Pterygium is a wing-shaped fibrovascular growth of the bulbar conjunctiva that encroaches on the cornea. Its prevalence is highest in the so-called pterygium belt within about 40 degrees of the equator, and chronic exposure to ultraviolet light is the main environmental cause [1,2]. A dose-response relation between sun exposure and pterygium has been shown [3], and outdoor work in early life carries the greatest risk [4]. Population studies from Asia have reported a high prevalence, particularly among older men, outdoor workers and rural residents [5-9]. Agricultural workers in rural Maharashtra share these risk factors.

Pterygium is now considered a disorder of the limbal stem cells. Altered limbal epithelial basal cells invade the cornea and express matrix metalloproteinases that dissolve Bowman's layer [10,11]. This explains why simple excision, which leaves the limbal defect bare, is followed by vigorous fibrovascular regrowth. Recurrent pterygium is usually more aggressive than the primary lesion and can cause astigmatism, restriction of ocular motility and symblepharon [12,13].

Many techniques have been described to reduce recurrence. The bare sclera technique is simple and quick but has been associated with high recurrence rates for decades [14]; reported rates range from about 24% to 89% [12]. Conjunctival autograft transplantation, popularised by Kenyon and colleagues [15,16], covers the bare sclera with healthy conjunctiva and restores a barrier at the limbus. Controlled trials in tropical populations showed much lower recurrence after conjunctival autograft than after bare sclera excision [17,18]. Adjuncts such as mitomycin C and amniotic membrane grafts have also been compared with conjunctival autograft [19-22]. Despite this evidence, bare sclera excision is still performed in many rural hospitals in India because it is quicker and needs less skill. The aim of this study was to compare the recurrence rate, time to recurrence and complications after pterygium excision with conjunctival autograft and with the bare sclera technique at a rural tertiary care hospital.

Study design and setting This prospective comparative study was conducted in the Department of Ophthalmology, Pravara Institute of Medical Sciences, Loni, a tertiary care teaching hospital serving a predominantly agricultural population. Written informed consent was obtained from every patient before surgery. Participants Patients aged 18 years or more with primary nasal pterygium extending at least 2 mm onto the cornea, or causing visual disturbance, cosmetic concern or chronic irritation, were included. Patients with recurrent or double-headed pterygium, pseudopterygium, glaucoma or previous ocular surgery, dry eye, active ocular surface inflammation, or a condition that might require the superior conjunctiva for future glaucoma surgery were excluded. One eye of each patient was included. Sample size Assuming a recurrence rate of 10% after conjunctival autograft and 35% after bare sclera excision, 43 eyes per group were required to detect this difference with 80% power at a 5% significance level, or 48 eyes per group allowing for 10% loss to follow-up. Sixty eyes were included in each group, giving a total of 120 eyes. Preoperative assessment All patients underwent recording of uncorrected and best corrected visual acuity, slit lamp examination, keratometry and dilated fundus examination. The horizontal extent of the pterygium onto the cornea was measured with the slit lamp beam. Pterygium morphology was graded as atrophic, intermediate or fleshy according to the visibility of the underlying episcleral vessels [18]. Surgical technique All operations were performed under local anaesthesia by experienced surgeons using an operating microscope. The head of the pterygium was dissected from the cornea and the body was excised together with the underlying Tenon's tissue, leaving an area of bare sclera of about 5 to 7 mm. The corneal bed and limbus were smoothed with a blade, and haemostasis was achieved with gentle wet-field cautery. In the bare sclera group, the defect was left uncovered. In the conjunctival autograft group, the size of the defect was measured with calipers and a free graft of thin conjunctiva, about 1 mm larger than the defect in each dimension and free of Tenon's tissue, was harvested from the superotemporal bulbar conjunctiva [15,16]. The graft was transferred with its limbal edge oriented to the limbus and its epithelial side up, and secured with interrupted sutures. The donor site was left to heal. No antimetabolite was used. Postoperatively, all patients received topical antibiotic and corticosteroid drops tapered over 6 weeks. Follow-up and outcome measures Patients were examined on day 1, at 1 week, and at 1, 3, 6 and 12 months after surgery. Recurrence was defined as regrowth of fibrovascular tissue extending 1 mm or more across the limbus onto the cornea at the site of excision; conjunctival recurrence without corneal involvement was recorded separately. The time to recurrence was noted. Complications recorded included graft oedema, graft retraction, subconjunctival haemorrhage, granuloma, dellen, scleral thinning and suture-related irritation. Keratometric astigmatism was measured before surgery and at the final visit. The primary outcome was the proportion of eyes with corneal recurrence within 12 months. Secondary outcomes were time to recurrence, the association of recurrence with age, sex, occupation, morphology and size of pterygium, complications and change in astigmatism. Statistical analysis Data were entered in Microsoft Excel and analysed with Python version 3.13 using the SciPy and statsmodels libraries. Categorical variables were compared with the chi-square test, or the Fisher exact test when any expected cell count was below 5, and continuous variables with the independent or paired t-test. Recurrence-free survival was compared with the log-rank test. Variables with p < 0.20 on univariate analysis were entered into a multivariable logistic regression model, and results were expressed as adjusted odds ratios (aOR) with 95% confidence intervals. A p-value below 0.05 was considered significant.

