The following decompensation factors were investigated:

Data were collected, entered, and analyzed using SPSS version 30.0. Qualitative variables were expressed as proportions, while quantitative variables were presented as means ± standard deviation. For ethical considerations, all data were collected anonymously, and the free and informed consent of all participants was systematically obtained. The information gathered was used exclusively for scientific purposes.

During the study period, a total of 120 diabetic patients were hospitalized, among whom 52 cases of diabetic ketoacidosis (DKA) were identified, representing a prevalence of 43%.

The study population consisted of 22 men and 30 women, yielding a sex ratio of 0.73 in favor of women. The mean age of patients was 50.2 ± 16.2 years, with a range from 22 to 77 years. Geographically, most patients (86.5%) resided in Conakry.

Regarding marital status, 84.6% of patients were married, followed by widowed individuals (9.6%) and single patients (5.7%). Concerning professional activity, most patients worked in the informal sector (57.7%), while 34.6% were employed in the formal sector, and 7.7% were unemployed.

Type 2 diabetes was the predominant form, accounting for 88.4% of cases, followed by type 1 diabetes (7.6%) and gestational diabetes (3.8%). In 32.6% of patients, diabetic ketoacidosis was the initial manifestation of diabetes. All patients presented with blood glucose levels exceeding 400 mg/dL on admission. Ketonemia was greater than 3.0 mmol/L in 42% of cases, and ketonuria of at least two crosses was observed in 98.1% of patients.

Regarding therapeutic education, half of the patients (50%) had never received training in diabetes management, and 70% lacked means for home blood glucose monitoring. Diabetes was diagnosed in the context of cardinal symptoms in 48% of cases, during routine health checks in 21.1%, and because of diabetic complications in 5.7%. Glycemic control was poor, with an average glycated hemoglobin (HbA1c) level of 13%.

Microangiopathic complications were dominated by diabetic peripheral neuropathy (56.7%), followed by diabetic retinopathy (15.3%) and diabetic nephropathy (10.6%). Macroangiopathic complications included obliterative arterial disease of the lower limbs (AOMI) in 12%, ischemic heart disease in 6.3%, diabetic foot in 9.6%, and stroke in 1.9%.

As for comorbidities, dyslipidemia was present in 71% of cases, arterial hypertension in 32%, and obesity in 6%.

Clinically, all patients presented with dehydration and Kussmaul dyspnea, a hallmark of metabolic acidosis. Cardinal symptoms of diabetes—polyuria, polydipsia, and weight loss—were observed in 57.6% of patients. Altered consciousness was noted in 8% of cases, including 5% with obnubilation and 3% in a comatose state.

Table 1 summarizes the main decompensation factors identified.

Treatment discontinuation was noted in 16 patients (30.7%). Among them, ten patients cited financial reasons for reducing or interrupting their treatment (insulin or oral antidiabetics), two patients cited unavailability of the drug in pharmacies, and four patients had abandoned their conventional treatment in favor of traditional medicine.

Infection was present in 37 cases (71.1%), mainly:

A dietary error was identified in 35 cases (67%), including eating only one meal a day, irregular mealtimes, and inappropriate snacking.

Use of poor-quality medication was noted in six patients, including two cases of expired medication and four cases of denatured insulin, linked to inadequate storage (lack of refrigeration).

Insulin therapy-related difficulties were observed in 16 patients (30.2%), including errors in injection technique, lack of rotation of injection sites, and poor dose adaptation.

In 16 cases (30.7%), multiple decompensation factors were associated with the same patient.

The average length of hospitalization was 7.5 ± 3.1 days. No deaths were recorded during the study period.

The hospital prevalence of diabetic ketoacidosis (DKA) in our study was 43%. This figure is consistent with those reported in similar African studies, such as Baldé et al. in Guinea (35%) [7] and Dionadji M. in N’Djamena (35.9%) [4] . However, significantly lower prevalence rates have been reported in other West African countries: 6.49% by Leye in Senegal [2] , 4.4% by Lokrou in Côte d’Ivoire [3] , and 8.24% by Ouédraogo in Burkina Faso [5] . These discrepancies may be attributed to differences in study inclusion criteria, case reporting methods, and variations in healthcare system organization across regions.

The mean age at the onset of DKA in our cohort aligns with findings from most African studies, where the average age is also around 50 years [1] - [3] . This likely reflects the predominance of type 2 diabetes, which is typically diagnosed in middle-aged or older adults.

Regarding sex distribution, our study showed a female predominance, resulting in a sex ratio of 0.73. This contrasts with findings from some African studies that report a male predominance. Although sex is not generally considered a major risk factor for DKA, such differences may stem from socio-cultural factors or better access to healthcare services for women in our study setting.

