· The dependent variable was non-adherence to therapy assessed using the Girerd’s scale [9] .

· The independent variables were:

○ Socio-demographic and economic characteristics: Age, sex, professional status, marital status, income, distance travelled.

○ Clinical data: Time of consultation, previous treatment, reason for consultation, mode of admission, general condition, blood pressure, heart rate, stages of chronic kidney disease, heart rate.

○ Therapeutic data: Quantity of medications.

· Operational definitions of variables were:

○ CKD was diagnosed according to the classification criteria for chronic kidney disease [8] (KDIGO, 2012). Stages 2, 3, 4 and 5 were retained for a glomerular filtration rate (GFR) between 60 and 89 ml/min/1.73 m², 30 and 59 ml/min/1.73 m², 15 and 29 ml/min/1.73m² and less than 15 ml/min/1.73 m² respectively.

○ The chronic kidney disease epidemiology collaboration (CKD-EPI) equation was used as the basis for calculating GFR [8] .

○ Past treatment was the type of treatment used prior to the consultation which could be phytotherapy when it is the practice of phytotherapy; traditional when it is the use of a traditional healer who uses rituals and invokes spirits to heal the patient. This later treatment can be accompanied by phytotherapy and medication when instituted by a health professional;

○ The consultation delay is the time elapsed between the onset of symptoms and the first contact with health services.

○ Blood pressure was measured in a seated subject after 15 minutes of rest, using an automatic digital blood pressure monitor fitted with a cuff (belonging to the list of blood pressure monitors validated by the French Society of Hypertension). It was taken on the bare right arm, and placed on a table with the palm facing upwards. Three measurements were taken at five-minute intervals; the average of the last two measurements was the blood pressure retained for each patient. Blood pressure is considered high when it is greater than or equal to 140/90 mmHg [10] .

○ Heart rate (HR) was considered normal for a number of heartbeats between 60 and 100 beats (bpm) per minute. It was taken 15 minutes after the patient’s arrival, on the bare right arm, placed on a table, palm upwards, with the same automatic digital blood pressure meter fitted with a cuff used for blood pressure measurement.

○ The Girerd’s [9] scale is used to determine therapeutic non-adherence. The items on Girerd’s [9] scale are: forgetting to take medication, running out of medication, delay in taking medication, amnestic stroke, feeling that treatment is doing more harm than good, and feeling that there are too many pills to take. It is validated in a patient when the number of affirmative responses is greater than or equal to 3. A total of 0 indicates good compliance. A total of 1 or 2 indicates a minimal compliance problem, and a total of 3 or more indicates poor compliance. In this study, the first two modalities (good compliance and minimal problem) were categorized as good compliance and the last one (poor compliance) as non-compliance.

○ The number of medications was the amount of oral medication patients were taking on the day of the adherence assessment.

Data were collected using a medical record tabulation form, covering patients’ socio-demographic, economic, clinical and therapeutic variables. Data were analyzed using R 4.3.0 software. Quantitative variables were expressed as means with standard deviations, and qualitative variables as proportions. Chi² and Fisher tests were used to determine the association between two qualitative variables. Student’s T-test was used to compare means. A statistically significant association was established for a p-value < 0.05. The variables associated with the univariate analysis were introduced into the multivariate analysis model. Potential predictor variables were identified from binary logistic regression analysis, with odds ratios (OR) and their 95% confidence intervals (95% CI) determined.

The survey protocol was validated by the Public health unit of the Cotonou Faculty of Health Sciences. Collection authorizations were obtained from the aforementioned unit and from the Head of the CHUD/B-A Nephrology Department. The anonymity and confidentiality of the data collected complied with the ethical principles applicable to medical research on human subjects contained in the Declaration of the World Medical Association of Helsinki.

The mean age of patients was 50 ± 14.9 years (CI 95% [48.4 - 51.6]), with extremes of 13 and 90 years. Patients aged between 40 and 60 were the most numerous (47.2%). Over half the patients (66.7%) were male, giving a sex ratio (M/F) of 2. The majority (36.8%) had no formal education. In terms of monthly income, 163 patients (47.2%) had monthly revenue of between 52,000 - 150,000 FCFA, most of them having come from less than 50 km away (61.7%) (Table 1).

Some 275 patients, or 79.7% of the total, were admitted with an altered general condition. According to the mode of admission to the department, 65.5% of patients were referred by consultants, the majority of whom (69.9%) were general practitioners. More than half of patients (71.6%) came for consultation at stage 5 of the disease. More than three-quarters (85.5%) of patients had received prior treatment before consulting the nephrology department. Of these, 65.2% had opted for a previous herbal treatment (phytotherapy). All patients (100%) received therapeutic education prior to treatment. The average number of oral medications taken by patients was 4.3 ± 1.7, with extremes of 2 and 9 molecules. Most patients were on 4 molecules. The average number of molecules differed statistically (p = 0.004) between the group of patients with good compliance (4.6 ± 1.7) and those with poor compliance (4.1 ± 1.7) (Table 2).

Evaluation of the various items on the Girerd’s scale showed that 60.6% of patients were late in taking their medication in relation to the usual times, and 58% mentioned medication breakdowns. The evaluation of therapeutic compliance showed that 197 patients were noncompliant, giving a prevalence of non-compliance equal to 57.1% (Figure 1, Table 3).

^a1 FCFA = 0.0016$.

The prevalence of therapeutic non-adherence was 57.1% at CHUD/B-A in 2022. This result is similar to the one of Yousfi et al. [11] in Tunisia in 2012, who reported prevalences ranging from 20% to 60% and depending on the number of medications taken per day. Similarly, Bettaieb et al. [12] , measuring therapeutic compliance in chronic hemodialysis patients in Tunisia, and Vigan et al. [7] in Cotonou reported prevalences of 57.35% and 48.67% respectively. However, our result was lower than the 89.2% found in Togo by Sabi et al. [13] and higher than the 10.4% reported by Parmier et al. [14] in France in 2012.

In our study, the factors associated with NOT in CKD patients were: income, mode of admission, number of drugs, and previous treatment with herbs or traditional healers. Our results are similar to those of Seck et al. [15] , who reported that factors contributing to poor compliance included adverse drug reactions and high drug costs.

Our study did not enable us to assess compliance with treatment in terms of follow-up appointments, prescription and paraclinical test renewals, hygienic and dietary measures, psychosocial aspects or quality of life. For a chronic disease with long-term treatment, the evaluation of these parameters could provide a holistic assessment of the handicaps to therapeutic compliance. A qualitative assessment of the reasons for poor compliance was also lacking.