The Tanzanian health system is a hierarchical health system, which is in tandem with the political-administrative hierarchy.²¹ At the lowest level of health care service delivery are dispensaries and health centres. Whereas dispensaries are found in every village, health centres are found at ward level and provide services to patients referred from the dispensary level.²² At the secondary level are district hospitals. Most of these hospitals are public (owned by the government), although private hospitals, faith-based organisations (FBO) or non-governmental organisations’ (NGO) hospitals may also be the designated district hospitals in a few areas. In each region, there is a regional referral hospital, which offers additional services to those at district level, including specialist services.²³ Tertiary or zonal hospitals offer specialist care at regional or national levels. The national referral hospital has the highest level of inpatient and outpatient services. Additionally, there are also some specialised hospitals that do not fit into this hierarchy and are therefore linked directly to the Ministry of Health.²⁴

The Kilimanjaro Christian Medical Centre (KCMC) in the Kilimanjaro region, northern Tanzania, is a prominent referral and consultant hospital, serving a catchment area of over 11 million individuals. As a recognised leading facility in the region for eye care services, KCMC boasts a robust eye department, employing more than 15 ophthalmologists and 8 optometrists.²⁵ The facility serves as a referral centre for patients requiring specialised medical attention beyond the capabilities of the district and regional referral hospitals. The well-established referral system between KCMC and the surrounding district and regional referral hospitals facilitates the seamless transfer of patients in need of advanced care, ensuring timely access to specialised services and expertise available at KCMC.²⁶

Ethical clearance was obtained from the Tumaini University Kilimanjaro Christian Medical College Ethical Clearance Committee (certificate number 2401) and the University of KwaZulu-Natal Biomedical Research Ethics Committee (reference number BE436/18).

A total of 1089 records of patients presenting with corneal abnormalities between the years 2016 and 2020 were analysed. The mean age of patients was 23.5 years (±9.9, 12–45), with female patients (51.9%) and those aged 12–24 years (60.8%) being in the majority of the 198 referred patients, and nearly half (46.5%) were referred from district hospitals. Further analysis revealed a prevalence of KC among all patients aged 12–45 years, presenting with any eye condition during the study period as 0.08% (109/140 523), and 10.0% (109/1089) among those with corneal abnormalities.

There was a higher proportion of male KC patients compared to female patients (57.8% vs. 42.2%). The mean age of patients was 21.6 years (± 9.0, 12–45), with male patients (20.2 ± 8.6 years) being younger than female patients (23.6 ± 9.4 years). There was a preponderance of male KC patients in the age group of 12–24 years, while the opposite was the case in the older age groups (Figure 1).

As shown in Figure 2, more than half of the KC patients (51.4%) were residents of the Kilimanjaro region, followed by Arusha (28.4%).

It was noted that the majority of patient record cards did not contain information on case history and clinical signs, mostly recording refractive data, which did not include pachymetry and keratometry readings in most KC patient records.

Among the 23 KC patients whose data on clinical signs were recorded, the majority (52.2%) had Munson’s sign (Figure 3).

There was no significant difference between the better and worse eye regarding unaided visual acuity, refractive error as measured by SE, astigmatism and mean-K reading. The mean values of symmetry indicate that the majority of KC patients had moderate visual acuity and myopia with high astigmatism.

Among the 109 KC patients, pachymetry was performed on only 4.6% and B-scans on 5.5%. Applying the K-readings < 45D (mild), 45D–52D (moderate), 52D – 62D (advanced) and > 62D (severe), on the worse eye¹⁹ to classify KC for the 13 patients who had keratometry performed and recorded, 71.4% had moderate or worse stages of KC (Figure 4).

Male patients with corneal abnormalities and those aged < 25 years were significantly more likely to present with KC than female patients (P = 0.033) and older patients (P = 0.015). Myopia, astigmatism, refractive errors, ocular allergies and vernal keratoconjunctivitis (VKC) were significantly associated with KC (P < 0.05) as shown in Table 1.

