Insulin-induced hypoglycemia

Insulin-induced hypoglycemia at Saint Barnabas Medical Center was not an isolated medical anomaly but a repeating pattern: patients receiving heparin drips suffered severe blood-sugar crashes, improved when the drips were removed, and deteriorated again when they were restarted. Laboratory tests eventually found insulin in the bags and evidence that the insulin in patients’ blood was foreign to their bodies. The pattern shifted the hospital’s concern from accidental treatment errors toward the possibility that someone was deliberately contaminating IV bags.

The First Suspicious Bag

On February 11, 1991, pharmacy nurse Pam Allen brought a suspicious IV bag to risk manager Karen Seiden. Its port looked as though it had been used, yet the bag was filled so full that it was leaking. Seiden shared her concern and contacted Thomas Arnold, the hospital’s assistant director of security and a former police officer. The bag was sent to the pathology lab. Although it was supposed to contain only saline and heparin, testing found insulin as well.

Three days later, Anna Byers, a patient in the Critical Care Unit, began a heparin drip. Within half an hour she was sweating, confused, nauseated, and weak. A blood test showed an extraordinarily high insulin level. Orange juice did not correct the crash, so nurses gave her dextrose directly into her bloodstream. That kept her alive, but the insulin continued to overwhelm the sugar they were giving her, and the crisis persisted through the morning, afternoon, and night.

Byers was scheduled for a heart catheterization the following morning. Her physician ordered the heparin drip removed in preparation. Once it was unplugged, her insulin problems eased. By two o’clock that afternoon she was stable, her blood-sugar crisis had disappeared, and she was ready to resume heparin. After the drip was restarted, however, the same unstable pattern returned. Nurses repeatedly gave her sugar as she alternated between stabilizing and crashing. By eleven that night, her blood sugar was too low to register. Her lines were disconnected and she was rushed to intensive care, but within twenty minutes off the drip she was feeling better again. The change with the drip’s removal, followed by the rapid recurrence after it was restarted, made the bag itself an increasingly important clue.

A Pattern Across Patients

Anna Byers was not the only patient affected. Down the hall, Fred Belf was also receiving heparin. By noon he was vomiting and unable to keep orange juice down; doctors ordered intravenous dextrose alongside his heparin. The two infusions continued in parallel through the day and night, as treatment for the falling blood sugar struggled against the process causing it.

By seven o’clock the next evening, Belf’s nurses had connected his deterioration to the heparin bag. They removed the drip, and he quickly began to feel better. Wearing gloves, they placed the bag in sterile plastic and sent it for testing. It tested positive for insulin. Microscopic examination of its exterior found tiny needle punctures, including three along the bag’s edges. Bags could be punctured while being connected to a stopcock, but punctures around the perimeter were unusual. The marks, together with the test result, made accidental damage appear less likely and raised the possibility that IV bags in the hospital’s storage room had been repeatedly and intentionally contaminated.

Arnold and Seiden now had two suspect bags, as well as the clinical experiences of Byers and Belf. They reviewed Critical Care Unit records to look for other unexplained insulin crashes. The resulting picture was incomplete: they had no established causal explanation for every episode, and the incidents were not confined to one unit or one shift. Even so, the records showed that patients had been crashing regularly for months. Codes sometimes overlapped, forcing nurses to leave one emergency to respond to another. Across intensive care, critical care, and cardiac units, patients were developing abrupt, diabetes-like crises without a known explanation.

Laboratory work strengthened the connection between the episodes and insulin exposure. The affected patients had exceptional insulin levels, and much of the insulin was identified as foreign rather than produced by their own bodies. That finding established that insulin had entered their systems; it did not, by itself, identify who had administered it or establish the cause of every crash.

From Medication Error to Investigation

The first explanation considered was an ordinary hospital mistake, such as a nurse misreading an order or a vial being mislabeled. Arnold and Seiden examined the charts but found neither a physician’s order for insulin nor a nursing notation that it had been given. The missing documentation left two troubling possibilities: an unrecorded error involving an unprescribed drug, or an act that was not an error at all. Either way, the records did not account for the insulin exposure.

Arnold took his findings to Joe Barry, Saint Barnabas’s senior vice president in charge of security, who had previously served as a decorated police officer and New Jersey State Police major. After ruling out patients’ visitors as possible suspects, Arnold and Barry compared nursing schedules with the dates and times of the patient codes. Three nurses had been working every code; among them, Charles Cullen drew particular attention. When Arnold interviewed staff, other nurses appeared anxious about their jobs, reputations, and patients. Cullen, in contrast, seemed unconcerned and defiant. His employment through a staffing service and his variable shifts made him difficult to schedule for an interview. When he was finally brought into a conference room, he refused to answer questions. Those circumstances made him a focus of the investigation, but the schedule comparison and his demeanor were investigative leads, not proof on their own.

The problem later remained visible in hospital procedures. Amy Loughren objected when her manager asked nurses to sign an insulin adjustment form recording their estimate of how much insulin remained in a vial. Previously, the insulin had simply been kept in a small refrigerator; the new form made nurses electronically attest to an amount they could only estimate. Loughren regarded the procedure as inaccurate and feared that nurses were being asked to stake their licenses on a guess. She also understood the change as a sign that something serious might have happened on her unit, possibly to patients. When she asked why insulin was being treated like a narcotic and what made it dangerous, her supervisor did not explain. Loughren refused to sign. Her resistance shows how a safety response could make a hidden danger perceptible to staff while withholding the information that would explain it.

Discovery, Trust, and Institutional Stakes

The insulin episodes formed part of the larger inquiry into Charles Cullen, but their significance lay in the way clinical observation became evidence. Byers’s improvement off the drip and relapse after its restart, Belf’s similar response, insulin-positive bags, and the unusual needle marks provided converging clues. Chart review and laboratory testing then connected individual crises to a pattern spanning units and months. At the same time, the investigation’s early limits mattered: the episodes were scattered, their records were anecdotal and confusing, and investigators had to distinguish a recurring signal from proof about particular acts.

For Loughren, the investigation altered a close working friendship into a conflict between loyalty and the possibility that Cullen had harmed patients. After a detective showed her Cullen’s Pyxis record from the night Reverend Florian Gall died, she could imagine that he had done something terrible, though she struggled to reconcile the evidence with her image of her friend. She was frightened of the investigation and worried that cooperating could cost her job. She later considered the risks of acting undercover: the detectives could be wrong, and Cullen was her friend. Her eventual assistance to investigators placed her nursing identity, professional security, and personal trust in direct tension.

The consequences extended beyond the immediate hospital inquiry. Families of Cullen’s victims or potential victims sued hospitals where he had worked; the New Jersey cases were settled out of court, with the files and settlement terms sealed. The New Jersey State Legislature subsequently passed the Patient Safety Act, requiring hospitals to report serious preventable adverse events to the state health department, and the following year added the Enhancement Act, which required limited reporting about health-care professionals and recordkeeping for complaints and disciplinary actions. These measures addressed reporting and accountability more broadly than insulin alone. In the book’s account, the insulin crashes are therefore both a medical emergency and an institutional warning: repeated, measurable harm can remain difficult to interpret when information is fragmented, explanations are withheld, and the people closest to the evidence are left to assemble its meaning.