Chloride Load Complicated Severe Diabetic Ketoacidosis in a Case Report

A 32-year-old woman with poorly controlled type 2 diabetes developed severe diabetic ketoacidosis that appeared to worsen even as her blood glucose and ketone concentrations fell. The case report identified two simultaneous acid-base processes: the high-anion-gap ketoacidosis that initially dominated and a treatment-associated normal-anion-gap hyperchloremic acidosis linked to repeated infusion of 0.9% sodium chloride. A delta-gap calculation supported the presence of mixed acidosis, while acute kidney injury and evidence of proximal tubular injury impaired the patient’s ability to handle the expanding chloride and acid load. Her condition deteriorated into coma, respiratory failure, suspected acute heart failure and worsening renal dysfunction. Continuous renal replacement therapy with bicarbonate-buffered replacement fluid was started 14 hours after admission, after which her blood pH returned to 7.42 within 12 hours. Because this was a single retrospective case, the report could establish a temporal association between the interventions and recovery, but it could not determine whether chloride loading, kidney injury or another mechanism was principally responsible for the refractory acidosis.

The patient, an Asian woman with a body-mass index of 25.4 kg/m², was admitted with five days of persistent mid-to-upper abdominal pain and six hours of progressive breathlessness, nausea, vomiting and deep, rapid breathing. She had consumed about 3.6 kg of durian with beer three days earlier without monitoring her glucose. Although she had received metformin, repaglinide, linagliptin and voglibose, the report said she had not adhered to treatment or monitored her blood glucose regularly. On arrival, she was somnolent and dehydrated, with a pulse of 128 beats per minute and a respiratory rate of 36 breaths per minute. Her finger-stick glucose was 32 mmol/L, ketones were 3.6 mmol/L, arterial pH was 6.974 and base excess was negative 28.5 mmol/L. Lactate was 1.9 mmol/L, serum chloride was 96.77 mmol/L and creatinine was 56.8 μmol/L. Abdominal computed tomography showed no pancreatic abnormality, and the clinicians diagnosed severe diabetic ketoacidosis.

Initial treatment consisted of fluid resuscitation with 0.9% sodium chloride and a continuous insulin infusion, with insulin administered at up to 15 units per hour. Glucose solution was introduced when blood glucose fell below 13.9 mmol/L, and potassium was supplemented when urine output was present. The early response appeared encouraging: after two hours, glucose had fallen to 16.4 mmol/L and ketones to 2.4 mmol/L. Yet six hours later, the patient developed worsening dyspnea and restlessness, oxygen saturation of 88% and a pulse of 134 beats per minute. Diffuse moist rales were heard throughout both lungs, and the treating team considered acute left heart failure; furosemide relieved the symptoms. At nine hours, blood pH was 6.9 and chloride had reached 115 mmol/L despite 6,550 ml of cumulative fluid intake and 4,600 ml of urine output. At 11 hours, 4.2 g of sodium bicarbonate was infused, but the measured pH did not normalize.

By 14 hours after admission, the patient was comatose. Her pH had fallen to 6.84, arterial oxygen tension to 65 mmHg and bicarbonate to 2.6 mmol/L, while chloride remained at 115 mmol/L and lactate was only 1.31 mmol/L. Creatinine had risen to 207.3 μmol/L, blood amylase to 754.4 U/L and NT-proBNP to 8,709 pg/mL. The PaO2/FiO2 ratio was 152.5 mmHg, supporting the presence of severe pulmonary compromise. The report characterized this stage as refractory acidosis complicated by acute heart failure and acute kidney injury. The combination of falling glucose, persistent severe acidemia, rising chloride and evidence of organ failure meant that a simple reading of DKA resolution based on glucose or ketone trends would have missed the broader deterioration. At this point, the patient’s condition was severe enough to require airway intubation and renal replacement therapy.

The clinicians used the delta-gap method to distinguish the components of the acid-base disorder. It compares the increase in anion gap with the decrease in bicarbonate: delta gap equals the measured anion gap minus the normal anion gap, minus the normal bicarbonate minus the measured bicarbonate. On admission, the measured anion gap was 16.1 mmol/L, compared with an assumed normal value of 10 mmol/L, while bicarbonate was 2.9 mmol/L against an assumed normal value of 24 mmol/L. The resulting delta gap was negative 15. The report notes that a value below approximately negative 6 suggests possible coexistence of high-anion-gap ketoacidosis and normal-anion-gap hyperchloremic acidosis. The large discrepancy indicated that the reduction in bicarbonate was much greater than the rise in anion gap could explain, consistent with a mixed disorder rather than ketoacidosis alone. The calculation was therefore used as a diagnostic clue within the clinical picture, not as a stand-alone test.

