This study was a retrospective single-centre case series at an urban district hospital in South Africa. All patients admitted with a diagnosis of DKA between July 2020 and July 2021 were identified from the hospital admissions register. During the study period, all admissions had nasopharyngeal and oropharyngeal swabs sent for identification of the novel coronavirus responsible for COVID-19 disease. Patients with a positive result on the real-time reverse transcription polymerase chain reaction (RT-PCR) assay were then included in the study.

The setting was an urban district hospital in South Africa. This busy facility serves a population of approximately 980 000 people and provides general medical, surgical and obstetric services. Approximately 100 patients with DKA are admitted annually. The high incidence of DKA has been attributed to the high prevalence of DM in South Africa.¹⁸ The hospital has a high care ward but not an intensive care unit (ICU). Patients admitted with DKA are usually managed by the Family Medicine Department in the high care ward.

This case series describes all adult patients (> 18 years old) admitted to the facility with DKA who were also infected with the novel coronavirus. The 12-month study period included the tail end of the first COVID-19 wave, the second wave and the early part of the third COVID-19 wave as experienced in South Africa.¹⁹ The criteria for diagnosing DKA were as follows:

Evidence of coronavirus infection was confirmed by laboratory result (RT-PCR). This test confirms presence of the viral genetic material in the body.²⁰ Patients with DKA who were managed as outpatients were excluded from this study. All eligible patients were included in this study.

A data collection tool was used to extract data from selected patient files. Medical records were reviewed, and data were extracted regarding patients’ demographic details, clinical progress and laboratory results.

Descriptive analysis was conducted on the data. Descriptive statistics were calculated for all variables, with mean ± standard deviation (s.d.) or median and interquartile range reported for continuous variables, and number and percentage of total were reported for categorical variables.

Consent was obtained from the hospital management and the University of KwaZulu-Natal Biomedical Research Ethics Committee (ref. no. BREC/00000907/2019).

A total of 106 patients were admitted with DKA at the district hospital during the study period (01 July 2020 – 31 July 2021). Ten of the identified patients with DKA also tested positive for the novel coronavirus (9.4%). Table 1 displays the clinical and demographic characteristics of the patient cohort.

As indicated in Table 1, six of the patients (60%) were under 45 years old. There were six male patients and four female patients. Nine of the patients were documented as African and one Caucasian.

Half of the patients (n = 5) were reported to have pre-existing type 2 diabetes mellitus (T2DM), for which they were on treatment. One of the known T2DM patients was on biphasic insulin only, and the other four were on metformin. None of the known DM patients were on two or more antidiabetic agents. Two of the patients with DM had pre-existing hypertension, and one DM patient was also human immunodeficiency virus (HIV)-positive. All of the patients with comorbid conditions were on treatment prior to admission.

The five patients with no previous history of DM reported no symptoms or signs suggestive of diabetes mellitus prior to their admission, nor did they have any other chronic conditions. Only one of the previously undiagnosed diabetic patients was noted to be obese, and two in the known diabetic group were obese. No other anthropometric measurements were recorded.

The biochemical markers and clinical outcomes for these patients are tabulated in Table 2. The investigations recorded include the HbA1c, the arterial blood gas results, CRP (C-reactive protein), urinalysis and the capillary blood glucose (glucometer scale 2 mmol/L – 30 mmol/L) on admission.

All of the patients were determined to have DKA as evidenced by the presence of hyperglycaemia, ketonuria and acidosis on arterial blood gas specimens. One of the patients had no documented bicarbonate levels on admission but was diagnosed with DKA on clinical features. The average pH on admission was 7.02 (range from 6.87 to 7.29). The patients with the two lowest pH values on presentation (pH of 6.88 and 7.00) both demised. The third patient who demised had an initial pH of 7.23.

The clinical presentation of patients: six of the patients presented with symptoms suggestive of COVID-19 infection (i.e. sore through, cough and shortness of breath). Hypoxia, as denoted by oxygen saturation of less than 90% on room air, was only present in two of the patients, both of whom demised.

Despite seven patients requiring step-up care in an ICU, as indicated by national guidelines, only one was successfully transferred to ICU, where he recovered in seven days and was discharged.²¹ Although it is a recommendation for patients with severe DKA to be managed in an ICU setting, only three of the non-COVID-19 DKA patients were transferred to ICU during the study period.²² Under normal circumstances, patients with DKA are managed in the high care ward at the district hospital. Only 1 of the 10 patients in this case series was managed in the high care ward; the rest were managed in the COVID-19 isolation wards. Three of the patients recovered in the district hospital ward and were discharged. Three of the patients were transferred to the field hospital from the district hospital, where management was continued. Three patients demised at the district hospital.

Two of the three patients who demised had severe DKA; the third patient who demised had features of combined DKA/hyperosmolar hyperglycaemic state (HHS).

The average length of stay in hospital was 3.3 days, ranging from 1 to 10 days. The short duration of hospital stay was highly influenced by the COVID-19 pandemic. The district hospital provided acute care and thereafter patients with COVID-19 were stabilised and transferred to one of the COVID-19 field hospitals.

Of note, all of the patients (both with known and previously undiagnosed DM) were noted to have abnormal HbA1c values on admission ranging from 9.7 to 13.8. Also, the known diabetic group had an average HbA1c of 11% and unknown diabetics HbA1c 12.5%, yet the previously undiagnosed diabetic patients had better outcomes (20% mortality) compared to the group of known diabetic patients (40% mortality). This could possibly be because of the relatively short duration of hyperglycaemia in the previously undiagnosed patients, and possible presence of comorbidities and diabetic complications in the known diabetic patients may have adversely affected their prognosis.

The sample size in this case series is small, and therefore the characteristics of this cohort of patients is not statistically significant. Furthermore, the patients selected were a convenience sample from one facility and may not be representative of all patients with DKA and COVID-19. The period of sampling occurred during the height of the COVID-19 pandemic and therefore other factors such as staff shortages and changes in facility resources may have influenced the inpatient management and clinical outcomes of the patients.