Splash Edo UD	N=10 N=13	55.5 ± 11.3 53.8 ± 17.9	3 months (4x/day +)	None	Schirmer’s, TBUT, Ocular surface staining Symptoms MG dropout aberrations	Schirmer’s, TBUT, Ocular surface staining	None	Lipid drops better for those with >50% MG dropout improving Schirmer’s & aberrations CLINICAL TRIAL NOT REGISTERED
Muntz et l 2020^Citation55	Randomized, double-masked, contralateral crossover, single-center study	Lipid, PG, HP-guar and mineral oil - Systane Complete PEG 400, PG and HP-guar - Systane Ultra	N = 28 (for all treatments)	29 ± 9	Single application – adverse environment	Symptoms, lipid layer quality, NIBUT with Systane complete	TMH Hyperaemia	Symptoms, NIBUT, Lipid layer quality only with Systane Complete	TMH Hyperaemia	Clinical Trial ACTRN12619000361101 Cochrane Risk of Bias R+C+M+O+I+S+B?
Nelson and Farris, 1998^Citation120	Randomized, double-masked, multi-center study	PVA 1.4% - Liquifilm SH 0.1%	N = 16 N = 20	52.3 ± 16.4 64.8 ± 10.8	8 weeks 8x/day +	-	Symptoms, Osmolality, TBUT, rose bengal staining, Schirmer’s I, impression cytology	Symptoms, Osmolality, TBUT, rose bengal staining, Schirmer’s I	Impression cytology	CLINICAL TRIAL NOT REGISTERED
Ousler et al 2007^Citation65	Randomized, double-masked crossover, single-center study	PEG & HP-Guar - Systane CMC - Refresh Tears CMC - Refresh Endura	N = 50	62.7	Single application	TBUT, Ocular protection index with Systane	Blink rate	No comparison presented		No difference between CMC products CLINICAL TRIAL NOT REGISTERED
Park et al 2017^Citation59	Randomized, investigator-masked, multi-center study	SH 0.1% SH 0.15% SH 0.3% Cyclosporine 0.05%	N = 43 N = 41 N = 47 N = 45	44.1 ± 13.9 46.2 ± 14.0 44.8 ± 16.2 45.2 ± 15.4	12 weeks (5–6x/day)	Schirmer’s (0.15% SH group)	Corneal & conjunctival staining TBUT	Corneal & conjunctival staining TBUT	Schirmer’s	AEs 13% 0.1% SH. 20% 0.15% SH, 13% 0.3% SH, 31% 0.05% CS group. Clinical Trial KCT0001796 Cochrane Risk of Bias R+C?M-O?I+S+B?
Perez-Balbuena et al 2016^Citation47	Randomized, double-masked, multi-center study	Xanthan gum 0.09% and chondroitin sulfate 0.1% PEG 400 0.4% and PG 0.3%	N = 76 N = 72	49.9 ± 16.0 45.5 ± 12.7	2 months (4x/day)	None	Schirmer’s, TBUT, Symptoms, Corneal & conjunctival staining	Schirmer’s, TBUT, Symptoms	Corneal & conjunctival staining	Clinical Trial NCT01657253 Cochrane Risk of Bias R+C?M+O?I+S+B+
Pinto-Bonilla et 2015^Citation42	Randomized, open-label, crossover, single-center study	Trehalose and SH 1.5mg/mL -Thealoz Duo PEG & HP-guar - Systane	N = 9 N = 8	45.3 ± 11.8 53.8 ± 14.6	1 week (5x/day) (Actual 3.7±0.9 / 3.5±0.9)	None	Symptoms, Corneal & conjunctival staining, Schirmer’s, TBUT	Symptoms	Schirmer’s TBUT, Corneal & conjunctival staining	CLINICAL TRIAL NOT REGISTERED
Postorino et al 2018^Citation53	Randomized, investigator-masked, single-center study	HA crosslinked + CoQ10 HA 0.15% crosslinked	N = 20 N = 20	60.2 ± 13.6 60.9 ± 12.5	3 months (4x/day)	Symptoms, MGD assessment, corneal / conjunctival staining, epithelial hyperreflectivity and keratocytes with HA + CoQ10	Symptoms, corneal aesthesiometry TBUT	OSDI MGD assessment, corneal / conjunctival staining, epithelial hyperreflectivity and keratocytes with HA + CoQ10 only	Corneal aesthesiometry TBUT	Fluorescein staining only Clinical Trial NCT03074344 - meibomian gland assessment and confocal additionally reported Cochrane Risk of Bias R+C+M-O?I?S-B?
Pult et al 2021^Citation77	Randomized, double-masked, crossover, multi-center study	Phospholipid 0.98% - Tears Again Phospholipid 0.12% - Ocuvers	N=30 (all treatments)	33.2±1.8	Single application	Symptoms, NIBUT with high concentration lipid	None	Symptoms, NIBUT with high concentration lipid only	None	CLINICAL TRIAL NOT REGISTERED
