Clinical profile and outcome of critically ill COVID-19 patients with diabetic ketoacidosis: A single-centre cross-sectional study in South India

INTRODUCTION India has among the world’s largest population of diabetic patients.[1] Literature suggests that diabetic ketoacidosis (DKA) is the most common hyperglycaemic crisis in COVID- 19 patients.[2] A retrospective matched cohort study in Canada, prior to the pandemic, reported the incidence of diabetic ketoacidosis (DKA) to be 4.59 per 1000 admissions in the Intensive Care Unit (ICU) (95% confidence Interval, 3.64–5.71).[3] A recent publication reported a prevalence of DKA among young Indians admitted to the hospital to be 6.6%.[4] With these factors in mind, this observational study aimed at assessing the incidence of DKA in severely ill COVID- 19 patients and on identifying the risk factors for developing DKA. MATERIALS AND METHODS We conducted a retrospective cohort study, in the COVID-19 ICU of a tertiary care centre in South India. The census method was used for sampling. The cohort included all patients admitted in ICU with laboratory confirmed COVID-19 infection and who developed DKA during the period of three months (15 May 2021 to 15 August 2021). Adult patients admitted in COVID-19 intensive care unit with a definite diagnosis of DKA, as per American Diabetes Association criteria, and all patients admitted in the COVID-19 ICU for more than 3 days with laboratory-confirmed COVID-19 infection were included. Patients with severe respiratory acidosis during the episode of DKA, patients who had DKA on admission, patients discharged against medical advice and patients with incomplete data were excluded. Ethical clearance was obtained from the institutional ethics committee. The data from the case records were collected using a data collection proforma. Patients were diagnosed as DKA if they had uncontrolled hyperglycaemia (>13.88 mmol/L), metabolic acidosis (pH <7.3, Sr HCO3 <18 mmol/L) and increased serum ketones. COVID-19 infection was confirmed by reverse transcription-polymerase chain reaction test. Study variables included demographic profile, clinical profile, laboratory parameters, treatment received, duration of stay, severity markers and survival status. One patient was excluded from the study due to incomplete data. Calculation of mean, standard deviation, frequency and percentages were carried out using SPSS software. RESULTS A total 20 patients, who had one or more episodes of DKA, were identified out of 156 patients in the period of 3 months [Figure 1]. The mean age of these patients was 54.6 years. The majority were males (M:F = 4:1). BMI was not assessed due to incomplete data.Figure 1: Consort diagramA total of 18 out of 20 patients were known diabetics. All of them had T2DM. The average glycosylated haemoglobin (HbA1c) on admission was 9.2. Preadmission control was primarily through oral hypoglycaemics; in two patients, this included sodium glucose co-transporter inhibitors (SGLT1), and one patient was on subcutaneous insulin. Two of the patients were not known diabetics. The coexisting comorbidities are shown in Table 1.Table 1: Patient demographics and clinical featuresAll patients received steroid treatment (dexamethasone dose of 6 mg once a day irrespective of weight), for an average duration of 12 days. Fifty per cent of the patients had severe DKA necessitating variable rate insulin infusions, and the remaining patients with mild/moderate DKA were managed with subcutaneous insulin. The number of discrete episodes of DKA per patient also varied. It was considered discrete if they were spaced 12 or more hours following biochemical resolution of the original DKA. One episode was documented in 8/20 and two episodes in another 8/20. Four patients had three episodes. One patient who was later excluded due to incomplete data had four episodes of DKA during the course of ICU stay. The biochemical parameters were as follows: Average highest sugar value reported during their DKA episode: 380.7 mg/dL, average blood ketone value was 12.2 mg/dL; average pH recorded during episodes were 7.26, lowest bicarbonate: 16.621 mmol/L, potassium: 5.03 meq/L. A total of 10 patients had D-dimer levels >500 mg/mL, and these data were missing in four patients. Sr ferretin levels were documented in 10 of these patients (average 814.5 ng/mL), CPK in 11 patients (average 9.74 mcg/L) and procalcitonin in 3 patients (average7.43 ng/mL) [Table 2].Table 2: Metabolic parametersAll the patients required oxygen support (100%), and 10/20 (50%) were on invasive ventilatory support. The average duration of stay in hospital was 12.4 days. The average duration of ICU stay was 8.5 days. There were no survivors in the sampled cohort. Most patients succumbed due to respiratory failure, four following acute kidney injuries and one each following acute myocardial infarction and hepatic encephalopathy. Therefore, the planned comparative statistical analysis between survivors and non-survivors was not possible. DISCUSSION The incidence of DKA in the adult COVID-19 ICU was 12.8%. Similar data from Western literature report an increased overall incidence of DKA admissions by 6–7% during the same period.