How the 72-Hour Fasting Test is Performed
The 72-hour fasting test is a supervised inpatient procedure that begins when the patient stops eating, requires discontinuation of nonessential medications, permits only calorie-free beverages, and involves serial blood sampling every 6 hours initially (then every 1-2 hours once glucose drops below 60 mg/dL) to measure glucose, insulin, C-peptide, proinsulin, and beta-hydroxybutyrate until specific endpoints are reached. 1
Test Initiation and Setup
- Mark the start time as when the patient last ate food
- Stop all nonessential medications before beginning
- Allow calorie-free beverages throughout the test (water, black coffee, unsweetened tea)
- Require physical activity during waking hours—the patient should not remain sedentary 1
Blood Sampling Protocol
Initial Phase (Glucose ≥60 mg/dL)
- Draw blood every 6 hours for:
- Plasma glucose
- Insulin
- C-peptide
- Proinsulin
- Beta-hydroxybutyrate (BHB)
Accelerated Phase (Glucose <60 mg/dL)
- Increase sampling frequency to every 1-2 hours once glucose drops below 60 mg/dL (3.3 mmol/L)
- Only send insulin, C-peptide, and proinsulin for analysis when glucose is <60 mg/dL to avoid unnecessary costs 1
Critical Endpoints for Stopping the Test
The test ends when ANY of these criteria are met:
Primary endpoint: Plasma glucose <45 mg/dL (2.5 mmol/L) AND the patient has symptoms/signs of hypoglycemia 1
Alternative endpoint: Plasma glucose <55 mg/dL (3.0 mmol/L) without symptoms IF Whipple's triad was previously documented unequivocally 1
Time endpoint: 72 hours have elapsed without symptoms 1
Newer metabolic endpoint: BHB >2.7 mmol/L, which indicates insulin suppression and effectively rules out hyperinsulinemic hypoglycemia 2
Critical Procedural Caveats
Use Laboratory-Measured Glucose, Not Point-of-Care Monitors
Do not base the decision to end the fast on bedside glucose monitors—these are insufficiently accurate for diagnostic purposes. Always use laboratory-measured plasma glucose concentrations to determine endpoints. 1
Do Not End Based on Low Glucose Alone
A low glucose without symptoms is insufficient to diagnose hypoglycemia, particularly in women and children who may have physiologically low glucose during prolonged fasting. The combination of low glucose AND symptoms is required (unless Whipple's triad was previously documented). 1
If Urgent Treatment is Needed
If severe symptoms require immediate treatment before completing the protocol, obtain all blood samples BEFORE administering carbohydrates 1
Final Testing at Endpoint
When the fast ends, immediately:
Collect final samples for:
- Plasma glucose
- Insulin
- C-peptide
- Proinsulin
- Beta-hydroxybutyrate
- Oral hypoglycemic agents (screen for factitious hypoglycemia)
Administer 1.0 mg glucagon IV
Measure plasma glucose at 10,20, and 30 minutes post-glucagon to assess glycemic response
Measure insulin antibodies (can be done later, not necessarily during hypoglycemia) 1
Modern Modifications to Reduce Hospitalization Time
Recent evidence suggests the test can be shortened without missing insulinoma cases:
BHB >2.7 mmol/L criterion: Adding this stopping point reduced average fast duration from 57.6 to 49.7 hours, saving 376.5 cumulative inpatient hours without missing any insulinoma cases (all 7 insulinoma cases developed hypoglycemia within 43 hours with BHB ≤1.2 mmol/L) 2
91.7% of insulinomas are diagnosed within 24 hours of starting the fast, suggesting shorter protocols may be adequate for many patients 3
Serial BHB measurements showing two successive values above the 18-hour baseline level strongly predict a negative fast 4
Practical Implementation Points
The test requires close supervision with nursing staff trained to recognize hypoglycemic symptoms and obtain samples at appropriate intervals. System-level interventions including standardized order sets and clinical decision support tools significantly improve test accuracy and reduce unnecessary resource utilization 5. The patient must remain active during waking hours to provoke hypoglycemia if an insulinoma is present 1.