A 56-year-old Chinese man with a history of poorly controlled type 2 diabetes mellitus was admitted to Ng Teng Fong General Hospital, Singapore, with a lower back abscess over 2 weeks. His last hemoglobin A1c was 8.9% (5 months prior to admission), and he was on oral hypoglycemic agents (metformin 750 mg 3 times a day and glipizide 5 mg 2 times a day) for control of type 2 diabetes mellitus, but he was not compliant to treatment. He sustained an abrasion over his back and noted worsening of swelling, increasing pain, and pus discharge over 2 days prior to admission. On examination, there was a 5 × 5-cm abscess with necrotic tissue over the mid lower back, which was associated with an area induration and bruising. Diabetic dermopathy was noted over bilateral lower limbs.
He was treated with intravenous (IV) Co-amoxiclav (amoxicillin/clavulanic acid) and surgical wound debridement with application of negative-pressure wound treatment (VAC dressing) under general anesthesia, followed by definitive antibiotic therapy after tracing complete culture results. Wound swab culture grew methicillin-sensitive Staphylococcus aureus and postdebridement tissue cultures grew methicillin-sensitive S. aureus and Klebsiella pneumoniae, which were both sensitive to cefipime. Hence, he was continued on IV cefipime to complete for 3 weeks.
He was monitored for blood glucose (BG) levels using a handheld point-of-care testing (POCT) glucometer (Accu-Chek Inform II; Roche Diagnostics, Basel, Switzerland) as part of routine nursing protocols for diabetic inpatients, during hospitalization. Upon hospitalization, he was restarted on oral hypoglycemics at old doses. However, doses were up-titrated because of poor BG control (metformin was increased to 1700 mg every morning and 850 mg every night. Glipizide was continued at the same old dose. He was later started on insulin (Mixtard) 12 U every morning and 8 U every night, which helped him achieve good glycemic control during acute illness.
During the hospitalization (on day 12 since admission), he was noted to have an unusually high BG reading. It read as 33.1 mmol/L during morning BG monitoring, and the test was repeated a few minutes later, which surprisingly showed a BG of 9.3 mmol/L. Once again, the next-day evening BG reading was noted to be 32.6 mmol/L, and it was immediately repeated twice, and values were 16.9 and 9.8 mmol/L, respectively (Fig. 1). On both occasions, BG levels were repeated as per nursing protocols and in view of poor clinical correlation and recent inpatient BG trends.
It was noted that the patient's glycemic control had been reasonable (4.6–11.4 mmol/L) over the preceding week, and he had been compliant to oral hypoglycemics and insulin regimen throughout hospitalization. Moreover, the patient was clinically and biochemically improving and was not septic at the time of abnormal BG reading. He did not have symptoms to suspect a hyperglycemic crisis either; hence, workup for hyperglycemic crisis was not carried out.
Point-of-care testing glucometer that was used to measure BG levels was checked and calibrated during investigation process. Accuracy and precision of the glucometer were well maintained. Furthermore, glucometer strips had been kept in airtight containers, at room temperature at all times, and they were not close to expiration date. Upon further investigation to rule out other possible preanalytical causes, we found out that the patient was eating preserved sweet plums (Prunus domestica) 1 to 2 hours before BG monitoring, on both occasions. According to the patient, he did not wash his hands after eating sweet plums. Therefore, we planned a simple experiment to replicate this clinical scenario, in order to validate our suspicion of possible spurious BG results (pseudohyperglycemia) secondary to contamination of fingertips with sugary foods prior to POCT glucose measurements.
A voluntary nondiabetic participant was recruited for the experiment. Fingertip BG levels were measured using the same handheld POCT glucometer that was used in this case scenario. Point-of-care testing glucose level was measured for the subject under 4 different conditions. First POCT glucose reading was taken during fasting state before handling sweet plums. Second reading was taken 5 minutes after handling sweet plums, but neither hand washing nor swabbing of fingertip with alcohol swabs was carried out. Third reading—after cleaning the fingertip with 1 alcohol swab—and similarly fourth reading were taken after cleaning the fingertip with 3 alcohol swabs (Table 1).
Pathophysiology of Hyperglycemia in Acutely Ill Patients
In a healthy individual, serum glucose level is tightly regulated. In a state of overnight fast, glycemia occurs as the rate of hepatic gluconeogenesis tallies with glucose uptake. Rising serum glucose levels following a meal are accompanied by a rapid increase in insulin level and a prompt decrease in glucagon level, which in turn reduces hepatic gluconeogenesis and increases peripheral glucose uptake, thereby preventing BG levels from exceeding 8.3 mmol/L, at its peak.1 During illness, stress increases the concentration of counterregulatory hormones such as glucagon, catecholamine, glucocorticoid and growth hormone, and cytokines. These hormones cause hyperglycemia by increasing hepatic gluconeogenesis and by decreasing peripheral glucose uptake. Furthermore, cytokines may have hyperglycemic effects through stimulation of counterregulatory hormone secretion.2
Causes of Hyperglycemia
Commonest cause of hyperglycemia is poorly controlled diabetes mellitus (both types 1 and 2 diabetes). Diabetic ketoacidosis and hyperosmolar hyperglycemic state (also known as hyperosmotic hyperglycemic nonketotic state) are 2 of the most serious complications of diabetes in which severe hyperglycemia is a feature.
