This multicenter cross-sectional study was conducted over a five-month period from January 1 to May 31, 2024, in the diabetology units of Garoua Regional Hospital and Garoua General Hospital. This two referral hospitals in Garoua have functional diabetology services, including therapeutic education units, nutrition services, outpatient consultation units, and inpatient wards. However, due to its status as a first-category facility within Cameroonian healthcare pyramid, Garoua General Hospital receives more socioeconomically diverse patients compared to Garoua Regional Hospital. All type 2 diabetic patients aged 35 or over who had given informed consent to participate in the study were included. Patients with a non-diabetic cause of peripheral polyneuropathy (alcoholism, chemotherapy, anticonvulsants, psychotropic, antiretrovirals drugs, anti-tuberculosis drugs…) were excluded.

For this in-hospital study, we used consecutive, non-probability, exhaustive sampling.

We obtained ethical clearance (N˚004/CERSH/NO/2024) and administrative authorizations (N˚2562/23/L/HRG/CM and N˚23-0005/AR/MSP/HGG/DG) prior to the start of the survey. Recruitment of participants began at the Garoua Regional Hospital, as this was the busiest in terms of patient attendance. Outpatients were approached by the principal investigator. After obtaining informed consent, a pre-established electronic questionnaire, prepared using Epi Data Entry 4.0.6.0 software, was administered to patients during a face-to-face exchange, enabling direct data capture during the survey. The following data were collected:

We assessed DPN using the DN-4 questionnaire, a validated screening tool for diabetic polyneuropathy. It consists of 10 items across four domains (pain characteristics, associated symptoms, pain location, and triggers): 7 interview-based and 3 physical exam-based items: 7 questioning items and 3 physical examination items scored by 0 or 1 according to the absence or presence of symptoms. A total score ≥ 4 indicated DPN, with a sensitivity of 80% and a specificity of 92% ( Table 1 ) [13] .

Data collected on Epi Data 4.0.6.0 were exported to IBM SPSS version 23 for statistical analysis. Frequencies and percentages were calculated for qualitative variables. Mean [standard deviation (SD)] and median [interquartile range (IQR)] were used for quantitative variables based on data distribution. For comparison of continuous variables, we used Student’s t test or its non-parametric equivalent. Proportions were compared using the Chi2 test. The prevalence [95% confidence interval (95% CI)] of diabetic peripheral polyneuropathy was calculated as the proportion of cases of diabetic sensitive polyneuropathy over the total number of diabetic patients tested. We used Logistic regression to investigate factors associated with DPN (age, gender, duration of diabetes, diabetes control, comorbidities, chronic complications). Univariate analysis variables with p < 0.05 were entered into a single multinomial logistic regression model to determine independent factors associated to DPN. The threshold of significance in this multivariate analysis was set at 0.05. The strength of the association was estimated using the adjusted odds ratio (ORa) and its 95% CI.

The study included 64 (31.5%) men and 139 (68.5%) women, with a male-to-female sex ratio of 0.46. The median age (IQR) of patients was 56 (48 - 63) years ranging from 35 to 80 years. The most represented age group was 51 - 60 years (35.50%). A large proportion of patients had no formal education 81 (39.90%) while only 17 (8.37%) patients had university-level education. Nearly half of the participants (45.3%) had a monthly income below 50,000 FCFA ( Table 2 ).

The median diabetes duration (IQR) was 5 (2 - 9) years. Almost half (47.29%) of the patients had had diabetes for less than 5 years, while 21.18% had diabetes for 10 years or more.

In our study population, 170 (83.74%) patients had performed a glycated hemoglobin within the last 3 months. The median (IQR) A1c was 8.80% (7.50% - 10.6%). Among these, only 14.71% had controlled diabetes (A1c < 7%), while 22.94% had moderately uncontrolled diabetes (A1c 7% - 8%), and 62.35% had poorly controlled diabetes (A1c ≥ 8%).

Most patients (65.22%) reported adhering to dietary recommendations and 50.74% practiced regular physical activity. Oral antidiabetic drugs were used by 72.41% of the patients.

Fundus examination was performed in 60.10% of the sample. Among them, 13 patients (10.66%) had diabetic retinopathy. Of the 158 patients with serum creatinine measured, 19.7% had an estimated creatinine clearance < 60 ml/min (Cockcroft formula) ( Table 3 ).

With regard to cardiovascular risk factors, 43.8% of patients had arterial hypertension. A total of 11.3% of patients reported regular alcohol consumption. Only 2.9% of patients were tobacco smokers, while 1.4% had quit smoking at least 3 years previously ( Table 4 ).