Robert et al 2016^Citation68	Randomized, investigator masked, multi-center study	Cationic Emulsion (Cation Norm) SH 0.18% - Vismed	N = 37 N = 37 Moderate to severe	60.0 ± 14.6 65.3 ± 11.1	3 months (4x/day)	Symptoms at 1 month with SH	TBUT, Schirmer’s, Corneal & conjunctival staining, Osmolarity, Impression cytology	Symptoms, Corneal & conjunctival staining	Schirmer’s, TBUT, Osmolarity, Impression cytology	AE’s 18% CE, 27% HS >10% drop-out Clinical Trial EudraCT 2011-A00955-36 Cochrane Risk of Bias R+C?M-O?I+S+B?
Safarzadeh et al 2017^Citation121	Randomized patient-masked, single-center study	Dextran 70, 1 mg/mL and HPMC – Tears Naturale Dextran 70, 0.1 mg/mL and 0.3 g HPMC – Tearlose	N = 41 N = 47	44.1 ± 6.3 45.8 ± 8.4	4 weeks (2x/day)	None	Symptoms, TBUT, Schirmer’s Corneal & conjunctival staining	Symptoms, TBUT, Corneal & conjunctival staining	Schirmer’s	Fluorescein staining only CLINICAL TRIAL NOT REGISTERED
Sanchez et al 2017^Citation39	Randomized, investigator-masked, single-center study	CMC 0.5% (Viscofresh) HA 0.15% (Lubristil)	N = 7 N = 8	51.8 ± 14.1 71.8 ± 12.2	1 month (4x/day)	TBUT, corneal staining, and HLA-DR with CMC	Schirmer’s Other inflammatory markers	HLA-DR, TBUT & corneal staining with CMC	Schirmer’s, Tear clearance,	No Aes CLINICAL TRIAL NOT REGISTERED
Schmidl et al 2015^Citation51	Randomized, double-masked, single-center study	Trehalose and SH 1.5mg/mL -Thealoz Duo SH, 0.15% - Hyabak NaCL 0.9% - Hydrabak	N = 20 N = 20 N = 20	43.6 ± 13.3 42.9 ± 12.0 41.8 ± 9.9	240 minutes Single application	Tear film thickness (SH+trehalose to 240min and SH to 40min only)	TBUT, Schirmer’s	Tear film thickness (both SH products)	TBUT, Schirmer’s	CLINICAL TRIAL NOT REGISTERED
Simmons and Vehige, 2007^Citation78	Randomized, double-masked, crossover and parallel groups, multi-center studies	CMC 1.0% CMC 0.5% (Refresh Tears)	N = 43 single application Parallel N = 53 N = 50	Mean 62 Not stated	60 minute (single-application) 1 month (4x/day)	Ocular protection index (low viscosity to 20min, high viscosity to 30min). Corneal & conjunctival staining with higher viscosity	Symptoms	Symptoms, Corneal and conjunctival staining	None	Fluorescein staining only. More AEs with high viscosity – visual disturbance 23vs4%; discharge 13vs2% CLINICAL TRIAL NOT REGISTERED
Simmons et al 2015^Citation79	Randomized, investigator-masked, multi-center study	CMC (Refresh Optive Advanced Sensitive), unit dose CMC (Refresh Optive Sensitive), unit dose CMC (Refresh Optive Advanced Sensitive), multi-dose CMC (Refresh Optive Sensitive), multi-dose	N = 105 N = 103 N = 51 N = 56	54.4 ± 14.8 55.8 ± 14.1 55.2 ± 14.5 53.5 ± 13.9	30 days (2x/day +)	None	Symptoms, TBUT, Corneal & conjunctival staining Schirmer’s	OSDI TBUT	Corneal & conjunctival staining, Schirmer’s	No clinically significant differences in safety, effectiveness, and acceptability between lipid and aqueous artificial tears Clinical Trial NCT01459588 Cochrane Risk of Bias R+C?M-O?I+S+B?
Simmons et al 2015^Citation49	Randomized, double-masked, multi-center study	CMC 0.5% + 0.1% HA (Optive Fusion) CMC 0.5% + 0.15HA CMC 0.5% (Refresh Tears)	N = 87 N = 87 N = 90	59.6 ± 14.5 59.2 ± 16.3 60.0 ± 13.3	3 months (2x/day +) (actual 4.3, 3.9, 3.8x/day)	Some symptoms with Fusion Corneal staining with Fusion vs Refresh	Conjunctival staining	Symptoms, Corneal & Conjunctival staining	None	Investigational formulations Clinical Trial NCT01294384 - visual disturbance questionnaire additionally reported Cochrane Risk of Bias R+C+M+O?I+S-B?
