A total of 174 subjects were included, of which 53 had diabetic retinopathy, then a frequency of 30.46%.

Table 3 shows the distribution of the frequency of diabetic retinopathy (DR) according to the history of peripheral neuropathy and nephropathy.

There was a statistically significant association between DR and the existence of nephropathy (p = 0.024). Patients with nephropathy were more likely to have DR than others (odds ratio 4.02 with a confidence interval of [1.13 - 14.22]).

Table 4 shows the distribution of the occurrence of DR according to the antidiabetic treatment.

There was no statistically significant association between diabetes treatment and DR.

Table 5 summarizes the distribution of the occurrence of DR according to anthropometric factors.

There was no statistically significant association between these factors and the existence of diabetic retinopathy.

HDM = Hygiene and dietetic measures; OAD = Oral antidiabetics.

NCEP* = National Cholesterol Prevention Program; IDF** = International diabetes federation.

Table 6 shows the distribution of the occurrence of DR according to the biological parameters.

Glycemic imbalance was strongly associated with the development of DR (p < 0.001). People who had glycated hemoglobin ≥ 7% were the most affected by this condition (odds ratio at 3.36 with a confidence interval of [1.60 - 7.04]). There was an association between 24-hour microalbuminuria and the existence of DR. Subjects with significant 24-hour microalbuminuria (>30 mg) were more likely to have DR (p = 0.066).

Sociodemographic characteristics

Table 7 shows the distribution of the frequency of DR by age, sex and ethnicity.

There was no statistically significant association between age, gender, ethnicity and DR.

Table 8 shows the distribution of the frequency of DR according to the history of high blood pressure and familial diabetes

There was a significant association between diabetic retinopathy and high blood pressure (p = 0.002). DR was much more common in people with hypertension (odds ratio of 3.45 with a confidence interval of [1.49 - 7.96]).

* = significative p.

* = significative p.

Table 9 shows the distribution of the occurrence of DR according to the age of discovery of diabetes, the type of diabetes and its duration.

Duration of diabetes was significantly associated with the occurrence of DR (p < 0.001). Subjects diagnosed with diabetes for more than 5 years were more affected by DR than others (odds ratio 5.02 with a confidence interval of [2.39 - 10.50]). There was also a statistically significant association between age of discovery of diabetes and DR (p = 0.029). Patients whose age was 48 years or less at the time of diagnosis of diabetes were the most affected by this condition (odds ratio 2.07 with a confidence interval of [1.06 - 4.04]).

*=significative p.

The frequency of DR was 30.46% in our study. It is close to the 33% of Nwosu et al. [6] in Nigeria in 2000 and 36.6% of Tchabi et al. [7] in Benin in 2012. However, Abouki et al. [8] in Benin in 2016, Kouassi et al. [9] in Côte-d’Ivoire in 2018 and Diallo et al. [10] in Burkina-Faso in 2014 found higher frequencies of 43.33%, 45% and 47.1% respectively. Rajoana et al. [11] in 2016 in Madagascar reported a higher frequency of 65.8%. These results illustrate the importance of DR within our populations and call for our actions to be focused on preventive measures. In contrast, Magulike et al. [12] in 2003 in Nigeria, Omolase et al. [13] in 2010 in Nigeria, Assavedo et al. [14] in 2016 in Benin, Djagadou et al. [15] in 2017 in Togo and Maammar et al. [16] in 2013 in Algeria found lower frequencies respectively 12.75%, 15%, 17.46%, 22.73% and 25.37%. The disparity in these frequencies may be related to the difference in size and to the sampling techniques that differ from one study to another.

Level of Education

In our study, the frequency of DR was slightly higher in people with a low level of education (out of school/primary) than in others (35.71% versus 28.79%). This result agrees with that of Abouki et al. [8] in Porto-Novo in Benin in 2016 who reported that this condition was predominant among illiterates (55.36%) and subjects with a primary education level (45.45%). In contrast, Cui et al. [17] in China in 2019 noted that DR was predominantly present in people with a high level of education (secondary/higher), i.e. 51.9%. In all cases, the level of education is not a risk factor for DR (odds ratio at 1.37 with a 95% confidence interval between 0.65 and 2.86; p = 0.395).

Some authors have found a higher proportion of DR among the unemployed. This is the case with Abouki et al. [8] in Porto-Novo in Benin (57.14%) in 2016 and from Ben et al. [18] in Tunisia (40%) in 2016. This result could be explained by the social situation of unemployed people who are unable to provide for correct and efficient management of their condition. It emerges from this study that occupation constitutes a risk factor for DR (for retirees, odds ratio at 2.02 with a 95% confidence interval between 1.04 and 3.93; p = 0.035).