Fluid composition was central to the proposed explanation. The 0.9% sodium chloride given during initial resuscitation contains 154 mmol/L chloride, far above the patient’s initial plasma chloride concentration of 96.8 mmol/L. As chloride rose to 115 mmol/L, the authors attributed much of the hyperchloremic component to this iatrogenic load. The patient also had evidence of renal tubular injury: urinary beta-2-microglobulin exceeded 4,000 μg/L and serum beta-2-microglobulin was 8,166.71 μg/L, findings considered consistent with proximal tubular injury that might reduce bicarbonate reabsorption. The authors could not confirm renal tubular acidosis because urinary pH, urinary anion gap and urinary electrolytes were not measured. Acute kidney injury was judged insufficient on its own to explain the initial anion gap or pH of 6.84. Gastrointestinal bicarbonate loss, lactic acidosis, methanol or ethylene glycol poisoning and chloride-containing medications were considered unlikely on the available evidence, although severe lipemia prevented calculation of the osmolal gap.

The report also proposed a possible “peripheral metabolic pool” contribution to the apparent paradox. In severe DKA, poor tissue perfusion may allow ketones and other acidic metabolites to accumulate locally in muscle, skin and other peripheral tissues. If circulation later improves, those sequestered metabolites can return to the central bloodstream and temporarily intensify acidosis even while measured glucose and circulating ketones are falling. The patient’s blood glucose and ketones were lower before renal replacement therapy than on admission, but ketones rose paradoxically during the early period of continuous renal replacement therapy, reaching 6.2 mmol/L 9.5 hours after treatment began. The authors viewed this as indirect, rather than definitive, support for washout of peripheral ketones. The explanation remains a hypothesis because peripheral ketone concentrations and tissue perfusion were not measured directly, and the single case cannot establish that this mechanism caused the persistent acidosis.

Severe hypertriglyceridemia and probable acute pancreatitis provided another possible link in the clinical course. Triglycerides were 28.21 mmol/L, well above the 11.3 mmol/L threshold cited for hypertriglyceridemia-associated acute pancreatitis. The authors therefore considered severe hypertriglyceridemia the principal precipitant of the pancreatic episode, while the stress of pancreatitis and DKA could have worsened insulin resistance, lipolysis and ketosis in turn. The pancreatitis diagnosis was based on abdominal pain and an amylase concentration of about 5.6 times the upper limit of normal, satisfying two of the Atlanta criteria cited by the report. Diagnostic certainty was limited because serum lipase and follow-up pancreatic imaging were not obtained. These findings indicate that the case was not simply an uncomplicated response to sodium chloride; it also involved an acute abdominal illness that may have increased the metabolic stress surrounding DKA.

With the patient in coma, respiratory failure and worsening renal function, the team began continuous renal replacement therapy in continuous veno-venous hemodiafiltration mode and intubated her airway. The replacement fluid included bicarbonate, and the procedure also provided controlled ultrafiltration and precise fluid removal. This mattered because additional rapid fluid administration could worsen heart failure, while glucose-containing fluids alone would not directly correct the hyperchloremic acidosis. After 12 hours of treatment, the blood pH had returned to the normal range at 7.42; after 25 hours, the patient regained consciousness and the endotracheal tube could be removed. Creatinine initially reached 302.9 μmol/L but later declined to 69.6 μmol/L by day 15. She was discharged on day 15 with reported normal glucose, lipid and renal-function results. The sequence supports the feasibility of using renal replacement therapy for refractory mixed acidosis in this highly selected setting, but it does not establish that a particular component of the treatment caused the recovery.

The case’s main clinical lesson is that falling glucose and ketones do not necessarily mean that acid-base status is improving. The report recommends calculating the delta gap when DKA patients develop persistent or worsening acidosis, interpreting it alongside the full clinical and laboratory picture rather than treating a threshold as an automatic indication for renal replacement therapy. It also suggests considering lower-chloride balanced crystalloids after initial resuscitation, a point consistent with the cited subgroup analysis in which median resolution time was 13.0 hours with balanced crystalloids and 16.9 hours with 0.9% sodium chloride. Important limits remain: this was one patient, data were retrospective, urine-electrolyte and titratable-acid testing was absent, and several mechanisms were proposed rather than directly demonstrated. The report nevertheless identifies a recognizable and potentially dangerous scenario—severe DKA complicated by rising chloride, acute kidney injury and multiorgan failure—in which standard measures of ketoacidosis resolution may not capture the full acid-base disturbance. The patient subsequently reported regular glucose monitoring and adherence to a diabetic diet, although the case itself cannot establish whether those changes prevented recurrence.

Subject of Research: Iatrogenic hyperchloremic mixed acidosis during severe diabetic ketoacidosis

Article Title: Case Report: Refractory mixed acidosis with iatrogenic hyperchloremic acidosis in diabetic ketoacidosis

Article References: Lin, J., Zheng, J., Xu, X., Zhou, C., Yang, G., & Zhu, R. (2026). Case Report: Refractory mixed acidosis with iatrogenic hyperchloremic acidosis in diabetic ketoacidosis. Frontiers in Endocrinology, 17, Article 1969157. https://doi.org/10.3389/fendo.2026.1969157

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1969157

Keywords: diabetic ketoacidosis, hyperchloremic acidosis, mixed acidosis, continuous renal replacement therapy, acute kidney injury, sodium chloride, balanced crystalloids, delta gap, acute pancreatitis, hypertriglyceridemia, Case, Report