Most patients included in the study resided in Conakry, a trend also observed in other African studies where urban centers, particularly capital cities, account for most hospital admissions [8] . This can be explained by two key factors: the location of the study site in the capital and the limited access to specialized diabetes care in rural areas, where healthcare infrastructure for managing acute diabetic complications is often lacking.

The predominance of type 2 diabetes in our study aligns with findings from other African studies, where acute decompensation—including diabetic ketoacidosis (DKA)—is also frequently observed in patients with type 2 diabetes. This may be related to a distinctive phenotype of type 2 diabetes in African populations, characterized by a tendency toward ketosis and, in some cases, a reversible need for insulin therapy [9] [10] .

Although DKA is classically associated with type 1 diabetes, it can also occur in the context of gestational diabetes, albeit much more rarely. In our setting, this could be explained by the lack of routine screening and delayed management of pregnant women with gestational diabetes. DKA during pregnancy represents a serious medical emergency, as it compromises both maternal and fetal outcomes.

In 32.6% of patients, DKA was the initial presentation leading to the diagnosis of diabetes. This high proportion of inaugural DKA among patients with type 2 diabetes may be related to the occurrence of atypical diabetes in African populations, which often presents in adulthood with features of ketosis [9] [10] .

Furthermore, 44% of patients had been living with diabetes for more than five years. This contrasts with the findings of Sidibé A.T. et al. [11] , who reported that in 78.6% of cases, diabetes had been present for less than five years. This discrepancy could reflect differences in patient profiles, prior access to care, or patient awareness of their diagnosis prior to the decompensation episode.

Diabetic ketoacidosis (DKA) typically develops progressively over several hours or even days. Its clinical presentation is variable and may include altered levels of consciousness. In our study, 8% of patients exhibited impaired consciousness—5% with obnubilation and 3% in a comatose state.

All patients presented with dehydration and Kussmaul dyspnea, which are characteristic features of DKA. Cardinal symptoms of diabetes—polyuria, polydipsia, and weight loss—were observed in 57.6% of cases. These signs reflect the natural course of poorly controlled or untreated diabetes. Our findings are consistent with those of Sidibé A.T. et al. [11] , who also reported 100% dehydration. However, this contrasts with data from Misaoui in Tunisia, where dehydration was reported in only 26.7% of patients [12] . According to the literature, water loss due to vomiting, diarrhea, and polypnea during DKA can range from 1 to 5 liters over 24 hours [13] .

Infections were the leading decompensation factor in our series, found in 71.1% of cases, followed by dietary errors in 63.4%. This predominance of infections is in line with findings from several African studies [2] - [9] [14] [15] . Infections—whether bacterial, viral, parasitic, or fungal—increase insulin resistance, disrupt carbohydrate metabolism, and worsen dehydration, thereby precipitating DKA.

A lack of patient education and awareness remains a major contributor to acute diabetic complications. In our study, only 50% of patients had received structured therapeutic education. Among known diabetics, 30.7% had interrupted treatment, 13.4% were on an inappropriate regimen, and 26.9% of insulin-treated patients reported difficulties with administration. These findings may be attributed to several factors, including limited patient knowledge, lack of trained healthcare personnel, unavailability of insulin, and economic challenges in maintaining long-term treatment adherence.

In 30.7% of patients, multiple decompensation factors were identified. No clear precipitating factor was found in 5.7% of cases—a phenomenon also reported in other African series [7] [16] . This may be due to the lack of systematic diagnostic tests such as blood cultures or urinalysis (ECBU), and the widespread practice of self-medication, including empirical antibiotic use, which can obscure clinical signs of infection.

Importantly, no deaths were recorded during the study period, suggesting that management at our center was relatively effective despite resource limitations.

We conducted a three-month cross-sectional study to analyze the factors involved in the decompensation of diabetic ketoacidosis at the Fellah polyclinic, a specialized diabetes management center in Conakry, Guinea. A total of 52 cases of diabetic ketoacidosis were documented. However, our study has some methodological limitations. On the one hand, the concentration of cases in a specialized facility could lead to an overestimation of the prevalence of diabetic ketoacidosis, which makes it difficult to extrapolate the findings to the entire Guinean population. On the other hand, the absence of certain biological examinations (pH, ionogram, bicarbonates, bacteriology) constitutes a limitation, as the severity of acidosis and infectious etiology could not be determined in all cases.

Despite these limitations, this study provides current data on the factors contributing to the decompensation of diabetic ketoacidosis in Conakry. This information is essential for a better understanding of the determinants of this acute complication and could serve as a basis for developing prevention and management strategies tailored to the Guinean context.