If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Euglycemic diabetic ketoacidosis is a metabolic condition characterized by relative euglycemia, ketonemia, and metabolic acidosis that occurs through mechanisms resembling starvation. Pancreaticoduodenectomy is a complex abdominal operation that subjects patients to a prolonged fasting and an inflammatory state. This study examined the incidence of euglycemic diabetic ketoacidosis and potential opportunities for early diagnosis and management in patients undergoing pancreaticoduodenectomy.
Methods
A single-institution retrospective review of 350 patients who underwent pancreaticoduodenectomy between 2017 and 2020 was performed. Primary endpoints were peak beta-hydroxybutyrate levels, peak lactate levels, lowest pH, peak base deficits, and urinary output within the first 24 hours, postoperatively. Additional endpoints included incidence of postoperative pancreatic fistula, delayed gastric emptying, total complications, postoperative hospital length of stay, readmission rates, and changes in insulin regimen at discharge.
Results
Of the 350 cases reviewed, 39 (11.1%) patients developed euglycemic diabetic ketoacidosis. Male sex and pancreatic cancer were associated with a risk for euglycemic diabetic ketoacidosis (P < .05). Patients with euglycemic diabetic ketoacidosis had significantly higher peak beta-hydroxybutyrate levels than patients without euglycemic diabetic ketoacidosis (mean difference = 19.8 mg/dL, 95% confidence interval = 14.7-24.9, P < .001), and were nearly four times more likely to require insulin at discharge (odds ratio 3.8, 95% confidence interval = 1.1–13.0, P < .05).
Conclusion
This is the first large descriptive study that investigates euglycemic diabetic ketoacidosis after pancreaticoduodenectomy. Euglycemic diabetic ketoacidosis after pancreaticoduodenectomy is associated with significantly higher beta-hydroxybutyrate levels and new or increased insulin requirement at discharge. Our study demonstrates potential markers for euglycemic diabetic ketoacidosis after pancreaticoduodenectomy, offering an opportunity to identify and successfully treat this disease in a timely manner.
Introduction
The pancreas is important for the endocrine regulation of key metabolic and physiologic processes in the body, most notably glucose homeostasis. Pancreatic surgery can result in acute derangements to this regulatory system. Euglycemic diabetic ketoacidosis (EDKA) is a metabolic condition that can occur in patients with or without diabetes.
It is traditionally characterized by relative euglycemia (blood glucose <250 mg/dL), ketoacidosis, and metabolic acidosis (pH <7.3 and serum bicarbonate <18 mEq/L).
It can arise in various settings including sepsis, pregnancy, pancreatitis, low caloric intake, treatment with sodium-glucose co-transporter 2 (SGLT-2) inhibitors, and after surgery.
Importantly, the lack of marked hyperglycemia in EDKA, in addition to varying clinical presentations, frequently leads to delays in diagnosis and treatment.
Pancreaticoduodenectomy (PD) is a complex abdominal operation that subjects patients to a prolonged fasting and inflammatory state, increasing the risk of EDKA.
Few cases of EDKA after PD (PEDKA) have been reported in the literature, but the condition is still poorly understood and under-recognized in this patient population.
Outside of limited case reports, there is no original research published in the literature evaluating EDKA after PD. This study aims to characterize the incidence of EDKA, its physiologic impact on postoperative recovery and patient outcomes, and potential areas for prevention and management in patients undergoing PD.
Methods
Study population
This study was a single-institution retrospective review of 350 consecutive patients undergoing PD at Thomas Jefferson University Hospital between 2017 and 2020. Permission for this analysis was obtained through the hospital Institutional Review Board (IRB# 21E.051). This study was performed using the data from the Thomas Jefferson University Hospital PD database and patient electronic medical records. Starting on postoperative day (POD) 0, patients with laboratory tests suggestive of acidosis were further assessed by measurement of arterial blood gases and plasma lactate. Cases presenting with acidosis (pH <7.32), increased base deficit (>3–4 mEq/L), and normal lactate levels in the presence of normal, or mildly elevated, blood glucose levels (<250 mg/dL) were sent for measurement of serum beta-hydroxybutyrate (BHB), serum osmolality, and urinalysis (for presence of glucose and ketones). Other previously described clinical signs such as increased urinary output (>150 mL/hour) were also assessed. Patients with BHB >10 mg/dL presenting in the aforementioned fashion were identified to have PEDKA and subsequently treated with continuous intravenous (IV) dextrose and insulin until resolution of their acidosis and ketonemia.