The frequency of DR was higher in subjects taking insulin (45.45%). The same observation was made by Pirie et al. [19] in 2014 in South Africa. In 2012 in Rwanda, Giraneza et al. [20] meanwhile, reported a lower proportion than ours, 37.17%. These differences could be explained by the sample sizes of these studies. Treatment regimen is not a risk factor for DR (odds ratio at 1.97 with a 95% confidence interval between 0.79 and 4.95; p = 0.395).

Diabetic retinopathy was more represented in people with diabetic nephropathy at 41.18% (vs. 14.81%). This result is similar to that of Assavedo et al. [14] in Parakou in Benin in 2014 who reported that 21.9% (compared to 5.8%) of subjects with diabetic nephropathy developed parallel DR, but also that of He et al. [21] in China (43.06% against 25.33%) in 2012. However, Rasoulinejad et al. [22] in Iran in 2015 objected that DR was present in almost equal proportion in patients already with the stage of diabetic nephropathy (67.74%) and in those who are not (64.53%). However, diabetic nephropathy is a risk factor for DR (odds ratio 4.02 with a 95% confidence interval between 1.13 and 14.22; p = 0.024).

DR was present in 40.96% of patients with poor glycemic control. This same observation was made by Djrolo et al. [23] in Benin (94.6%) in 2014, Lopez et al. [24] in Spain (16.82%) in 2017. A UKPDS study in type 2 diabetics demonstrated the role of glycemic imbalance in the incidence and progression of DR [25]. Diabetes imbalance is indeed a risk factor for DR (odds ratio at 3.36 with a 95% confidence interval between 1.60 and 7.04; p < 0.001).

In our study, DR was predominantly present at 39.39% in diabetics who had significant microalbuminuria (>30 mg/24h). This result is comparable to those of Abouki et al. [8] in Porto-Novo in Benin (66.67%) in 2016 and Berkia et al. [25] in Morocco (53.1%) in 2014 but in higher proportions. However, microalbuminuria is not a risk factor for DR (odds ratio at 2.99 with a 95% confidence interval between 0.90 and 9.85; p = 0.066).

DR was more common in type 2 diabetes in our study at 31.14% versus 14.29%. In fact, it usually occurs after 7 years of diabetes in type 1 diabetics, and 20% of type 2 diabetics had it when their diabetes was discovered [26]. However, in 2015 in Rabat, Morocco, Andaloussi et al. [27] obtained similar results for the two types of diabetes (40.5% for type 1 and 38.9% for type 2) with regard to the occurrence of this condition. However, the type of diabetes is not a risk factor for DR (odds ratio 2.71 with a 95% confidence interval between 0.31 and 23.11; p = 0.676).

Many studies have shown an increase in the frequency of DR with the age of diabetes. Mallika et al. [28] noted in Malaysia in 2011 a faster development of DR when the age of diabetes is greater than 15 years. Cui et al. [17] in China in 2019 found as proportions of DR 15.02%, 43.59% and 66.67% respectively in subjects diagnosed for less than 5 years, 5 to 10 years and more than 10 years. The duration of diabetes is therefore a risk factor for DR (odds ratio 5.02 with a 95% confidence interval between 2.39 and 10.50; p < 0.001).

People diagnosed with diabetes at younger ages (≤48 years) had more DR (37.78%). This result was similar on that obtained in 2016 by Zou et al. [15] in China who found that young age at discovery of diabetes (<45 years) is a factor favoring the development of DR. In contrast, Ahmed et al. [29] in 2016, Saudi Arabia found that DR was more common when the age of discovery of diabetes was over 45 years (70.6% vs. 58%). This difference could be explained by the fact that this study only took into account type 2 diabetics. Thus, the age onset of diabetes is a risk factor for DR (odds ratio at 2.07 with a 95% confidence interval between 1.06 and 4.04; p = 0.029).

Based on the literature, there is a strong link between DR and high blood pressure. According to a study by the UKPDS (United Kingdom Prospective Diabetes Study), strict balancing of blood pressure in type 2 diabetics was highly beneficial because it would reduce the incidence of microvascular complications by 37% and reduce the increase in DR by 34% [9]. In Senegal in 2008, De Médeiros-Quenum et al. [30] noted that 77.77% of hypertensive patients presented with DR. Conversely, in Porto-Novo in Benin in 2016, Abouki et al. [8] objected that 50% of non-hypertensive diabetics developed DR while 40.91% of hypertensive patients were affected. In our study, DR was more found in hypertensive than in non-hypertensive (37.50% versus 14.81%). High blood pressure is also a risk factor for DR (odds ratio at 3.45 with a 95% confidence interval between 1.49 and 7.96; p = 0.002).