Euglycemic diabetic ketoacidosis (EDKA) is characterized by normal glucose level (< 200 mg/dL), metabolic acidosis (pH < 7.3 and serum bicarbonate < 18 mM/L), and ketonemia [1]. EDKA is often discovered incidentally during surgery due to unexplained metabolic acidosis in patients with risk factors. Changes in acid-base status may prompt the evaluation for EDKA. To diagnose EDKA, arterial blood gas analysis (ABG) for high anion gap metabolic acidosis (HAGMA) and serum glucose and ketone testing are essential in diabetic patients [2].

However, the traditional anion gap (AG) approach may be unsuitable for complex acid-base disorders. An alternative assessment of acid-base disorders is a mathematical model based on the physicochemical principles described by Stewart [3]. However, the Stewart approach is a complex method for clinical use. Thus, several studies [4,5] have attempted to simplify the Stewart approach to acid-base physiology and combine it with the base excess approach. Story et al. [6] described the simplified Fencl–Stewart approach for the partitioning of the standard base excess (SBE). This approach uses four equations to estimate the base excess effects of the important components: the strong ion difference (sodium and chloride), the total weak acid concentration (albumin) and unmeasured ions. Therefore, this approach allows the components of acid-base disorders to be quantified individually, providing a better understanding of the pathogenesis.

We present a case of a patient with type 2 diabetes who was diagnosed with EDKA, even though the AG approach failed to identify HAGMA during surgery. The patient exhibited normal AG metabolic acidosis according to ABG. However, subsequent partitioning of the SBE revealed the presence of unmeasured anions (UA) and a significant contribution of hyperchloremia to the total acidifying base deficit. We will explore the limitations of the AG approach and the advantages of SBE partitioning in the context of the diagnosis of HAGMA with hyperchloremic acidosis and the administration of bicarbonate therapy. Written informed consent has been obtained from the patient for the publication of this paper.

A 73-year-old female with a height of 141 cm and a weight of 64 kg underwent surgical adhesiolysis and enterostomy for intestinal obstruction. Her comorbidities included type 2 diabetes mellitus and dyslipidemia diagnosed 15 years prior, treated with metformin, linagliptin, glimepiride, and rosuvastatin. She visited the emergency room five days ago due to vomiting and abdominal pain, and an abdominal computed tomography scan revealed that she had a small bowel obstruction. Laboratory tests after admission were as follows. Blood tests showed hemoglobin of 11.5 g/dL, hemoglobin A1c of 7.5%, sodium of 136 mM/L, potassium of 4.4 mM/L, chloride of 106 mM/L, glucose of 309 mg/dL, albumin of 3.98 g/dL, bicarbonate of 18.5 mM/L, and lactate of 3.4 mM/L. The urine test results revealed that glucose was 3+, and protein and ketone were negative. She was hospitalized for intravenous fluid resuscitation and nasogastric decompression. Her home medications for diabetes mellitus were discontinued and sliding scale insulin was ordered, resulting in finger stick blood glucose (FSBG) levels decreasing to 104–152 mg/dL. The results of her laboratory tests measured the day before surgery were as follows. Blood tests showed hemoglobin of 10.1 g/dL, sodium of 134 mM/L, potassium of 3.4 mM/L, chloride of 102 mM/L, glucose of 104 mg/dl and albumin of 3.53 g/dL. The urine test revealed that glucose was 1+, protein was weakly positive, and ketone was negative. Other laboratory findings, including renal and hepatic functions, were unremarkable. She did not receive insulin and glucose overnight. The FSBG revealed 84 mg/dL on the morning of the surgery, and the infusion of regular insulin and dextrose solution was not given with concern of intraoperative hypoglycemia.

General anesthesia was induced with propofol 40 mg and rocuronium 50 mg. The patient was intubated, and anesthesia was maintained with sevoflurane, remifentanil, and rocuronium. In addition to standard anesthesia monitoring, which includes pulse oximetry, electrocardiogram, non-invasive blood pressure, capnography, temperature, etc., an arterial line was established for hemodynamic monitoring and blood sampling. The results of ABG with serum glucose and lactate 60 minutes after the initiation of operation were (Table 1).

In a simplified Fencl-Stewart approach for the partitioning of SBE [6,7], we employed the equation

SBE_NaCl indicated hyperchloremic acidosis. The difference between SBE and SBE_NaCl suggested that HAGMA might be developing. To confirm this, serum albumin should be measured. However, the actual serum albumin level was not measured during surgery as albumin has a lag time of about 60 min for the identification of the result in our laboratory. Therefore, based on the preoperative level of albumin, it was ‘estimated’ that the SBE_Albumin was ≥ +1.7 mM/L, and the SBE_UA was ≤ –4.4 mM/L, strongly indicating the development of HAGMA. This prompted testing for β-hydroxybutyrate. The measured serum β-hydroxybutyrate was 5.2 mM/L. The corrected AG (AGc) was estimated to be ≥ 14.3 mM/L, which is within the normal range of 8 to 16 mM/L.

Next, we estimated the quantitative contribution of hyperchloremia to the total acidifying base deficit, based on the previous studies [7,8]. The total acidifying base deficit (by excluding alkalinizing SBE partitions), as proposed in a study [7], was calculated as follows: SBE_NaCl + SBE_UA = –10 – 4.4 = –14.4 mM/L. Thus, the contribution of hyperchloremia (SBE_NaCl) to the total acidifying base deficit (SBE_NaCl + SBE_UA) was calculated to be 69% (–10/–14.4).