Szegedi et al 2018^Citation122	Randomized, patient-masked, single-center study	SH 0.18% + triglycerides, and phospholipids SH 0.18% - Vismed sodium chloride 0.9% - Hydrabak	N = 20? N = 20? N = 20?	34.6 ± 11.7 40.5 ± 9.9 39.2 ± 12.8	40 minutes Single-application	Tear film thickness 40min vs 20min vs 0min with phospholipids	TBUT, Corneal staining, Lipid thickness	Tear film thickness, TBUT, Corneal staining, Lipid thickness	None	Clinical Trial NCT03161080 Cochrane Risk of Bias R?C?M-O-I+S+B?
Tomlison et al 2013^Citation123	Randomized, double-masked, crossover, single-center study	CMC 0.5% - Refresh Tears CMC 0.5%/castor oil - Optive Plus Glycerin 1%/castor oil - Refresh Ultra	N = 18 with dry eye N = 19 controls For all treatments	41 ± 14 30 ± 12	2 weeks 3x/day	Evaporation for both CMCs	Symptoms, TBUT, NIBUT (except for controls), osmolarity	Symptoms, evaporation, TBUT, NIBUT (except for controls), osmolarity	Lipid thickness	Measures taken after adaptation to environmental centre CLINICAL TRIAL NOT REGISTERED
Troiano and Monaco, 2008^Citation46	Randomized, patient-masked, crossover, single-center study	HA 0.4% 300mOsm/L HA 0.4% 150mOsm/L	N = 28 For all treatments	55.5 ± 7.3	7 days 4x/day	Foreign body and dryness symptoms and ocular surface staining with 150mOsm/L	None	Symptoms, hyperaemia, ocular surface staining	None	Reducing osmolarity effective Rose Bengal staining only CLINICAL TRIAL NOT REGISTERED
van Setten et al 2020^Citation124	Substitution, open-label, multi-center study	High molecular weight HA 0.15% - Comfort Shield Over habitual controls	N = 44 N = 40	57.7 ± 14.4 59.5 ± 12.5	8 weeks Actual 8.2 vs 6.5			Symptoms, Visual acuity, nerve fibre length with high molecular weight HA	Corneal staining, TBUT, Schirmer’s, Lid wiper epitheliopathy, mucotaneous junction, osmolarity	Change from habitual optimal artificial tears. No change with controls CLINICAL TRIAL NOT REGISTERED
Waduthantri et al 2012^Citation125	Randomized, double-masked, single-center study	CMC 0.5% - Refresh Tears PEG 400 0.4% / PG/HP-guar 0.3% - Systane Ultra	N = 15 N = 15	54.7 ± 5.8 55.9 ± 5.3	6 weeks 4x/day	None	Symptoms Schirmer’s TBUT, Corneal staining	Symptoms	Schirmer’s TBUT, Corneal staining	Clinical Trial NCT00796926 - meibography, osmolarity and tear meniscus height not reported Cochrane Risk of Bias R+C?M+O+I+S-B+
Wang et al 2007^Citation73	Randomized, open label, single-center study	Carbomer - Vidisic Ophthalmic Gel Cellulose - Artelac Ophthalmic Solution Mineral oil (lanolin) -Duratears Ointment	N = 22 N = 23 N = 22	55.9 ± 15.7 50.1 ± 14.3 60.3 ± 11.2	4 weeks (4x/day for Carbomer and Cellulose) (1x/day before sleep for mineral oil)	Schirmer’s with Carbomer and Cellulose & TBUT with Carbomer	Schirmer’s	Symptoms, TBUT, Schirmer’s		Fluorescein staining only, but not reported in results CLINICAL TRIAL NOT REGISTERED
Wang et al 2010^Citation126	Randomized, open label, single-center study	Carbomer + lipid gel - Liposic Ophthalmic Liquid Gel HP-guar gel - Systane Lubricant Eye Drops	N = 15 N = 15	40.4 ± 15.0 49.5 ± 12.2	2 months (4x/day)	Symptoms & Schirmer’s with Carbomer + lipid	TBUT	Symptoms Schirmer’s TBUT	None	Fluorescein staining only, but not analysed in results CLINICAL TRIAL NOT REGISTERED
Xiao et al 2008^Citation72	Randomized, investigator-masked, single-center study	Carbomer‐based 0.4% gel CMC 1.0%	N = 30 N = 30	46.7 ± 2.3 46.6 ± 2.1	3 months 3x/day +	Symptoms, TBUT, Schirmer’s, corneal staining, ocular residence time with carbomer gel	None	Symptoms, TBUT, Schirmer’s corneal staining (but no statistics presented)	None	Method relating to precorneal residence time missing. Fluorescein staining only. CLINICAL TRIAL NOT REGISTERED

