53-Year-Old Man With Dyspnea

A 53-year-old man with type 2 diabetes mellitus treated with metformin, pioglitazone, and empagliflozin presented to the emergency department with acutely worsening dyspnea for 1 day. Five days before onset of dyspnea, he developed chills followed by a nonproductive cough. He had not had any recent known sick contacts, recent travel, or medication changes. In the emergency department, vital signs included a temperature of 37 ºC, heart rate of 103 beats/min, blood pressure of 136/76 mm Hg, respiratory rate of 28 breaths/min, and oxygen saturation of 99% on room air. Physical examination found tachypnea without other notable findings. Because of his dyspnea, he was unable to complete a full sentence without taking a breath. He had no notable murmurs and pulmonary examination was without adventitious sounds. A complete blood count revealed a leukocyte count of 16.3×109 /L (reference [ref], 3.4 to 9.6 ×109/L), hemoglobin of 13.8 g/dL (ref, 13.2 to 16.6 g/dL), and a platelet count of 277×109 (ref, 135 to 317 ×109/L). A basic metabolic panel showed sodium 130 mM (ref, 135 to 145 mM), chloride 92 mM (ref, 98 to 107 mM), blood urea nitrogen 24 mg/dL (ref, 18 to 24 mg/dL), potassium 5.1 mM (ref, 3.6 to 5.2 mM), bicarbonate (HCO3) less than 5 mM (ref, 22 to 29 mM), creatinine 1.04 mg/dL (ref, 0.74 to 1.35 mg/dL), and a glucose level of 166 mg/dL (ref, 70 to 140 mg/dL). Urinalysis found a glucose level of 500 mg/dL and ketonuria with no reference range provided for either of these values. Chest x-ray was notable for patchy airspace opacities in the right lower lobe. A severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) polymerase chain reaction nasal swab was positive. He was started on ceftriaxone and doxycycline and was admitted to the general medical service with diagnosis of coronavirus disease 2019 (COVID-19) with possible bacterial superinfection.1.What is the best next test for assessing the most likely underlying cause of this patient’s shortness of breath?a.Electrocardiogramb.Blood gas analysisc.D-dimerd.Computed tomography (CT) angiogram of the cheste.No further workup is required, symptoms are likely secondary to his pneumonia and/or COVID-19 The differential diagnosis for shortness of breath is broad and involves multiple organ systems; therefore, the next step in evaluation should follow the leads provided by the history, physical exam, and initial test results. Generally speaking, an electrocardiogram should be obtained when there is concern for a new myocardial infarction or symptomatic arrhythmia as a cause of shortness of breath. In this patient, with no prior cardiac history, no syncopal events, no chest pain, and 5 days of symptoms, a cardiac cause for his presentation is unlikely. The undetectable HCO3 on the basic metabolic panel should lead to further acid-base analysis. As such, blood gas analysis is the correct answer. That said, in this case, workup for a pulmonary embolism was pursued first, which delayed this patient’s care—as is common with this particular diagnosis. D-dimer could be considered; however, with the patient’s significant degree of respiratory distress, one would expect some degree of hypoxia if pulmonary embolism was the cause; therefore, pulmonary embolism is less likely. Although a pulmonary embolism should be considered, the patient is low risk for pulmonary embolism by Well’s criteria—CT angiogram of the chest would not be the next best test. D-dimer was obtained with a Wells score of 1.5 and a revised Geneva score of 5 and was elevated at 1,921 ng/mL (ref, <500 ng/mL). He was sent for a CT angiogram, which was negative for pulmonary embolism and re-demonstrated patchy airspace opacities in the right lower lobe. Computed tomography showed no changes of other pulmonary parenchymal or airspace disease. Given this patient’s respiratory distress, severe acidosis, and symptoms out of proportion to findings on chest imaging, simply continuing antibiotics without further workup would be inappropriate. Blood gas analysis was found with pH 7.09 (ref, 7.35 to 7.45), partial pressure of carbon dioxide (PaCO2) less than 15 mm Hg (ref, 35 to 40 mm Hg), and partial pressure of oxygen (PaO2) 111 mm Hg (ref, 83 to 108 mm Hg) on room air. A repeat basic metabolic panel revealed sodium 129 mM (ref, 135 to 145 mM), chloride 99 mM (ref, 98 to 107 mM), blood urea nitrogen 21 mg/dL (ref, 18 to 24 mg/dL), potassium 4.6 mM (ref, 3.6 to 5.2 mM), HCO3 less than 5 mM (ref, 22 to 29 mM), creatinine 0.72 mg/dL (ref, 0.74 to 1.35 mg/dL), and a glucose level of 163 mg/dL (ref, 70 to 140 mg/dL).2.What is the most likely cause of this patient’s shortness of breath and tachypnea?a.COVID-19 infectionb.Pulmonary embolismc.Bacterial pneumoniad.Respiratory compensation for metabolic acidosise.Chronic obstructive pulmonary disease exacerbation Although this patient was found to test positive for COVID-19, attributing his shortness of breath to his viral infection would likely represent early closure