The major findings of this study included the following. First, patients with HHS were older (median age: 75 years), and infection was the most common trigger, followed by steroid use, enteral nutrition, oral antipsychotic use, and total parenteral nutrition. In addition, approximately two-thirds of the patients were not taking diabetes medication at the onset of HHS. Second, of the 84 patients with HHS, 70 were classified into the isolated HHS group and 14 into the mixed DKA/HHS group. The 30-day mortality rate was significantly higher in the isolated HHS group than in the mixed DKA/HHS group. Third, among the isolated HHS group, those with metabolic acidosis other than ketoacidosis had a particularly poor prognosis.

In our study, 17.9% of patients were newly diagnosed with diabetes at the onset of HHS. The remaining 82.1% had known type 1, type 2, pancreatogenic or other diabetes. Kitabchi et al. reported that the proportion of patients with undiagnosed diabetes before HHS was 20% [15], which is consistent with the results of our study. Of all 84 patients with HHS, 69 (82.1%) were previously diagnosed with diabetes, but only 26 of the 69 patients were receiving diabetes medications, while the remaining 43 were not. Therefore, regular blood glucose monitoring and appropriate diabetes medications in outpatient clinics may be important to prevent the development of HHS. Infections have been widely reported as the primary precipitating factors of HHS, with pneumonia being the most common, followed by urinary tract infections [16]. Consistent with previous reports, our study identified infection as the leading cause of HHS, accounting for 51.2% of all HHS cases. Therefore, when managing patients with HHS, it may be essential to consider antibiotic therapy in parallel with insulin-based blood glucose control, considering the possibility of underlying infections.

In our study, the 30-day mortality rate for HHS was 21.5%, which is higher than the 10–20% reported in previous studies [17,18,19,20]. In a hospital-based retrospective cohort study in Taiwan [17], involving 978 patients with HHS hospitalized from 1991 to 2005, the 28-day mortality rate was 18.83%. The mean age of these patients was 67.9 years, 7 years younger than our sample. When stratified by age, the mortality rate was reported to be only 3.6% for those aged below 45 years, 5.8% for those aged 45–69 years, and 17.9% for those over 69 years. Thus, age had a strong influence on the HHS mortality rate. Another study, the Danish nationwide residence-based cohort study [18], comprising 634 adult patients with HHS from 2016 to 2018, reported a hospitalization mortality rate of 14.0%, the mean age of the patients was 69 years, 6 years younger than our patients. In addition, the comorbidity rate of dementia in that study was 5.5%, lower than the 22.6% observed in our study, while the comorbidity rate of cancer was 12.9%, lower than 34.5% observed in our study. Therefore, the higher mortality rate in our study may have been due to older age and higher rates of comorbidities. In a recent multicenter retrospective cohort study of 21 acute care hospitals in Japan on hyperglycemic crises [21], 226 adult hospitalized patients with HHS from 2012 to 2016 were analyzed. The reported in-hospital mortality was 7.1%. The mean age of the patients was 71.9 years, 3 years younger than our sample. Infectious diseases were identified as a precipitating factor for HHS in 42.7% of the patients, lower than the 51.2% observed in our study. The steroid use rate was 6.2%, which is also lower than the 21.4% observed in our study. Furthermore, we compared the frequency of ischemic heart disease and stroke as precipitating factors for the development of HHS between our study and this Japanese multicenter retrospective cohort study. Notably, ischemic heart disease was identified as a precipitating factor for HHS in 1.3% of patients in that study, which is lower than the 3.6% observed in our cohort. The prevalence of stroke was 1.8%, which is also lower than the 6.0% observed in our cohort. Our hospital is a regional university hospital, where patients with serious emergencies are transported by the local emergency system. The differences in patient backgrounds may have contributed to the higher mortality rate in our study.