Notes: Grey box = no statistical comparison made; ? = not certain from paper. Cochrane Risk of Bias Tool Rating:³⁵ “+”: low risk; “-”: high risk; “?”: unclear risk; for Random sequence generation (selection bias), allocation Concealment (selection bias); Masking of participants/researchers (performance bias), masking of Outcome assessment (detection bias), Incomplete outcome data (attrition bias), Selective reporting (reporting bias), other Biases (respectively). CLINICAL TRIAL NOT REGISTERED - based on a search of the main registries and paper. © Aston University.

Abbreviations: CoQ10, coenzyme Q10; CMC, carboxymethylcellulose/ carmellose sodium; HA, hyaluronic acid/ hyaluronan; HPMC, Hydroxypropyl methylcellulose; HP-Guar, Hydroxypropyl guar; NaCL, sodium chloride; NIBUT, non-invasive tear breakup time; PEG, polyethylene glycol; PG, propylene glycol; PVA, polyvinyl alcohol; SH, sodium hyaluronate; TBUT, fluorescein tear breakup time; TMH, tear meniscus height; LIPCOF, Lid parallel conjunctival folds.

Figure 1 PRISMA 2020 flow diagram of the systematic review search results.

Notes: PRISMA figure adapted from Liberati A, Altman D, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Journal of clinical epidemiology. 2009;62(10). Creative Commons.¹²⁷

All studies are prospective (as expected) and involve parallel groups (unless stated otherwise) of dry eye patients (diagnosed using National Eye Institute, arbitrary or recently TFOS DEWS II criteria). However, less than half (20 out of 42) are registered with a clinical trials database and even those that are have high risk of bias characteristics,³⁵ hence the certainty of the result is generally low. The lack of a definitive severity classification has been identified as a factor in differentiating the effectiveness of the available artificial tears,³¹ but previous attempts at a severity matrix table in TFOS DEWS I³⁶ led to patients being graded at different levels of severity by different tests and was abandoned in TFOS DEWS II;³⁷ severity to a dry eye patient is based on symptoms whereas it is more likely to be based on signs on the ocular surface to a cataract surgeon for example. While the intention of many of the analysed studies is to demonstrate non-inferiority compared to an established treatment, some are underpowered (see TFOS sample size recommendations)³⁷ and/or both eyes included without accounting for the correlation between the two eyes³⁸ of an individual.^39–43 In most studies, fluorescein sodium is used for assessing corneal staining (although an appropriate blue light with a peak around 395nm [not cobalt blue whose peak is ~450nm] and yellow filter with a cut off around 500nm is often not stated).⁴⁴ Most studies use lissamine green for conjunctival staining (unless otherwise stated in Table 1) which is the recommended practice,³⁷ but few state the brand which can dramatically affect the staining observed.⁴⁵ Some studies^46,47 report differences even when they do not meet the standard criteria of p < 0.05 and therefore any “difference” should be considered as noise in the data. While many trials comparing artificial tears are manufacturer initiated or sponsored, unless the research was conducted by the company or not conducted by a reputable research organisation, this should not lead to concerns regarding bias.