Besides, hyperglycemia in acutely ill patients is a common clinical scenario. Metabolic stress of surgery, sepsis, or trauma contributes to hyperglycemia in nondiabetic patients.3 Hyperglycemia can also result from provision of excessive calories, as well as dextrose infusions that are commonly used for fluid resuscitations and delivery of IV medications. Medications formulated in fat emulsion such as propofol provide equal amount of calories to that of 10% fat emulsion, which is used in parenteral nutrition. It is important not to forget the medications that could precipitate hyperglycemia, which include corticosteroids, sympathomimetics, and immunosuppressants.1
Causes of Pseudohyperglycemia
Pseudohyperglycemia is spuriously elevated BG results. Causes of pseudohyperglycemia could be described under 3 main categories: preanalytical, analytical, and postanalytical. Poor hand hygiene after handling sugar-containing food products such as fruits is a rare but a known preanalytical cause of pseudohyperglycemia.4,5 However, it was not observed when fingertip was cleaned properly with alcohol swabs or hand washing was carried out.4 Luckily, instant hand sanitizer (contains alcohol) is compatible with the results of POCT glucometers,6 but, use of 10% povidone-iodine solution for cleaning resulted in overestimated BG values.7
Coding procedure (most modern glucometers no longer require a coding step), inappropriate storage of glucometer strips, expired strips, and use of machine-incompatible strips are also common preanalytical causes of pseudohyperglycemia. Practically, emptying old strips into newly opened containers by hospital staff could result in using expired strips for patients. Furthermore, heating of glucometer strips or changes of temperature from low to room temperature and high relative humidity could cause pseudohyperglycemia.8
Analytical Causes. Low oxygen partial pressure associated with high altitudes (usually >2000 m) causes both underestimation and overestimation of BG results.9 Electromagnetic interferences from mobile phones also have been reported to impact BG measurements.8 In addition, most handheld POCT glucometers show positive bias to lower hematocrit levels, low oxygen levels (eg, patients with severe COPD), very high uric acid levels (patients who undergo chemotherapy, radiotherapy, or certain cancers with rapid cell turnover), and certain medications.8–11 Elevated plasma acetaminophen levels may cause pseudohyperglycemia.8 Besides, ascorbic acid, dopamine, maltose, and mannitol have been found to interfere POCT glucometer measurements.12
Glucometers can measure nonglucose carbohydrate molecules (galactose, maltose) as well. Icodextrin is a glucose polymer used in peritoneal dialysis, which is partially absorbed (20%–30%) into blood circulation and metabolized to oligosaccharides (maltose) and known to cause pseudohyperglycemia.8,13 Pseudohyperglycemia was also reported in an infant, with classic galactosemia, which resolved after weaning and initiation of feeding with a low-lactose formula.14
Erroneous transcription of BG results is the only possible postanalytical cause of pseudohyperglycemia, which is minimized with enhanced connectivity of newer POCT glucometers.
In a nutshell, causes of spuriously elevated BG levels are quite limited, but mostly related to POCT glucometers, which are commonly used in hospital, outpatient departments, other health care agencies, and home glucose monitoring settings. These unreliable results may adversely affect clinical decision, resulting in potential harm to patients, and it carries a high risk of overtreatment, which in turn could cause hypoglycemia and its potential lethal consequences.
The authors thank Ms Lim Sian Foong for her support in identifying this case and during the investigation process.
1. Montori VM, Bistrain BR, McMahon MM. Hyperglycemia in acutely ill patients. JAMA
2. McCowen KC, Malhotra A, Bistrian BR. Stress-induced hyperglycemia. Crit Care Clin
3. Corl DE, Greenfield L, Hoofnagle A, et al. Measuring point-of-care blood glucose in critically ill patients. Nurs Crit Care
4. Olamoyegun MA, Oloyede T, Adewoye OG, et al. Pseudohyperglycemia
: effects of unwashed hand after fruit peeling or handling on fingertips blood glucose monitoring results. Ann Med Health Sci Res
5. Hirose T, Mita T, Fujitani Y, et al. Glucose monitoring after fruit peeling: pseudohyperglycemia
when neglecting hand washing before fingertip blood sampling. Diabetes Care
6. Mahoney JJ, Ellison JM, Glaeser D, et al. The effect of an instant hand sanitizer on blood glucose monitoring results. J Diabetes Sci Technol
7. Lipshutz AM, Hawes EM. Probable glucometer
interference caused by topical iodine solution test site preparation. J Clin Pharm Ther
8. Erbach M, Freckmann G, Hinzmann R, et al. Interferences and limitations in blood glucose self-testing: an overview of the current knowledge. J Diabetes Sci Technol
9. Ginsberg BH. Factors affecting blood glucose monitoring: sources of errors in measurement. J Diabetes Sci Technol
10. Tang Z, Judith HL, Louie RF, et al. Effects of different hematocrit levels on glucose measurements with handheld meters for point-of-care testing
. Arch Pathol Lab Med
11. Bishop ML, Fody EP, Schoeff LE. Clinical Chemistry: Principles, Procedures, Correlations
. Baltimore, MD: Lippincott Williams & Wilkins; 2005.
12. Tang Z, Du X, Louie RF, et al. Effects of drugs on glucose measurements with handheld glucose meters and a portable glucose analyser. Am J Clin Pathol
13. Perera NJ, Stewart PM, Williams PF, et al. The danger of using inappropriate point-of-care glucose meters in patients on icodextrin dialysis. Diabet Med
14. Özbek MN, Öcal M, Tanriverdi S, et al. Capillary bedside blood glucose measurement in neonates: missing a diagnosis of galactosemia. J Clin Res Pediatr Endocrinol