[5] However, these data are not confined to ICU patients. There are no similar published Indian data as yet. A related study reported that 23.1% of DKA patients admitted to a tertiary care centre in North India during the same period were COVID-19 positive.[6] A prospective study on Indian population reported a significant incidence of new-onset DKA following parenteral glucocorticoids in Type 2 DM (T2DM) patients with COVID-19, especially in those with body mass index <25.56 kg/m2, HbA1c% >8.35% and IL-6 levels >50.95 pg/ml at admission.[7] In a series of 658 patients, Li et al.[8] reported that 42 patients (6.4%) presented with increased urine or serum ketones, of which only three had acidosis, meeting the criteria for DKA, making the incidence of DKA 2.3%. In contrast, a retrospective single centre study in the UK reported the incidence of DKA as 9.2% among diabetics with COVID 19. Another study in the UK reported a prevalence of 5.9%.[9,10] The higher incidence in our study could be accounted by the sampled population which included only ICU patients with severe disease. Given that diabetes is the commonest precipitating comorbidity, the very high incidence in Indian population could also explain the data.[11] The incidence Of DKA in the ICU is also higher than noted in the prepandemic era, 12.8% versus 0.4%.[3,4] This suggests the infection with the SARS-COV-2 virus has a contributory role. Predisposition to hyperglycaemic crises during COVID-19 infection has also been attributed to a greater release of counter regulatory hormones, namely glucocorticoids and catecholamines.[11] The use of intravenous steroids as the mainstay in the management of COVID -19 pneumonia further added insult to injury.[12] However, there are only isolated case reports attributing DKA to steroid therapy.[13] The higher incidence of DKA in patients with Type 2 DM is explained by the increased susceptibility in these patients under stressful conditions and the higher prevalence of Type 2 DM,[14] and 2/20 were previously normoglycaemic adults. COVID-19-provoked ketone prone diabetes (KPD) is increasingly reported in literature.[15] The most common comorbidities among these patients were diabetes, hypertension and chronic kidney disease, and seven patients had uncontrolled diabetes with HbA1C level more than 7%. These findings are similar to those in a meta-analysis published recently.[16] Two patients were on SGLT2 inhibitors. These drugs have been reported to increase risk of DKA in Type 1 diabetes.[17,18] A multivariate risk analysis reported that the only parameter associated with a significantly increased risk of hospitalization was a preinfection HbA1c≥9%.[19] In our study, the average value was >9.2%. The incidence of presenting symptoms is consistent with what is reported till date.[20] There are no reliable data regarding severity and number of episodes of DKA per patient during ICU stay. There are reports suggesting prolonged ketosis and duration of DKA for up to 35 hr in patients with COVID-19 infection.[21,22] The biochemical parameters noted were not consistent with what is reported in a systematic review and meta-analysis by Alhumaid et al.[16] The median random blood sugar levels we observed was 21.15 mmol/L which is lower than 28.14 mmol/L reported in the review. The difference can be explained by the difference in severity of infection and the population sampled. A recently published study concluded that patients with diabetes and COVID-19 are at a higher risk for severe outcomes and ICU admission, primarily due to DKA.[23] Mortality rates associated with DKA and COVID vary and are largely unknown. While Chamorro-Pareja et al.[24] report a mortality rate of 50% in patients with diabetes and COVID-19, Alkundi et al. suggest that the development of DKA in patients with diabetes and COVID-19 is associated with a greater survival compared to those who do not develop DKA.[9] A retrospective study in China reported the mortality rate among those with diabetes which was significantly higher than that among the non-diabetic group (16.5% vs. 0%).[25] The 100% mortality in our study can be attributed to the fact that the study population included only severely ill patients, lack of access, availability and overwhelming of ICU facilities at the height of the pandemic. We conclude that the incidence of DKA is higher in ICU patients infected with SARS-COV-2 virus. Male patients over 50 years of age, known diabetics, who received steroid treatment were more likely to develop DKA. The high incidence of both DM and COVID-19, delay, overwhelming and lack of access to healthcare in our country could account for this. Risk stratification, prevention of DKA and early detection are crucial to effectively allocate resources, manage the disease and improve outcomes. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.