To alleviate concurrent hyperchloremic acidosis, we infused 80 mEq of sodium bicarbonate, targeting an increase in serum bicarbonate to 18 mM/L (a criterion for diagnosing EDKA), as calculated by the equation 0.5 × weight × [target serum bicarbonate – present serum bicarbonate] [9]. This resulted in a satisfactory decrease in metabolic acidosis (Table 1) and in the hyperchloremic contribution to the total acidifying base deficit (from 69% to 53%). Following the bicarbonate therapy, the surgery ended, and the patient was extubated and transferred to post anesthesia care unit (PACU). The operative time was 100 minutes, and blood loss was estimated between 100 and 150 mL. Intraoperatively, 600 mL of balanced salt solutions (Hartmann’s solution [JW Pharmaceutical] and Plasma Solution-A^® [HK innoN]) were administered, with a urine output of 100 mL. The ABG results checked in the PACU are shown in Table 1. Hyperventilation occurred because of metabolic acidosis.

The patient was transferred to the general ward and started on a regular insulin infusion in 5% dextrose. Immediately after arrival on the ward, follow-up serum albumin and β-hydroxybutyrate levels were 2.93 g/dL and 3 mM/L, respectively. The urine test conducted on the next day showed a weak positive for ketones. Despite this finding, the postoperative course was uneventful, and the patient was discharged in good health a week after surgery.

The diagnosis of EDKA often begins with the verification of HAGMA. However, when HAGMA and hyperchloremic acidosis occur together, the traditional AG (including AGc) approach may be unreliable [7]. Hyperchloremia can decrease the AG [10], making it insufficient to identify HAGMA, even if the blood ketones are high. In our case, partitioning of SBE was used to avert this problem and was more suitable for identifying HAGMA than the AG approach.

Partitioning of the SBE may also be advantageous in quantifying the contribution of hyperchloremic acidosis to mixed acidosis. SBE_NaCl was found to be more reliable for evaluating the effect of chloride on the acid-base status compared to other methods such as SBE_Cl [11]. By calculating SBE_NaCl and SBE_UA (mostly due to ketoacid anions), we were able to quantify the acidifying severity of each partition, including chloride and non-chloride ketoacid anions. This showed a significant contribution of hyperchloremic acidosis to the severity of mixed metabolic acidosis in this patient. The high degree of the hyperchloremic base deficit as a percentage of the total acidifying base deficit may suggest the necessity of treating hyperchloremia to alleviate the severity of acidosis.

There has been considerable debate regarding the pros (even in mild to moderate metabolic acidosis, pH < 7.3 and BE < –4mEq/L) [12,13] and cons [14] of bicarbonate therapy for the treatment of hyperchloremic acidosis. In our opinion, the treatment of hyperchloremic acidosis may have some advantages for the following reasons when hyperchloremic acidosis significantly contributes to the total acidifying base deficit. First, metabolic acidosis may cause discomfort for patients, such as hyperventilation, dyspnea, nausea, vomiting, hypotension, and tachycardia [9]. Bicarbonate therapy may facilitate the resolution of hyperchloremic metabolic acidosis [15]. As a result, bicarbonate infusion during surgery may shorten or alleviate the patient’s discomforts after surgery. We believe that hyperventilation would have been more severe after surgery if bicarbonate was not administered during surgery in this patient and waiting for spontaneous resolution may not be reasonable. Second, it has been reported that acid-base status is a factor that determines the severity of DKA [16]. Hyperchloremic acidosis is not usually dangerous, but it increases the severity of metabolic acidosis. In addition, hyperchloremia has been linked to prolonged hospital stay, and reduced kidney function, in DKA [17]. Hyperchloremia can lead to adverse clinical outcomes even in the absence of severe acidosis [15]. Therefore, treating hyperchloremic acidosis may be preferable to not treating it as it could help reduce the severity of mixed acidosis.

Some may argue that using preoperative albumin level would miscalculate SBE_UA and hyperchloremic contribution to the total acidifying base deficit because the intraoperative albumin level may differ from the preoperative level. Because it took a long time to obtain results, albumin levels were not measured during surgery, which is a limitation of this study. Due to potential surgery-induced leakage of albumin [18] and fluid therapy, it is more likely that the level decreased rather than increased or remained unchanged. The decrease in serum albumin after surgery in this patient suggests that the albumin level during surgery would have been between the preoperative (3.53 g/dL) and postoperative (2.93 g/dL) levels. This could have decreased the SBE_UA (between –4.4 and –5.9 mM/L) and the contribution of hyperchloremia to the total acidifying base deficit (between 53% and 45%). Therefore, it is necessary to take this into account when calculating partitions of SBE.

In summary, partitioning of the SBE can be advantageous for detecting UA and may be used to determine the necessity of intraoperative bicarbonate therapy when HAGMA and hyperchloremic acidosis occurred together. Bicarbonate therapy may potentially reduce the severity of mixed metabolic acidosis and any related symptoms. Treating hyperchloremic acidosis may be better than not treating it to reduce the severity of mixed acidosis.