Data collection
Demographic information was collected for all patients in this study. Past medical history, use of hypoglycemic agents and insulin at home, history of SGLT-2 inhibitor use, and oncologic data (tumor site and histology) were also collected. The primary endpoints of this study were peak BHB levels, perioperative urinary output, peak lactate levels, lowest pH, and peak base deficits within the first 24 hours postoperatively. Additional endpoints collected included incidence of postoperative pancreatic fistula (POPF), delayed gastric emptying (DGE), total complications, postoperative hospital length of stay (LOS), readmission rates, and changes in insulin regimen at discharge. Classifications for DGE and POPF according to the 2016 update of the International Study Group for Pancreatic Surgery have previously been described in the literature.
The International Study Group of Pancreatic Surgery definition of delayed gastric emptying and the effects of various surgical modifications on the occurrence of delayed gastric emptying after pancreatoduodenectomy.
Documentation of the precise date of patients’ return to solid diet was not available in the electronic medical records due to maintenance on the institution-specific enhanced recovery pathway in the immediate postoperative period, thus a modified DGE scale based on time to unrestricted full liquid diet was employed.
Enhancing patient outcomes while containing costs after complex abdominal operation: a randomized controlled trial of the Whipple accelerated recovery pathway.
Statistical analysis was performed using SPSS, version 20 (IBM Corp, Armonk, NY) software and R-Studio, version 1.2.5042 (R-Studio Inc., Boston, MA) with the ‘ggplot2’ graphics package. For continuous variables, Mann-Whitney tests and t tests were used for comparisons based on the normality of distribution. Categorical data were compared using Fisher exact test and the χ2 analysis to evaluate differences between qualitative variables. If identified as predictive, the variable was evaluated by odds ratio (OR), 95% CIs, and significance. Receiver operator curve (ROC) analysis was used to assess the robustness of associations between variables.
Results
PEDKA is associated with pancreatic ductal adenocarcinoma and male sex
In our cohort of 350 patients undergoing PD, we identified 70 (20%) patients who demonstrated clinical signs of PEDKA and were evaluated via serum BHB testing. Of the 70 patients monitored in this fashion, 39 (56%) patients were identified to have PEDKA (Table I). The overall incidence of EDKA in our population of patients undergoing PD was 11.1%. Twenty-nine (74.4%) patients had peak BHB levels on POD 0, 5 (12.8%) on POD 1, 2 (5.13%) on POD 2, 1 (2.56%) on POD 4, 1 (2.56%) on POD 8, and 1 (2.56%) on POD 10. In patients who required serum BHB monitoring, the rate of PEDKA in females was 42.9% compared with 68.6% in males. There was an association between patient sex and the rate of PEDKA after BHB testing, with male patients being nearly 3 times more likely to have PEDKA (OR 2.909, 95% CI 1.093–7.739, P < .05). Analysis of the surgical specimen pathology showed pancreatic ductal adenocarcinoma (PDAC) to be associated with PEDKA compared with non-PDAC pancreatic pathologies (OR 2.715, 95% CI 1.008–7.316, P < .05), as seen in Figure 1. Of note, there were no significant associations between PEDKA and BMI, previous history of DM, Hgb-A1c, or SGLT-2 inhibitor use.
Table IPatient demographics of the 70 patients for whom serum BHB values were collected. Note that there is a greater proportion of male patients and PDAC pathology associated with PEDKA.
Severity of PEDKA ketonemia levels correlates with serologic and physiologic critical measurements
Overall, BHB levels were 19.8 mg/dL higher in patients with PEDKA than in patients without PEDKA (approximately 2-fold increase compared with the PEDKA diagnostic threshold of 10 mg/dL). Patients who had a diagnosis of PDAC had significantly higher peak BHB levels than other pathologies (Mean difference = 10.7, 95% CI = 4.24–17.1, P < .001). Peak BHB positively correlated with base deficit (r = 0.32, P = .011), and negatively correlated with peak lactate levels (r = –0.31, P = .013) and total urinary output (r = –0.28, P = .026) as seen in Figure 2. There were no associations between peak serum BHB levels and previous use of SGLT-2 inhibitors or preoperative BMI.
