Six of the pages which were found to be ranking for this question state the same two things. Preservatives irritate the eye. Redness-relief drops cause the redness to return after use worse than before the bottle was opened.
Between them, those two pages contain approximately twelve thousand words and cite, in a form the reader could actually go and look up, not one peer-reviewed paper. This would be less concerning if the two claims had equal support, but they most certainly do not. One has a 4,107-patient survey and a visible dose-dependence to its effect, while the other is unsubstantiated case reports and the only controlled experiment ever done on the two molecules in question found no rebound effect at all.
This is the part where most people would have made the wrong choice, and the shelf offers them no help.
The two bottles are not the same product
Lubricating drops, sold as artificial tears, provide volume and viscosity to the tear film. Decongestant drops constrict the blood vessels in the conjunctiva, lightening the eye for a few hours, and have no effect on dryness. Two different problems, two different molecules, one shelf.
Same 10 ml plastic bottle. Frequently the same manufacturer, often the same packaging.
If your eyes have been dry and scratchy by the fourth call in the afternoon and you have been reaching for the bottle that promised to get rid of the redness, you have been using the drug class for the wrong diagnosis. This is worth knowing before you proceed to read further.
The best evidence for the effects of preservatives comes not from the artificial tears but from glaucoma patients, who have been using the same drops every day for years, meaning that any cumulative effect would be visible in the population.
What benzalkonium chloride does, and at what dose
A prospective survey published in 2002 in the British Journal of Ophthalmology and covering 4,107 patients with assessment by 249 ophthalmologists found the rates of discomfort on instillation to be 43% for preserved drops and 17% for preservative-free drops, burning or stinging to be 40% vs 22%, and dry eye sensation to be 23% vs 14%, all at p<0.001 (PubMed).
Two notes in particular stand out from this paper, the first being that the higher the number of preserved drops a patient used, the higher the chance they were to report any of the above symptoms, which is a sign of a dose-dependent effect. The second is that the symptoms and signs fell in patients who switched to preservative-free drops, meaning that the majority of the effect was not permanent.
The paper itself is careful to note the limitations of its design: both patients and the examiners knew which drops the patients were using, so the ones who switched could easily have believed their symptoms to be improving when they were not, and the selection bias is evident, with the patients who enrolled in the study being more likely to have concerns about their medication. The lack of an untreated control group of similar age and the inability to control the dosing as one would in a clinical trial are also worth noting.
And finally, the confounding factor which may explain all of the above. The concentrations and the dosing regimens were glaucoma-appropriate and used daily for years, while the paper does not claim to be able to extrapolate its results to the use of a lubricant twice a day.
The goblet-cell study which is frequently invoked is much more specific, much more narrow, and much more limited.
A 2021 in vitro study published in Biomedicine Hub exposed human conjunctival goblet cells cultured from at least three donors to four different commercial eye drops containing 0.005-0.02% benzalkonium chloride for 30 minutes and six hours of exposure (PubMed).
Six hours in a dish is not four drops a day
Goblet cells are responsible for producing mucin, which is one part of the tear film necessary for its even distribution over the ocular surface. Their loss is associated with the disruption of the tear film, not direct irritation.
The results of the study show that at 30 minutes, survival was at 93% for the 0.005% solution and 69% for the 0.02% solution, and at six hours, 74% and 31%, respectively. The survival rates decreased with increased concentration and exposure time. Finally, the 0.005% solution did not significantly decrease survival at either time point.
The most important caveat, as noted by the authors themselves, is that the cells were exposed to a constant concentration of the solution, while in reality, the tear film is constantly diluted and refreshed. The comparison between different eye drops was done using solutions with different active ingredients diluted in cell culture medium, not in tears, and the results were obtained using cells from only three donors. The authors themselves advise caution in clinical extrapolation and further research.
The most honest summary of the results would be that benzalkonium chloride kills these cells, and the higher the concentration and the longer the exposure, the more cells die. What effect, if any, this has on a person who uses two drops a day is a question this study does not answer, and it does not pretend to answer it.