Table 5 summarizes the anthropometric parameters of our study population. The median waist circumference (IIQ) in our sample was 91 cm (82 - 102 cm). Women had a median waist circumference (IIQ) of 89 cm (78 - 98 cm) and men 99 (92 - 107 cm).

Participants’ median BMI (IIQ) was 26.51 (22.75 - 30.83) kg/m² and 56.65% patients had a BMI ≥ 25 kg/m².

Using the American NCEP-ATP III classification for the gender distribution of our sample, we note that 47.7% of patients were android obese ( Table 5 ).

BMI: Body mass index.

Burning was the main symptom, with a frequency of 90.3% in the DPN population. Followed by tingling sensation, present in 80.6% of patients ( Table 6 ).

The prevalence (95% CI) of DPN in our study population was 45.81% (38.82 - 52.96)%.

Multivariate analysis was carried out on a sample of 101 patients, due to the fact that some of our patients did not undergo complementary examinations. The independent factors associated [ORa (95% CI), p] with DPN were female gender [3.65 (1.62 - 8.23), p = 0.037]; age ≥ 60 years [4.57 (2.11 - 9.86), p = 0.048]; duration of diabetes progression [2.93 (1.16 - 7.45), p = 0.006]; hyperglycemic imbalance with HbA1c ≥ 7% [3.99 (1.25 - 5.72), p = 0.043] and chronic renal failure [4.79 (2.22 - 7.54), p = 0.009] ( Table 7 ).

In our study, the prevalence (95% CI) of DPN was 45.81 (38.82 - 52.96)%. This result is similar to the results reported by Dehayem et al. in Cameroon on a sample of 376 patients and Aynaou et al. in Morocco on a cohort of 391 patients, which were 44.1% and 40% respectively. Both used the same diagnostic tool, the DN-4 questionnaire [8] [11] .

Studies using other diagnostic criteria reported higher prevalence rates. Using the Michigan Neuropathy Screening Instrument (MNSI), Elleuch et al. found a prevalence of 57.8% [4] . Conversely, studies that employed electroneuromyography (ENMG) found lower rates. Mojaddidi et al. reported a prevalence of 16.4% in Saudi Arabia using ENMG, identifying subclinical DPN [14] .

Thus, diagnostic tools significantly influence reported prevalence rates. This was confirmed in a 2022 Saudi Arabian meta-analysis by Owolabi et al., which included 12 studies using different diagnostic methods. Studies relying on questionnaires and clinical exams showed a pooled prevalence of 48% [46% - 50%], while those using nerve conduction studies reported a prevalence of 26% [15% - 36%] [6] .

The study of associated factors with DPN concluded that the female sex is involved. Many authors reported similar results. These include Hode et al. (p = 0.022) and Halawa et al. (p = 0.001) [15] [16] . However, some studies reported opposite results. Sani et al. found an association between male sex and DPN in their study (p = 0.017), explained by the predominantly male prevalence in their population [17] .

We noted a significant relationship between advanced age (60 years) and DPN. This has also been reported by Sani et al. and Halawa et al. [16] [17] .

In univariate analysis, both overweight and abdominal obesity were significantly associated with DPN, likely due to sensory dysfunction in small nerve fibers [18] . However, in multivariate analysis, BMI and waist circumference were not independent predictors.

Long-standing diabetes is a well-established major associated factor. Moreover, there was a positive correlation between the duration of diabetes and DPN, with a clearer correlation from 10 years onwards. This was demonstrated by logistic regression. The correlation has been confirmed by several studies, notably those by Bassanguen et al. and Halawa et al. [16] [19] .

The role of hyperglycemia is clearly established in the onset and aggravation of DPN. Vinik et al. have demonstrated that poor glycemic control (A1c ≥ 7%) increases the risk of developing DPN by more than fivefold. According to our work, hyperglycemic imbalance is linked to the occurrence of DPN, with an aOR = 3.99, in case of A1c ≥ 7%. Similar results reported by Hode et al. [15] .

In our study, chronic kidney disease was a significant predictor of the occurrence of DPN. Bassanguen et al. and Sani et al. highlighted this relationship, which is similar to our data [17] [19] .

Many patients faced financial barriers to completing certain paraclinical investigations. Thus, almost all the data collected was based on a pre-established questionnaire, completed during the consultation, with all the hazards inherent to this environment, notably the often stressful hospital atmosphere, which could alter the quality of patients’ answers. Screening for DPN using this questionnaire was based on subjective responses, influenced by the patient’s psychic disposition at the time of the interview.

The exclusive use of the DN-4 questionnaire to diagnose DPN could be responsible for overestimating the prevalence of DPN. The use of nerve conduction studies, even on a sub-sample of participants, would have improved diagnostic accuracy and strengthened the validity of the results.