Figure 1Peak peak-beta hydroxybutyrate on postoperative day 0 is increased in patients with pancreatic ductal adenocarcinoma compared with patients with all other pathologies (P < .001). PDAC, pancreatic ductal adenocarcinoma.
Figure 2Associations between peak-beta hydroxybutyrate and physiologic markers in the 24-hour perioperative period. Graphs demonstrating, from top to bottom, the association of peak peak-beta hydroxybutyrate with lowest pH, maximum base deficit, maximum lactate, and total urinary output volume. Note decreased association with maximum lactate, and decreased association with urinary output.
PEDKA is associated with increased requirements for insulin upon discharge
The median postoperative hospital LOS for our patient population was 5 (95% CI 5–6) days. Complications occurred in 55.7% of our patient population. More than 11% of our patients were considered to experience severe complications (Accordion Scale ≥3). Clinically relevant POPF and DGE were diagnosed in 14.3% and 8.6% of patients, respectively. The LOS, ICU LOS, total complications, severe complications, and 30-day readmission rates did not differ significantly between patients with PEDKA and those without PEDKA. No association was found between PEDKA and DGE or POPF (Table II).
Table IIHospital course/complications of the 70 patients for whom serum BHB values were collected. Note that there is no significant difference between PEDKA and clinically relevant postoperative pancreatic fistula or delayed gastric emptying. Note the significant association between PEDKA and insulin requirement at discharge (P = .029).
PEDKA (n = 39)
Non-PEDKA (n = 31)
P value
Median LOS (d)
5.00 (95% CI = 5.00–6.00)
5.00 (95% CI = 5.00–6.00)
.650
Mean total complications
1.31 ± 1.89
1.23 ± 1.54
.965
Mean total severe complications (Accordion >3)
0.21 ± 0.66
0.23± 0.67
.931
Mean total ICU stay (d)
1.67 ± 1.36
1.48 ± 0.77
.970
Clinically relevant POPF
6 (15.4%)
4 (12.9%)
.768
Clinically relevant DGE
3 (7.7%)
3 (9.7%)
.768
Wound infection
2 (1.4%)
0 (0.0%)
.499
Insulin requirement at discharge (yes/no)
14 (35.9%)
4 (12.9%)
.029
Mean insulin dose at discharge (units/d)
4.26 ± 7.5
1.6 ± 5.65
.042
Readmissions 30 d
7 (17.9%)
7 (22.5%)
.630
Readmission owing to POPF
2 (5.1%)
3 (9.7%)
.649
BHB, peak-beta hydroxybutyrate; ICU, intensive care unit; LOS, length of stay; PEDKA, euglycemic diabetic ketoacidosis after pancreaticoduodenectomy; PGE, prostaglandin E; POPF, postoperative pancreatic fistula.
Patients with PEDKA were nearly four times more likely to require insulin at discharge (OR 3.780, 95% CI 1.097–13.026, P < .05) and on average required higher insulin doses than those without PEDKA (Table II). New use of insulin or increased insulin dosage was required at discharge in 17.9% of patients (7 out of 39) who had PEDKA and none of the patients (0 out of 31) who did not have PEDKA (P = .015). The ROC analysis revealed BHB levels >17 mg/dL as the optimal threshold for predicting new or increased requirement for insulin at discharge (AUC = 0.86, P = .005). There was no significant association between preoperative insulin use at home and incidence of PEDKA.
Discussion
Euglycemic diabetic ketoacidosis is a frequently missed and under-recognized endocrine emergency. It is associated with prolonged starvation and major surgery and can impact postoperative clinical management. Pancreatic surgery presents an even greater clinical challenge as the fasting and physiologic stress that occurs during this major operation is combined with the endocrine derangement of the glycemic homeostasis system regulated by the pancreas. Clinical symptoms—including abdominal pain, nausea, vomiting, and dehydration—are similar to diabetic ketoacidosis (DKA) and nonspecific in the setting of any major abdominal surgery. Current literature surrounding PEDKA has only been limited to individual case reports and, as far as we are aware, our study is the first to investigate PEDKA on a larger scale.
This study has found that PEDKA was relatively common and was identified in 11.1% (39/350) of patients undergoing PD between 2017 and 2020. Such prevalence of PEDKA highlights the importance of awareness of the necessary preoperative and postoperative signs, including clinical findings and laboratory results, to ensure timely recognition and treatment of PEDKA during the postoperative period.