DKA is characterized by the triad of hyperglycemia, elevated blood and/or urine ketone body levels, and metabolic acidosis. The pathogenesis of DKA results from a combination of absolute or relative insulin deficiency and elevated levels of counteracting hormones (such as glucagon, epinephrine, norepinephrine, cortisol, and growth hormone). In contrast, HHS has residual insulin secretion to a degree that minimizes ketosis but fails to control hyperglycemia [9]. However, in clinical practice, one-third of patients with hyperglycemic crises show characteristics of both HHS and DKA [9, 10]. Therefore, in our study, HHS was separated into HHS without DKA and HHS with DKA, with the former being defined as “isolated HHS” and the latter as “mixed DKA/HHA”. Notably, the 30-day mortality rate in the isolated HHS group was higher than that in the mixed DKA/HHS group, indicating a significantly worse prognosis in the isolated HHS group. A report of the clinical profile of 144 patients admitted with hyperglycemic emergencies from 2012 to 2019 at a tertiary medical center in Kumamoto, Japan, showed that 87 were classified as having DKA alone, 38 as HHS alone, and 19 as mixed DKA/HHS. The in-hospital mortality rates were 5.7%, 13.2%, and 5.3%, respectively, with the HHS alone group having the poorest prognosis. The HHS alone group reported significantly higher serum Na+ levels and effective osmolarity than the mixed DKA/HHS group; hence, they may have been more dehydrated [20]. However, our study did not find differences in serum Na + levels and effective osmolarity between the isolated HHS and mixed DKA/HHS groups, although both tended to be higher in the isolated HHS than the mixed DKA/HHS group. The reason for the lack of statistically significant differences may be due to the small sample size. In the Danish national residence-based cohort study, 634 adult patients with HHS were divided into the isolated HHS group (n = 394) and mixed DKA/HHS group (n = 240) [18]. The in-hospital mortality rates were 17.3% and 8.8%, respectively, with a significantly worse prognosis in the isolated HHS group than in the mixed DKA/HHS group. Of note, the prevalence of type 1 diabetes in the Danish study was higher in the mixed DKA/HHS group (38.8%) than in the isolated HHS group (5.8%), a trend that was similarly observed in our study, with 42.9% in the mixed DKA/HHS group and 7.1% in the isolated HHS group. We hypothesized that the HHS of type 1 diabetes was more likely to be “mixed HHS + DKA” than “isolated HHS” and that the patients would be able to visit a hospital before their general condition became severe, since type 1 diabetes quickly progresses to DKA after insulin deficiency, whereas it takes several days for type 2 diabetes to progress to HHS. In fact, as shown in Table 1, the level of consciousness (JCS) at admission was significantly milder in the mixed DKA + HHS group, which had a higher proportion of type 1 diabetes, than in the isolated HHS group. Finally, although the difference was not statistically significant, steroid use was less frequent in the mixed DKA/HHS group than in the isolated HHS group. Additionally, the absence of enteral and total parental nutrition in the mixed DKA/HHS group may have contributed to the lower mortality rate observed in our study. As for the living environment prior to admission, the proportion of patients who lived with family was significantly higher in the mixed DKA/HHS group than in the isolated HHS group, but the reason for this is not clear. However, considering that 6 of the 14 patients with the mixed DKA/HHS had type 1 diabetes and were of older age, the need for continuous support in elderly patients with ketone-prone type 1 diabetes may have been a contributing factor to living with the family prior to admission.

Among the 84 patients with HHS in this study, only 34 individuals (40.5%) were discharged to their homes, indicating a substantial proportion of patients required alternative care arrangements post-discharge. In our study, approximately half (51.2%) of the patients with HHS had serious infections such as pneumonia and urinary tract infections that required long-term antibiotic therapy, resulting in an extended average hospital stay of 23 days. In elderly patients, the onset of HHS triggered a subsequent decline in activities of daily living and cognitive function, leading to increased disuse syndrome. These factors may have contributed to the difficulty of discharging patients to their homes in our study. Thus, even when death was averted, the functional prognosis was considered poor. Therefore, early recognition and prompt treatment are essential for improving outcomes in patients with HHS.

Finally, we found that the prognosis of HHS was particularly poor when it was accompanied by metabolic acidosis other than ketoacidosis. HHS is a serious condition that often coexists with acute illnesses, and metabolic acidosis can occur owing to disease progression or acute renal injury. In our study, 11 of 70 isolated HHS cases presented with metabolic acidosis other than ketoacidosis. Among the 11 patients, 6 died within 3 days of admission, and the causes of death were ischemic heart disease, pulmonary embolism, cerebral infarction, and NOMI, all of which were serious acute diseases. Hyperlactatemia reflects metabolic disturbances such as tissue hypoxia, anaerobic metabolism, and impaired lactate clearance. Hyperlactatemia has been reported to correlate with illness severity and is a predictor of mortality in trauma and non-trauma patients [22]. In our study, the lactate level in the isolated HHS with acidosis group was markedly higher at 14.6 mmol/L; hence, the higher mortality in this group was not surprising. Additionally, among the 11 patients with isolated HHS and metabolic acidosis, two were taking metformin. One patient died of ischemic heart disease on the first day of hospitalization. The other patient was in septic shock due to a urinary tract infection and was transferred to post-acute care hospital after 51 days. Then, the cause of their metabolic acidosis was thought to be a severe acute illness rather than metformin-associated lactic acidosis. Consequently, based on these results, the prognosis of patients with HHS with metabolic acidosis other than ketoacidosis is extremely poor. As for the living environment prior to admission, the proportion of patients who lived with family was significantly higher in the group with acidosis than in the group without acidosis, but the reason for this is not clear. Among the 11 patients in the isolated HHS with acidosis group, three experienced ischemic heart disease and one each experienced pulmonary embolism, cerebral infarction, and NOMI. All of these patients died within 3 days of hospitalization. In cases of severe and rapidly progressive acute arterial embolism, timely access to medical care would likely be impossible for patients living alone, and this may have contributed to the higher proportion of patients living with family observed in the group with acidosis.

This study had a few limitations. First, this was a single-center study with a small sample size. Second, this was a retrospective study with no long-term outcomes. Hence, future multicenter follow-up studies are needed to validate our findings in larger cohorts of patients.

In conclusion, our study demonstrated the characteristics and prognoses of patients with HHS at a single university hospital in Japan. HHS commonly occurred in elderly patients, many of whom were not receiving appropriate diabetes medications. Infection was the most common trigger for the development of HHS. Furthermore, isolated HHS had a worse prognosis than mixed DKA/HHS; moreover, the prognosis was worse when accompanied by metabolic acidosis other than ketoacidosis.