as it is unlikely for a viral infection to cause this degree of dyspnea with normal oxygen saturation and PaO2. Additionally, his CT scan did not have prominent features of viral pneumonia. Pulmonary embolism has been effectively ruled out based on the CT scan results. Similarly, although this patient may have bacterial pneumonia or aspiration pneumonitis, it would be unlikely to cause this degree of dyspnea without other findings such as hypoxia. To compensate for metabolic acidosis, respiratory drive will increase to remove carbon dioxide from the circulation and is the likely driver of this patient’s tachypnea and dyspnea; therefore, respiratory compensation for metabolic acidosis is the correct answer. Exacerbation of chronic obstructive pulmonary disease is unlikely in a patient without a smoking history, wheezing, or evidence of emphysema or bronchitis on chest imaging. Additionally, this would generally lead to hypercarbia and respiratory acidosis, which is not observed in this patient’s presentation. Given the patient’s clinical appearance and lab abnormalities at this point in his presentation, he was transferred to the medical intensive care unit for further treatment and evaluation.3.Which of the following is the most likely acid-base status of this patient?a.Anion gap metabolic acidosis with respiratory compensationb.Non-anion gap metabolic acidosisc.Primary respiratory acidosisd.Primary respiratory alkalosise.Anion gap metabolic acidosis with respiratory alkalosis Many algorithms have been proposed for determining the acid-base status of a patient.1Berend K. de Vries A.P. Gans R.O. Physiological approach to assessment of acid-base disturbances.N Engl J Med. 2014; 371: 1434-1445Crossref PubMed Scopus (166) Google Scholar Briefly, after clinical evaluation, it is prudent to start by looking at the pH followed by the PaCO2 to determine the primary driver of an acid-base disturbance. That said, strict pH parameters may not apply in mixed acid-base disorders. In this patient, the pH was 7.01, the PaCO2 was low, and the HCO3 was low, indicating a primary metabolic acidosis with respiratory compensation. In a primary metabolic acidosis, Winter Formula can be used to determine the anticipated respiratory compensation (expected PaCO2=1.5∗HCO3+8±2).2Albert M.S. Dell R.B. Winters R.W. Quantitative displacement of acid-base equilibrium in metabolic acidosis.Ann Intern Med. 1967; 66: 312-322Crossref PubMed Google Scholar In this case, with an estimated bicarbonate of 4, the expected PaCO2 would be 14±2 mm Hg which correlates with the findings on arterial blood gas and indicates anion gap metabolic acidosis with respiratory compensation without a secondary respiratory acid-base disturbance. A non-anion gap acidosis would be characterized by an elevated chloride and low bicarbonate resulting in the maintenance of a normal anion gap. Respiratory acidosis is generally identified with a pH less than 7.35 and a PaCO2 greater than 40 mm Hg, whereas a respiratory alkalosis would have a pH greater than 7.45 and a PaCO2 less than 40 mm Hg. If this patient did not have appropriate respiratory compensation, then the PaCO2 would not be appropriately low. In this case, the patient’s PaCO2 is appropriately low, reflecting normal respiratory compensation in the setting of a primary metabolic acidosis. The identification of an anion gap acidosis, with a calculated anion gap in this patient of 26, lead to a structured differential diagnosis. Upon arrival at the medical intensive care unit, he was alert, interactive, and tachypneic.4.Which test is most likely to explain the etiology of this patient’s metabolic acidosis?a.Beta-hydroxybutyrateb.D-lactatec.Methanol leveld.Thiamine levele.5-oxoproline Historically, two mnemonics have been popular for recalling the differential diagnosis of an anion gap metabolic acidosis: MUDPILES (methanol, uremia, diabetes, paraldehyde, iron, lactate, ethylene glycol, and salicylate) and KUSMALE (ketoacidosis, uremia, salicylate, methanol, aldehyde, lactate, and ethylene glycol). As excessive aldehyde is becoming increasingly rare, three more organic anion-gap producing acids have become widely known (5-oxoproline, D-lactate, and propylene glycol) and a new mnemonic, GOLD MARK, has been more widely used.3Mehta A.N. Emmett J.B. Emmett M. GOLD MARK: an anion gap mnemonic for the 21st century.Lancet. 