The pathophysiology of EDKA starts mostly in a fasting or caloric-restricted state. In this state, carbohydrate deficits result in low levels of serum insulin. Meanwhile, counter-regulatory hormones such as epinephrine and norepinephrine result in increased lipolysis and increased fatty acid production. Fatty acids are broken down to acetyl-CoA that is then further broken down into ketoacids including BHB and acetoacetate, which accumulate in the blood stream causing ketoacidosis. Decreased effectiveness of gluconeogenesis and depleted hepatic glycogen stores during the fasting state only contribute to the euglycemia and worsening ketoacidosis.
Treatment focuses on correcting dehydration using IV fluids and correcting ketoacidosis using IV dextrose and insulin, similar to the treatment of DKA.
Once a diagnosis was established, all of our patients were treated successfully with a combination of IV dextrose and insulin. After our review of the findings, we have established and implemented a protocol to assist in the identification and timely management of PEDKA (Figure 3).
In this study, we have identified preoperative risk factors related to PEDKA. Male patients and patients with a diagnosis of PDAC were shown in our study to have a significantly higher risk of developing PEDKA. Although there is no literature regarding an association between PEDKA and sex, there are studies and reports regarding EDKA in other medical settings. A 2015 case series reporting 9 cases of EDKA in patients taking SGLT2 inhibitors, showed that 8 out of 9 patients were women.
Plewa et al additionally reported a worse prognosis in pregnant women compared with the general population, with increased maternal and fetal mortality.
Unfortunately, we were unable to find reports explaining our observations of higher rates in males.
Euglycemic diabetic ketoacidosis has been previously described in patients with lung and colon malignancies, but aside from the previous case series from our institution, we are not aware of any literature discussing EDKA in PDAC.
Nonetheless, these reports suggest that a cancerous state may predispose a patient toward EDKA. A few mechanisms may be at play in promoting EDKA in cancer patients: cancers can be associated with cachexia and high physiologic stress; PDAC, especially, is commonly linked with new-onset diabetes and worsening of existent diabetes; and many cancers are associated with pro-inflammatory states that can affect endocrine homeostasis.
We have found a statistically significant positive correlation between postoperative peak BHB ketonemia levels in patients with PEDKA and increasing base deficit; an inverse correlation was also shown with urinary output and peak lactate levels. Additionally, we identified that PEDKA is associated with new or increased insulin requirement at discharge. ROC analysis of our data suggested that this need for new or increased insulin at discharge can be predicted by setting a threshold for BHB levels >17 mg/dL.
Current literature about EDKA in the postoperative setting of pancreatic surgery is extremely limited to a small number of case reports. We identified 3 patients across two case reports who had undergone pancreatic surgery with reported EDKA discovered by a high anion-gap metabolic acidosis with a normal lactate level and serum glucose level, elevated serum BHB, and elevated urine ketones in the setting of polyuria.
All 3 patients had type 2 diabetes before surgery and were managed with an SGLT2 inhibitor, which was discontinued 1 day before surgery. The mechanism of EDKA in all 3 patients was attributed to the use of SGLT2 inhibitors in the setting of physiologic stress secondary to their operations. Additionally, in the literature there are several cases that discuss EDKA in the setting of bariatric and cardiac surgery. Moreover, the mechanism of EDKA in these case reports was thought to be owing to SGTL2 inhibitor use in the setting of physiologic stress attributed to surgery, dehydration, and decreased carbohydrate intake perioperatively.
Empagliflozin induced euglycemic diabetic ketoacidosis in a patient undergoing coronary artery bypass graft despite discontinuation of the drug 48 hours prior to the surgery.
Interestingly, in our study, only 10% of patients that developed PEDKA were taking SGLT2 inhibitors at home and nearly all cases had these medications held for 1 to 5 days preoperatively. These findings suggest that SGLT2 inhibitors may be a key component of the mechanism behind EDKA, particularly if the SGLT2 inhibitor is continued up until the day of surgery.
Our finding of significantly decreased urinary output in the early postoperative period in patients experiencing PEDKA is intriguing as it seems paradoxical. In patients with a serum glucose of ≥180 mg/dL, polyuria would likely be expected because the glucose excretion rate is greater than the reabsorption in the proximal tubules of the kidney.
in: Walker H.K. Hall W.D. Hurst J.W. Clinical Methods: The History, Physical, and Laboratory Examinations. 3rd edition. Butterworth-Heinemann,
Boston (MA)1990
Due to the mild hyperglycemia in PEDKA, we expected that patients would have normal or slightly increased urinary output resulting from increased urinary osmotic forces. It is unclear whether PEDKA drives decreased urinary output, or a separate process such as hypovolemia drives both decreased urinary output and PEDKA. In the context of relative euglycemia, one proposed mechanism could be that glucosuria-driven osmotic diuresis is challenged by serum hyperosmolality from a high concentration of ketone bodies.