A total of 120 eyes of 120 patients were included: 60 in the conjunctival autograft group and 60 in the bare sclera group. All eyes completed 12 months of follow-up. The mean age was 45.8 ± 15.9 years, 52.5% were men and 46.7% were outdoor or agricultural workers. The two groups were similar in age, sex, occupation, morphology and size of pterygium and preoperative astigmatism (Table 1).

Table 1. Baseline characteristics

SD: standard deviation; D: dioptre. Chi-square test for proportions; independent t-test for means.

Corneal recurrence within 12 months occurred in 11 eyes (18.3%) after conjunctival autograft and in 31 eyes (51.7%) after bare sclera excision (p < 0.001). The difference was already significant at 1 month and remained so at every visit. Conjunctival recurrence without corneal involvement was seen in 2 and 3 eyes respectively (Table 2).

Table 2. Recurrence after surgery

Cumulative numbers. Chi-square or Fisher exact test; independent t-test for time to recurrence among eyes with recurrence.

Most recurrences appeared early: of the 42 recurrences, 57.1% were detected by 3 months and 71.4% by 6 months, with a mean time to recurrence of 5.5 ± 4.4 months. Recurrence-free survival at 12 months was 81.7% after conjunctival autograft and 48.3% after bare sclera excision (log-rank p < 0.001). Recurrence was more common in outdoor or agricultural workers (48.2% versus 23.4%) and was somewhat more frequent in fleshy pterygia (42.6% versus 30.1%), but was not related to age below 40 years, sex or corneal extent (Table 3).

Table 3. Univariate analysis of factors associated with recurrence

Chi-square test.

On multivariable logistic regression, bare sclera excision (aOR 5.34, 95% CI 2.20 to 12.97; p < 0.001) and outdoor or agricultural work (aOR 3.00, 95% CI 1.29 to 6.99; p = 0.011) were independent risk factors for recurrence, while fleshy morphology was of borderline significance (aOR 2.21, 95% CI 0.93 to 5.29; p = 0.074) (Table 4).

Table 4. Multivariable logistic regression of risk factors for recurrence

Variables with p < 0.20 on univariate analysis were entered into the model. OR: odds ratio; CI: confidence interval. Model fit: Hosmer-Lemeshow p = 0.19.

Complications were recorded in 27 eyes (45.0%) after conjunctival autograft and 4 eyes (6.7%) after bare sclera excision. In the autograft group they were minor and self-limiting: suture irritation in 12 eyes, subconjunctival haemorrhage in 10, graft oedema in 9, graft retraction in 3, dellen in 1 and granuloma in 1. No graft was lost. In the bare sclera group, dellen developed in 3 eyes and scleral thinning in 1 (Table 5). Mean keratometric astigmatism fell from 2.31 D before surgery to 1.17 D at the final visit (paired t-test p < 0.001), with no difference between the groups.