2008; 372: 892Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar This pneumonic includes glycols (both ethylene and propylene), oxoproline, L-lactate, D-lactate, methanol, aspirin, renal failure, and ketoacidosis. This patient had ketoacids in his urine and is on a sodium-glucose cotransporter-2 inhibitor (SGLT2i); as such, euglycemic diabetic ketoacidosis should be considered, and measurement of beta-hydroxybutyrate would be helpful. The patient had no history that was concerning for D-lactic acidosis, such as short bowel or suspicion of bacterial overgrowth. Without known ingestion or alcohol use disorder, methanol toxicity is unlikely. Thiamine deficiency may contribute to type B lactic acidosis via the inability of pyruvate dehydrogenase to shuttle pyruvate through aerobic metabolism; however, its measurement is not frequently pursued, and empiric supplementation may be more cost-effective if there is high clinical concern.4Patel J.J. Bergl P. Thiamine and difficulties in differentiating type A from B lactic acidosis.Crit Care Med. 2019; 47: e434-e435Crossref Scopus (2) Google Scholar 5-oxoproline is a metabolite that can build up after long-term acetaminophen use. This patient had no history of chronic pain or over-the-counter pain medication use. The beta-hydroxybutyrate in this patient was measured at 4.3 mM.5.Which treatment is the most appropriate to start next?a.Balanced intravenous fluid resuscitationb.Insulin infusionc.Insulin bolusd.Bicarbonate ampulee.Potassium supplementation With an HCO3 of less than 18 mM, anion gap of greater than 12, pH less than 7.3, and beta-hydroxybutyrate of greater than 3 mM, this patient likely has shortness of breath secondary to high respiratory drive to compensate for euglycemic diabetic ketoacidosis (DKA) in the setting of acute illness while on an SGLT2i. Regardless of the underlying cause, the first treatment for any patient with DKA involves rapid correction of dehydration and electrolyte abnormalities.5Long B. Lentz S. Koyfman A. Gottlieb M. Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management.Am J Emerg Med. 2021; 44: 157-160Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar Depending on the patient’s sodium level, volume expansion with a balanced solution, such as PlasmaLyte or Lactated Ringer’s, is generally recommended as 0.9% saline may lead to additional hyperchloremic non-anion gap metabolic acidosis and may prolong DKA resolution.6Self W.H. Evans C.S. Jenkins C.A. et al.Clinical effects of balanced crystalloids vs saline in adults with diabetic ketoacidosis: a subgroup analysis of cluster randomized clinical trials.JAMA Netw Open. 2020; 3e2024596Crossref Scopus (36) Google Scholar After initial fluid resuscitation, ketogenesis must be stopped and reversed with continuous insulin and dextrose infusion, even if the blood glucose is within normal limits. A bolus of insulin is often given before initiating an insulin infusion in patients with hyperglycemia but should be avoided in euglycemic DKA as it may cause hypoglycemia. Bicarbonate bolus is rarely needed and is controversial in ketoacidosis but can be considered if the pH is less than 6.9.7Chua H.R. Schneider A. Bellomo R. Bicarbonate in diabetic ketoacidosis — a systematic review.Ann Intensive Care. 2011; 1: 23Crossref PubMed Google Scholar If a patient’s measured potassium level is lower than 3.5 mM, it is generally recommended to supplement with potassium before initiating insulin to avoid profound hypokalemia. The patient received balanced crystalloid resuscitation followed by insulin infusion with subsequent resolution of his metabolic acidosis and dyspnea. The US Food and Drug Administration first approved SGLT2i in 2013; it reduces glucose reabsorption in the proximal tubule of the kidney leading to lower blood glucose and improved hemoglobin A1C8Musso G. Gambino R. Cassader M. Pagano G. A novel approach to control hyperglycemia in type 2 diabetes: sodium glucose co-transport (SGLT) inhibitors: systematic review and meta-analysis of randomized trials.Ann Med. 2012; 44: 375-393Crossref PubMed Scopus (229) Google Scholar; and SGLT2i treatment has been increasing in popularity over the last decade due to recent evidence showing therapeutic benefit in diabetes, chronic kidney disease, and heart failure.9Perkovic V. Jardine M.J. Neal B. et al.Canagliflozin and Renal Outcomes in Type 2 Diabetes and Nephropathy.N Engl J Med. 2019; 380: 2295-2306Crossref PubMed Scopus (3279) Google Scholar Euglycemic DKA (eDKA) was first described in 1973 by Munro et al.10Munro J.F. Campbell I.W. McCuish A.C. Duncan L.J. Euglycaemic diabetic ketoacidosis.Br Med J. 1973; 2: 578-580Crossref PubMed Google Scholar There is an increased risk of eDKA with SGLT2i, which is augmented by surgery, excessive alcohol intake, physical exertion, and dietary—especially carbohydrate—restriction.11Goldenberg R.M. Berard L.D. Cheng A.Y.Y. et al.SGLT2 inhibitor-associated diabetic ketoacidosis: clinical review and recommendations for prevention and diagnosis.Clin Ther. 2016; 38: 2654-2664.e1Abstract Full Text Full Text PDF PubMed Scopus (185) Google Scholar Because of the increased risk for DKA, SGLT2i treatment is not recommended in patients with type 1 diabetes despite the cardiovascular and renal benefits. To reduce the risk of eDKA, some have recommended prophylactically holding SGLT2i during acute illness, when on restricted diets, when at risk for dehydration, during excessive alcohol intake, and for up to 3 days before planned surgeries. Measurement of urinary ketones can also be obtained to screen for eDKA; however, this is generally not recommended in patients receiving SGLT2i because their measurement can be misleading and SGLT2i may increase renal reabsorption of ketones.11Goldenberg R.M. Berard L.D. Cheng A.Y.Y. et al.SGLT2 inhibitor-associated diabetic ketoacidosis: clinical review and recommendations for prevention and diagnosis.Clin Ther. 2016; 38: 2654-2664.e1Abstract Full Text Full Text PDF PubMed Scopus (185) Google Scholar,12Handelsman Y. Henry R.R. Bloomgarden Z.T. et al.American Association of Clinical Endocrinologists and American College of Endocrinology position statement on the association of SGLT-2 inhibitors and diabetic ketoacidosis.Endocr Pract. 