From the studies summarised to date (with the caveat that the effects might be affected by dry eye severity and full artificial tear formulation as well as the patient demographic) it would appear from direct comparisons between artificial tears that:

• Combination formulations are more effective than single active ingredient artificial tears.

  • The combination of CMC with hyaluronic acid is more effective than either in isolation.^48,49

  • Hyaluronic acid⁵⁰ and sodium hyaluronate⁵¹ benefit from the addition of trehalose.

  • CMC is enhanced by the addition of glycerine.⁵²

  • CoQ10 enhances the effectiveness of hyaluronic acid.⁵³

  • Newer versions of Systane (Complete and Balance) outperform earlier versions with less complexity (Ultra).^54,55

• Some studies suggest sodium hyaluronate could be more effective than CMC⁴⁰ and carbomers,⁵⁶ while others find no difference,^57,58 the optimal percentage is not clear.^59,60

• PEG containing artificial tears are more effective than those containing CMC^61–65 and HPMC.^66,67

• Cationic formulations are more effective than sodium hyaluronate (for objective signs)⁶⁸ and polyvinyl alcohol.⁶⁹

• Hyaluronic acid containing artificial tears might be better than those with HPMC,⁷⁰ but worse than those with CMC.³⁹

• Carbomer containing artificial tears might be more effective than those based on PVA⁷¹ or CMC⁷² or cellulose/mineral oils,⁷³ but less^56,74 or as effective⁴³ as sodium hyaluronate.

• Most studies recommend 4x/day use, but reported/measured use is generally less than that advised.⁴²

• Long-term compliance is needed to improve ocular surface signs rather than just symptoms⁴ and symptoms benefit from 4x/day compared to “as needed” dosing.⁷⁵

• Higher liposomal concentration increases effectiveness.^76,77

• Lower osmolarity drops increase the effectiveness of an artificial tear drop.⁴⁶

• Higher concentration (viscosity) CMC is more effective in reducing corneal and conjunctival staining, but caused more reports of visual disturbance.⁷⁸

• While drops targeting individual layers of the tear film seem equally effective,^41,79 studies have shown that the most effective drop for an individual can be predicted from their baseline classification; drops containing phospholipids are more effective in those with evaporative dry eye^3,4 and osmoprotectants benefit those with high tear film osmolarity.³

• Artificial tears may not be effective for as much as one-third of patients, but this can be predicted by one month of compliant use.⁴

These findings can inform clinical dry eye practice; in summary: non-preserved or soft preserved artificial tears being appropriate to prescribe to patients, regardless of the severity of their DED; patients with evaporative dry eye should be prescribed artificial tears containing a high concentration of liposomes; one month’s compliant use 4x/day is recommended to determine whether an artificial tear can manage the patients’ symptoms in the longer-term; signs of ocular surface disease typically take up to 4 months to start improving so patience is needed; artificial tears with multiple active ingredients (especially with PEG) seem to outperform more basic previous generation drops; ability to use different types of artificial tear bottles/sprays varies⁹ and should be part of the prescribing consideration. While the efficacy of artificial tears is well established for managing DED, its use in ocular surface disease without symptoms to improve post-surgical symptomology and to reduce refractive ‘surprises’ from poor ocular biometry⁸⁰ is less well established. The data available as reviewed in this study is limited by the definition of dry eye disease applied in published studies being variable as well as the disease severity examined and compliance with artificial tears being rarely quantified.

Other Therapeutic Functions of Artificial Tears

As well as being a management option for dry eye disease and the ocular surface, artificial tears can also be utilised for a wide range of therapeutic functions such as in the treatment of anterior eye trauma, infection, inflammation and disease as well as contact lens management.

Corneal Abrasion and Wound Healing

Corneal abrasions can be caused by foreign bodies, trauma, and trichiasis, and may result in pain, redness, lacrimation, and photophobia. Artificial tears improve epithelial healing.⁸¹ Ideally, preservative free drops are used as they tend to be associated with better ocular surface health and tolerability.⁸² The most common treatment for perioperative corneal abrasions is artificial tears followed by a combination of artificial tears and antibiotic ointment.⁸³ Most artificial tears contain hydrogels; these are known to activate the epidermal growth factor (EGF) receptor which promotes the healing of corneal epithelial wounds.⁸⁴