The claim is that the vasoconstrictors naphazoline and tetrahydrozoline, found in decongestant drops, cause the blood vessels to dilate beyond their initial state when the vasoconstrictive effects wear off, causing the rebound effect.
This is a reasonable claim to make. It is repeated frequently, and it has been tested, if not thoroughly.
The rebound story, and the experiment nobody cites
A 1984 study published in Ophthalmology and conducted on 11 healthy volunteers tested the effects of naphazoline 0.02% and tetrahydrozoline 0.05% on the conjunctival redness, both as a single dose and as a ten-day treatment, measuring the degree of whitening, its duration, tolerance, and rebound effect (PubMed).
Both vasoconstrictors caused the redness to decrease below baseline after a single dose, with naphazoline being more effective than tetrahydrozoline, and the effect lasted for eight hours. After ten days of treatment, only naphazoline remained effective, with the authors noting that the decreasing efficacy of tetrahydrozoline may lead to its misuse.
On the subject of the rebound effect, their conclusion is as follows: neither of the tested vasoconstrictors produced a rebound vasodilation after discontinuation.
11 people, healthy eyes, ten days of treatment, redness measured by eye rather than by photography or other objective measures.
A single negative result does not invalidate a claim, but it does make it weaker, and this result, along with the lack of other evidence, is not enough to support the claim of a rebound effect. It does, however, rob the pages repeating this claim of the ability to use the word "proven". The rebound effect is a much less documented issue than the preservatives, and the few pages which describe it do not go into enough detail to make the claim reliable.
The more documented issue, the one which is described in greater detail and with more supporting information, is the fact that the vasoconstrictive effect simply wears off, and the drops which promised to get rid of the redness stop working, becoming a temptation to use them again.
Read the label you already own
Turn the bottle over.
If the active ingredient is listed as naphazoline, tetrahydrozoline or brimonidine, you are looking at a decongestant, regardless of what is written on the front of the carton. Hypromellose, carboxymethylcellulose, glycerin or polyethylene glycol indicate lubricating drops.
The preservative, if there is one, is listed under the inactive ingredients, and benzalkonium chloride is not a short name for anything, as it does not have any. Single-dose vials, the small twist-off ampoules which are sold in strips of five or ten, contain no preservatives at all, which is why they are sold that way and why they are frequently much more expensive than their multi-dose counterparts.
A few minutes spent reading the fine print will tell you which of the two issues mentioned above is relevant to you.
A herb on the front does not change the back
Household eye care in India is not particularly exotic. Rose water dabbed on the eyelids after a long day, a cold compress, a triphala rinse in a household which keeps triphala in the kitchen. Most people doing these things would not consider them to be treatment.
Two things are worth noting. First, that home-made rinses and even rose water are not sterile, and eye-care professionals advise against using non-sterile liquids in the eye, which is worth considering. Second, that a commercially available Ayurvedic preparation is a manufactured multi-dose preparation, and as such, it is subject to the same preservative requirements as any other multi-dose preparation.
The traditional preparation noted on the front of the carton is not exempt from the chemistry noted on the back.
What this actually leaves you with
The preservative issue is a strong one, with a clear mechanism of action, one large human survey, and a few limitations which apply to your situation specifically. The rebound issue is weaker, unsubstantiated by controlled trials, and better documented in its lack of evidence.
None of this, however, adds up to a specific number of drops.
The four-a-day threshold which will be mentioned as the point at which preservatives become a concern is a clinical recommendation, not a research result, and anyone citing it as such is filling a gap the research has not filled.
If you are on a prescription eye drop, none of this is a reason to stop using it, and the decision to change or stop should be made in consultation with the physician who prescribed it. The 2002 survey which was mentioned several times above was a comparison between drops used by the patients, never about whether to treat at all.
Stinging after instillation is normal and should pass within a few seconds. If it persists for more than half a minute, or if the drops have made your eyes drier than before you instilled them, it is worth booking an appointment rather than reaching for the next bottle on the shelf.