The serum hyperosmolality subsequently could promote the release of ADH, which stimulates the reabsorption of water in the distal convoluted tubule and collecting duct, resulting in a net decrease in urinary output. Additionally, the total urinary output may be decreased given that filtered loads of acetoacetate and BHB are positively correlated with their renal absorption.
Alternatively, in patients with extremely high serum ketone levels, the renal ketone reabsorption may encounter similar limitations to glucose reabsorption, possibly resulting in osmotic diuresis in these cases. The combined glucosuria and ketonuria temporarily result in an increased urinary output. However, over time as hypovolemia ensues through this osmotic diuresis, there would be a decrease in urinary output. The lack of literature surrounding the mechanisms behind this phenomenon highlights the need for further studies to verify the existence of hyperosmolality in patients with PEDKA and its correlation with urinary output.
There are several limitations to this study. As this was a retrospective study in which all identified patients were also promptly treated, we are unable to discuss the clinical course of untreated PEDKA. Furthermore, because BHB was measured only in patients presenting with signs of acidosis, it is possible that some cases were missed that could have been identified if all patients were screened for BHB levels. The immediate treatment of PEDKA also renders our study unable to examine the relationship between PEDKA and DKA, and if one devolves into the other. Certainly, the incidence of DKA postoperatively is low, with only one large-scale study identified that reported 2 cases of DKA in 141 patients undergoing total pancreatectomy.
Further studies are required to compare the incidence of PEDKA and postoperative DKA, their prognoses, and if one can serve as a risk factor for the other. Lastly, the study only had data on the insulin requirements of the study subjects in the immediate postoperative period; therefore, it is unclear how PEDKA affects insulin requirements over an extended period after surgery. Follow-up of insulin requirements over time is important in understanding the long-term consequences of PEDKA and should be included in future studies.
In conclusion, EDKA is a serious endocrine derangement that usually presents in the early postoperative period after PD. This is the first large-scale descriptive study to investigate PEDKA. Our study demonstrates that PEDKA manifests as ketoacidosis with elevated BHB, increased base deficit, and relative euglycemia. PEDKA also correlates with decreased urinary output and increased short-term and long-term insulin requirements. Timely identification of this clinical entity is crucial as appropriate treatment may reduce the risk of postoperative complications and improve patient outcomes.
Funding/Support
This research did not receive any specific funding from any agencies in the public, commercial, or not-for-profit areas.
Conflict of interest/Disclosure
The authors have no conflicts of interests or disclosures to report.
The International Study Group of Pancreatic Surgery definition of delayed gastric emptying and the effects of various surgical modifications on the occurrence of delayed gastric emptying after pancreatoduodenectomy.
Enhancing patient outcomes while containing costs after complex abdominal operation: a randomized controlled trial of the Whipple accelerated recovery pathway.
Empagliflozin induced euglycemic diabetic ketoacidosis in a patient undergoing coronary artery bypass graft despite discontinuation of the drug 48 hours prior to the surgery.
in: Walker H.K. Hall W.D. Hurst J.W. Clinical Methods: The History, Physical, and Laboratory Examinations. 3rd edition. Butterworth-Heinemann,
Boston (MA)1990
Euglycemic diabetic ketoacidosis (EDKA), which was first described by Munro et al1 in 1973, is a clinical syndrome characterized by euglycemia (blood glucose <250 mg/dL) in the presence of severe metabolic acidosis (arterial pH <7.3; serum bicarbonate <18 mEq/L) and ketonemia. In EDKA, delayed diagnosis and treatment (due to the absence of hyperglycemia) might induce serious and severe and life-threatening complications. In particular, EDKA after pancreaticoduodenectomy (PD) is still poorly understood and under-recognized, primarily because no large studies have specifically evaluated EDKA after PD; current literature is limited to case reports.
00536-0&id=gr1.jpg){kind=link}
00536-0&id=gr2.jpg){kind=link}
00536-0&id=gr3.jpg){kind=link}