Pain and Inflammation Management

Artificial tears are commonly used in the management of ocular pain and inflammation. In the treatment of episcleritis, the combination of artificial tears and cold compresses provide symptomatic relief.⁸⁵ No significant differences have been observed in the signs or symptoms of idiopathic episcleritis when either artificial tears or topical ketorolac (NSAID) is used.⁸⁶ Following photorefractive keratectomy (PRK) surgery, the application of preservative-free artificial tears reduces postoperative ocular discomfort and increases visual recovery.⁸⁷ Cooled artificial tears have been shown to reduce corneal and conjunctival sensation, with 4°C being the most comfortable temperature.⁸⁸ In contrast to this, Bitton et al found no improvement in perceived patient comfort when refrigerated Systane Ultra artificial tears were used for mild to moderate dry eye sufferers.⁸⁹ It is also worth noting that pain complaints can be associated with contrasting subjective responses,⁹⁰ and in some patients artificial tears are not effective in relieving uncomfortable symptoms.⁹¹

Conjunctivitis

Allergic conjunctivitis causes ocular itching, watery discharge, lid oedema and conjunctival chemosis. Bilkhu et al exposed 18 participants (who had a known allergy to grass pollen) to grass pollen, and found that artificial tears and cold compresses improved the signs of allergic conjunctivitis and provided symptomatic relief.⁹² However, if symptoms are persistent, short-term use of topical antihistamines and mast cell stabiliser drops is recommended.⁹³

Viral (non-herpetic) conjunctivitis causes redness, discomfort, and watering. Follicles on the palpebral conjunctiva and punctate epithelial lesions on the cornea may also be observed. It has been shown that 0.5% topical ketorolac,⁹⁴ 0.45% ketorolac tromethamine,⁹⁵ and 1% prednisolone acetate⁹⁶ are no better in relieving signs or symptoms of viral conjunctivitis compared to artificial tears.

Bacterial conjunctivitis causes redness, discomfort, and produces a sticky discharge with crusting of the eyelids. Bacterial conjunctivitis usually self-resolves, but the application of artificial tears and eye bathing aids ocular comfort and hygiene. If bacterial conjunctivitis persists after 3–4 days, the application of topical antibiotics is usually recommended.⁹⁷

Keratitis

Keratitis is an inflammation of the cornea and has several different aetiologies including viral (Herpes Simplex), bacterial (marginal keratitis), fungal, contact-lens associated and unprotected exposure to ultraviolet radiation (photokeratitis). In dry eye and photokeratitis,⁹⁸ the application of artificial tears has been recommended. In herpetic keratitis, marginal keratitis, fungal keratitis, and contact-lens associated keratitis, artificial tears are advised (for lubrication and symptomatic relief) alongside additional treatment such as topical antivirals, topical and/or oral antibiotics, and antifungals.

Contact Lens Rewetting and Removal

Contact lens wearers commonly use preservative free artificial tears for ocular lubrication, comfort and contact lens rehydration.^99–101 Towards the end of wear, contact lenses become drier and fit tighter. The application of artificial tears reduces friction against the cornea and can facilitate safe lens removal.

Foreign Body Removal

Corneal foreign bodies can cause irritation, lacrimation, blurred vision, and redness. Loose foreign bodies can be irrigated away with normal saline or artificial tears. Upon successful removal of a foreign body, prophylactic antibiotics,¹⁰² analgesia and artificial tears are advised.¹⁰³

Summary

Artificial tears are the mainstay of DED management, but also have a role in corneal abrasion and wound healing, pain and inflammation management, conjunctivitis, keratitis, contact lens rewetting and removal, and foreign body removal. A review of randomized controlled trials comparing artificial tears identified 64 papers. There is good evidence that artificial tears improve symptoms of DED within a month of regular use, applied ~4x a day, but signs generally take several months. Not all patients with DED benefit from artificial tears, so if there is no benefit over a month, alternative management should be considered. Combination formulations are more effective than single active ingredient artificial tears. PEG containing artificial tears are more effective than those containing CMC and HPMC. Those classified as having evaporative DED, benefit from artificial tears with liposomes, especially of higher concentration.

Disclosure

JSW is on the executive of the Tear Film and Ocular Surface Society and the Aston University Optometry Research Group have received research funding from Alcon, the Eye Doctor, Scope Ophthalmic and Thea Pharmaceuticals. No funding was received to conduct this review. The authors report no other conflicts of interest in this work.