2016; 22: 753-762Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar As these medications continue to be prescribed at increased rates, we anticipate that there will be more cases of eDKA. Unfortunately, the lack of hyperglycemia often leads to a delay in diagnosis and effective treatment, as was observed in this case.5Long B. Lentz S. Koyfman A. Gottlieb M. Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management.Am J Emerg Med. 2021; 44: 157-160Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar In general, the pathophysiology of DKA is well established and involves a relative (increased metabolic demand without increase in available serum glucose and insulin) or absolute insulin deficiency with an associated increase in glucagon production. This leads to mobilization and metabolism of free fatty acids with resultant ketoacidosis. Common precipitants for DKA include surgery, strenuous exercise, myocardial infarction, stroke, infection, fasting state, and other metabolically stressful events.12Handelsman Y. Henry R.R. Bloomgarden Z.T. et al.American Association of Clinical Endocrinologists and American College of Endocrinology position statement on the association of SGLT-2 inhibitors and diabetic ketoacidosis.Endocr Pract. 2016; 22: 753-762Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar For this patient, we posit that his underlying COVID-19 infection, in combination with reduced oral intake, lead to his DKA presentation. In eDKA, specifically, there is also reduced glucose availability or production; this is often observed in a fasting state leading to further difficulty differentiating starvation ketosis from eDKA. This euglycemic presentation is most frequently found in patients with type 1 diabetes, pregnant patients, patients on a ketogenic diet, patients who received insulin before presentation, and patients losing glucose through their urine secondary to SGLT2i.5Long B. Lentz S. Koyfman A. Gottlieb M. Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management.Am J Emerg Med. 2021; 44: 157-160Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar This loss of glucose in the urine also leads to excessive osmotic diuresis; hence, these patients may present with profound hypovolemia. Additionally, SGLT2i may also reduce the clearance of ketone bodies, which may contribute to their association with eDKA.5Long B. Lentz S. Koyfman A. Gottlieb M. Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management.Am J Emerg Med. 2021; 44: 157-160Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar Although many hospitals have clear procedures for DKA, there may be some confusion when treating eDKA. As with DKA, monitoring and repleting potassium and volume are of the utmost importance. The initial treatment for most patients presenting with DKA is intravenous fluid resuscitation, which remains true in the setting of eDKA. When treating eDKA, it is imperative to stop any offending agent and reverse the underlying pathologic insulin insufficiency by treating with insulin therapy.12Handelsman Y. Henry R.R. Bloomgarden Z.T. et al.American Association of Clinical Endocrinologists and American College of Endocrinology position statement on the association of SGLT-2 inhibitors and diabetic ketoacidosis.Endocr Pract. 2016; 22: 753-762Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar As these patients are euglycemic, insulin bolus should be avoided, and insulin drip should be started. To prevent hypoglycemia, a continuous infusion of dextrose containing solution should be administered as well. Simultaneous initiation of dextrose-containing fluids and intravenous insulin led to rapid improvement in this patient’s acidosis and dyspnea. After his anion gap had improved, he was able to be taken off the insulin drip and was transitioned to subcutaneous insulin with blood sugar monitoring to avoid hypoglycemia. Ideally, short-acting subcutaneous insulin is started with a meal and the insulin infusion can be stopped 2 hours later. Notably, eDKA in the setting of SGLT2i use is not a contraindication to future use, and the medication can often be restarted after hospitalization.5Long B. Lentz S. Koyfman A. Gottlieb M. Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management.Am J Emerg Med. 2021; 44: 157-160Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar The authors report no competing interests.