Binary logistic regression indicates that myopia, astigmatism, VKC and ocular allergies are significant predictors of KC (Chi-Square = 94.0, P < 0.001). The other predictors, viz., gender, age, region of residence, and refractive error are not significant. All eight predictors describe 18.6% (Nagelkerke R = 0.186) of the variability of KC. Myopia, astigmatism, VKC and ocular allergies are significant at the 5% level (myopia: Wald = 9.1, P = 0.003; astigmatism: Wald = 31.4, P < 0.001; VKC: Wald = 8.8, P = 0.003; ocular allergies: Wald 12.7, P < 0.001). As shown in Table 2, the adjusted odds ratio for astigmatism (aOR = 5.66;95% CI: 3.09–10.39) and VKC (aOR = 2.68;95% CI: 1.40–5.15) were the highest.

The overall KC prevalence revealed among patients attending the KCMC eye department during the study period (2016–2020) was 0.08%. However, this prevalence could be underestimated as the study revealed poor record keeping with relevant diagnostic tests not conducted or recorded. The notable lack of advanced diagnostic equipment, such as topographers and tomographers, has most likely resulted in patients being missed or misdiagnosed. Further, very few patients underwent keratometry, topography, pachymetry and B-scans. The prevalence is similar to another hospital-based study conducted at a tertiary hospital in Ghana (0.05%),⁷ but much lower than that reported in other similar studies: 13.7% among patients presenting at a provincial eye hospital in South Africa,⁸ 0.5% among children aged up to 15 years in Nigeria,²⁷ and 0.9% among patients diagnosed with allergic conjunctivitis of all ages in Gambia.²⁸

The prevalence of KC among the 1089 patients diagnosed with corneal abnormalities aged 12–45 years was 10.0%, which is lower than that reported in a study in Syria (18.19%) among 1479 patients seeking refractive surgery,⁵ in the Jazan region, Saudi Arabia (18.72%) among patients seeking refractive surgery as detected by corneal topography,²⁹ and in Kenya among patients presenting with allergic conjunctivitis (14.6%).³⁰ However, it is higher than the 1.7% reported by Rashid et al.⁹ in their population-based study conducted among 3051 high school students aged 13–25 years in Kenya. The differences in prevalence rates among the various studies could be attributed to differences in the study populations and diagnostic modalities. Appropriate empirical data are needed, with comparative diagnostic protocols and study populations. It is recommended that population-based studies be undertaken to appropriately inform eye care planning in Tanzania.

The mean age (21.6 years) at diagnosis of KC found was indicative that the disease is more prevalent in younger populations, reportedly starting at puberty,⁹ with associated progression. It is similar to a hospital-based study conducted in Central China, Jordan, and South Africa, which found mean ages of 20.98 years,³¹ 20–24 years,^32,33 and 20.64 years,³² respectively. Keratoconus was diagnosed in a higher proportion of men than women, which was consistent with previous findings.^31,34

Munson’s sign was the most common corneal sign observed in KC patients. The possible reason for this could be that the majority of patients had moderate to severe KC, making it easily observable even with the naked eye. However, the prevalence may be underestimated as only 23 out of 109 (21.1%) KC patients in the study had records on signs as shown in Figure 3. This further highlights poor record keeping and the possibility of late presentation to the facility by patients, warranting practitioner training and KC advocacy among lower-level referral centres and within communities. Generalised late presentation was also attested to by Naderan et al.,³⁵ who reported Munson’s sign as one of the most common clinical signs in KC, observed in over 50% of KC patients, and Nkoana,³² who reported a prevalence of 21.8%. However, one study in Malaysia reported a small proportion of KC patients with Munson’s sign,³⁶ perhaps indicating better access to early diagnosis. It is advised that in countries where most patients present with Munson’s sign, a sign of moderate-advanced KC, strategies for early detection of the disease need to be developed.

The study found that the majority of KC patients exhibited moderate myopia and high astigmatism (Table 3). Progressively steepening corneal curvature can contribute to high myopia and astigmatism, leading to corneal distortion and resulting in irregular astigmatism, a known precursor to KC. Similar associations have been reported in other studies.³² High refractive errors lead to poor quality of life and require optical interventions at their early stages of development.