Table 5. Postoperative complications

In this prospective study from a rural tertiary care hospital, recurrence after pterygium excision with conjunctival autograft was 18.3%, compared with 51.7% after bare sclera excision. After adjustment, bare sclera excision increased the odds of recurrence about five-fold. Most recurrences occurred within the first 6 months, and outdoor or agricultural work was an independent risk factor. The high recurrence after bare sclera excision in our study agrees with the literature. Youngson described frequent recurrence after simple excision [14], and Hirst's review found reported rates between about 24% and 89% [12]. In a controlled trial in Singapore, Tan and colleagues found far more recurrences after bare sclera excision than after conjunctival autograft, and fleshy pterygia recurred most often [18]. Lewallen showed the benefit of conjunctival autograft in a randomised trial in the tropics [17], and long-term results from Turkey also showed the highest recurrence rate after bare sclera excision [23]. Our recurrence rate of 51.7% after bare sclera excision lies within this range and reflects the high ultraviolet exposure of our largely agricultural population. Our recurrence rate of 18.3% after conjunctival autograft is higher than the lowest rates in the literature but still much lower than after bare sclera excision. Kenyon and colleagues reported a low recurrence rate with this technique for advanced and recurrent pterygium [15], and Prabhasawat and colleagues found that conjunctival autograft gave fewer recurrences than amniotic membrane graft or primary closure [20]. A randomised trial in Thailand also found fewer recurrences after conjunctival autograft than after amniotic membrane graft [22], while Ma and colleagues found amniotic membrane graft, conjunctival autograft and mitomycin C to give comparable results [21]. Results vary between centres, and Ti and colleagues showed that the success of conjunctival autografting depends on surgical experience and technique [24]. A thin graft free of Tenon's tissue, correct orientation of the limbal edge and wide excision of the subconjunctival fibrovascular tissue are important [16]; with extended removal and a large graft, Hirst achieved very low recurrence rates [25]. Refinement of our technique along these lines could reduce our recurrence rate further. Mitomycin C is an alternative adjunct, but its use has been associated with scleral thinning and other sight-threatening complications. In randomised trials, conjunctival autograft and mitomycin C gave comparable recurrence rates [19,26], and limbal conjunctival autograft was also effective for recurrent pterygium [27]. Conjunctival autograft avoids the risks of antimetabolites and should be preferred in primary pterygium [13]. The complications after autograft in our study were minor, mostly suture irritation and subconjunctival haemorrhage, and none threatened the eye. Fibrin glue can shorten surgery and reduce discomfort compared with sutures [28], and might reduce suture-related irritation, but it adds cost. Recurrence was an early event in our study, with 71.4% of recurrences detected by 6 months. In previous series, most recurrences appeared within the first few months, and nearly all within one year [29,30]. A follow-up of 12 months is therefore sufficient to compare techniques, and patients should be reviewed closely during the first 6 months. Outdoor or agricultural work tripled the odds of recurrence in our patients, which agrees with the role of ultraviolet exposure in the development of pterygium [3,4]. Fleshy pterygia, which recurred more often in previous work [18], showed a similar trend in our study. Young age was not a risk factor in our patients. The limbal stem cell theory of pterygium [10,11] explains why restoring a limbal barrier with a graft is more effective than leaving bare sclera. These findings have clear implications for practice in rural India. Conjunctival autografting requires an operating microscope and some additional skill, but it is inexpensive, uses the patient's own tissue and adds only a few minutes to surgery. Training residents in this technique and discouraging bare sclera excision would reduce the burden of recurrent pterygium, which is more difficult to treat. Outdoor workers should be advised to use sunglasses and hats to reduce ultraviolet exposure after surgery. Strengths and limitations The strengths of this study are its prospective design, complete 12-month follow-up of all eyes, a clear definition of recurrence, and grading of pterygium morphology. It also has limitations. Allocation was not randomised and the examiners were not masked to the technique. More than one surgeon performed the operations. Follow-up of 12 months may miss a small number of late recurrences, and the effect of the procedures on the tear film and ocular surface was not studied.

Pterygium excision with conjunctival autograft gave a significantly lower recurrence rate than bare sclera excision (18.3% versus 51.7%), with only minor complications. Most recurrences appeared within 6 months, and outdoor or agricultural workers were at greater risk. Bare sclera excision should no longer be used for primary pterygium, and conjunctival autograft should be the standard procedure in rural hospitals.

DECLARATIONS

Acknowledgement: The author thanks the faculty, residents, nursing staff and operation theatre staff of the Department of Ophthalmology, Pravara Institute of Medical Sciences, Loni, and all the patients who took part in the study.

Funding: None.

Conflict of interest: The author declares no conflict of interest.

Author contribution: Dr. Shyam Popatrao Bawage designed the study, operated on and followed the patients, analysed the data, wrote the manuscript and approved the final version.

Data availability: The data supporting the findings of this study are available from the corresponding author on reasonable request.