Univariate analysis demonstrated that directly related significant factors for the development of KC were being a male patient, resident of the Arusha region, referred from lower-level health care facilities and diagnosed with myopia, astigmatism, VKC, refractive errors and ocular allergies. Patients with other corneal abnormalities were less likely to develop KC. Multivariate analysis with binary logistic regression demonstrated that determinants of the development of KC were found to be residence in Arusha region, referral from lower-level health facilities, and diagnosis of ocular allergies and VKC. Diagnosis of other corneal abnormalities was found to be protective for the development of KC.

Although in multivariate analysis (Table 2), gender was not a determinant for the development of KC, univariate analysis in studies conducted elsewhere has reported a similar finding on gender differences, with a significant predominance of male KC patients over female patients.^{7,13,31,32,34} However, other studies have reported a significant predominance of female KC patients compared to male patients,^37,38 while some studies have not demonstrated a significant gender difference in KC patients,^5,14,33 requiring further studies with larger sample sizes.

When compared to patients aged 35–45 years presenting with corneal abnormalities, those aged 12–24 years were significantly more likely to develop KC. This finding aligns with that demonstrated in a study by Hashemi et al.³⁹ where the mean age for KC patients was significantly lower than that of non-KC patients in a population-based study in Sharoud. However, studies by Shilpy et al.³⁷ and Das et al.³¹ conducted among refractive surgery patients in India and a population-based study among high school students in Kenya³⁷ found no statistically significant difference in age between KC and non-KC patients.

Patients residing in the Arusha region were significantly more likely to develop KC. The possible reason for this could be eye rubbing, which has been shown to be a risk factor for the development of KC.^9,35 Inhabitants in the Arusha region are predominantly the Masai and Waarusha, engaged mainly in herding livestock, and during the dry season, the area is dry and dusty which may trigger allergies and eye rubbing.

Our study demonstrated that astigmatism was significantly associated with KC, with the majority of KC patients having high astigmatism. A study by Ibrahim et al.⁴⁰ on the prevalence of KC in Egyptian subjects with high astigmatism found that KC was significantly more prevalent in the group with high astigmatism than in the low astigmatism group. Similar findings were reported by Serdarogullari et al.³⁸ and Shakir et al.⁴¹ where astigmatism was also found to be a determinant of KC. As reported by Ismail et al.⁴² regarding myopia, the study found that more than one-third of the KC cases had high myopia in either eye and that myopic patients were significantly more likely to develop KC. As expected, myopia remained significant even on multivariate analysis as it has been shown that KC leads to visual impairment by inducing myopia.⁴³ Studies have shown that changes in the position and tension of the eyelid, corneal collagen fibrils, Descemet membrane, and extraocular muscles can significantly impact corneal shape. These alterations may result in orthogonal astigmatism, characterised by principal meridians of the cornea positioned at right angles, leading to variations in light refraction.⁴⁴ Understanding these relationships is crucial for diagnosing and managing conditions related to corneal shape, such as KC. Also, research has indicated a link between steeper corneal power and higher levels of myopia,³⁸ with other researchers such as Tideman et al.⁴⁵ and Ojaimi et al.⁴⁶ agreeing that the curvature of the cornea plays a modest role in the overall magnitude of myopia.

Of note was that, unlike the expected KC profile, there was poor asymmetry between the two eyes in this study (Table 3). This supports the notion that patients only present when the visual acuity in the originally better eye has worsened enough to affect functionality. Therefore, early detection strategies should be investigated and corrective strategies implemented.

Vernal keratoconjunctivitis and ocular allergies were found to be significantly associated with the development of KC in both univariate and multivariate analysis. This finding is consistent with findings in other studies.^47,48 It has been shown that atopy, which includes ocular allergies and VKC, is a result of predisposition to allergen hypersensitivity mediated by CD4+Th2 differentiation and overproduction of immunoglobulin E (IgE),^9,47 and is highly associated with eye rubbing. Keratoconus patients have a significantly higher incidence of elevated serum IgE.