The release of forensic toxicology results for Memorial patients’ remains turned a difficult investigation into a more concrete, but still disputed, evidentiary inquiry. On October 6, 2005, laboratory director Robert Middleberg told investigators that nine of eighteen samples contained morphine and about five or six contained Midazolam. The results aligned with a list of deaths already considered suspicious, although the degraded condition of the remains and disagreement about how to interpret postmortem drug levels prevented the findings from answering, on their own, how or why the patients died.

From recovered remains to laboratory testing

The inquiry began with a problem of evidence preservation. The bodies had remained in the heat for more than a week before recovery, and Orleans Parish coroner Frank Minyard initially told Special Agent Virginia Rider that an autopsy would be of little use: decomposition made blood sampling impossible, eliminating one familiar way to test for potentially lethal drug levels. Rider and prosecutor Butch Schafer needed evidence beyond witness accounts and empty morphine vials if they were to establish that a crime had occurred.

A volunteer pathologist pointed out that drugs might remain detectable in more resilient tissues, including liver, brain, and muscle, or in fluid that collects in the abdomen after death. At the attorney general’s request, Minyard agreed to ask a federal disaster mortuary team to conduct autopsies and collect tissue samples from people who had died at hospitals and nursing homes in Orleans Parish after Hurricane Katrina. The number initially appeared to be about one hundred; allegations of wrongdoing had arisen at several facilities.

On September 22, state and federal investigators flew north carrying samples from eighteen Memorial Medical Center bodies. They delivered them to National Medical Services, Inc., a forensic toxicology laboratory in Willow Grove, Pennsylvania. There, technicians prepared the tissues for analysis by gas chromatography and mass spectrometry. The laboratory’s work offered a route around the absence of blood samples, but it could not remove the interpretive problems created by decomposition.

The October 6 report

Two weeks after the samples were delivered, Middleberg called the attorney general’s office with preliminary results. Rider put the call on speakerphone so Schafer and another special agent could listen. Middleberg emphasized that the samples were less than ideal because of decomposition and that the testing was not complete; he was providing an update, not a final account of every case.

His preliminary count was striking: nine of the eighteen cases were positive for morphine, and approximately five or six more were positive for Midazolam, a drug also known by the brand name Versed. Middleberg described Midazolam as a medication generally used in operating rooms or when a patient was going to be intubated, rather than one routinely prescribed for home use. He also noted that the laboratory had found common medications, including antidepressants and drugs for gastric reflux, as would be expected in patient remains. The presence of Midazolam concerned him, and he stressed that investigators would need to examine the medical records closely.

When Schafer asked whether Midazolam could accumulate in a patient, Middleberg said that repeated exposure would not ordinarily be expected without repeated intubation and extubation or repeated trips to an operating room. The lab then provided the names associated with morphine-positive samples: Harold Dupas, Hollis Alford, Wilda McManus, Elaine Nelson, Emmett Everett, Alice Hutzler, Rose Savoie, Ireatha Watson, and George Huard. Dupas was among the LifeCare patients whom staff had believed were injected by Dr. Anna Pou or two Memorial nurses who came to the seventh floor on September 1.

The names were important not merely because of the positive results, but because investigators had not told the laboratory which patients were on their list of suspicious deaths. The eighteen tested cases that were positive for morphine or Midazolam corresponded exactly to the names on that list. Middleberg also characterized some of the morphine concentrations as very high. For Rider and Schafer, the match was a powerful investigative lead; it was not, by itself, a complete account of the medication each patient received, the timing of administration, or the cause of death.

Interpreting postmortem evidence

The subsequent review broadened the picture and sharpened the cautions. Pou and nurses Cheri Landry and Lori Budo had been arrested in connection with four deaths, but twenty-three of the forty-one bodies from Memorial and LifeCare analyzed by Middleberg’s laboratory tested positive for morphine, Midazolam, or both. The larger number did not automatically settle what had happened in any individual case. Postmortem tissue testing was not equivalent to measuring drug levels in a living patient’s blood, and prolonged exposure to heat might have changed concentrations. Middleberg advised the reviewers to consider each person’s clinical history alongside the laboratory findings.

The experts therefore examined the cases one by one. They considered whether a result might reflect a single large dose shortly before death or repeated doses that accumulated, especially in patients whose kidney or liver disease could impair drug processing. They also asked whether the drug combination was contraindicated for particular patients, whether breathing-suppressing doses would be dangerous for someone not on a respirator, and whether any estimated effect on breathing or blood pressure fit the known timing of death.

Alice Hutzler’s case illustrated why the laboratory result and the medical record had to be read together. Hutzler, a ninety-year-old LifeCare nursing-home resident treated for pneumonia, had morphine and Midazolam detected in liver, brain, and muscle tissue. Her chart, current until a few hours before her death was recorded on September 1, contained no prescription for either drug. It noted that she had been resting comfortably the previous afternoon, and her nurses had not documented pain or distress that night that would indicate a need for them. Those details raised specific questions about the medications, but the toxicology result alone did not establish who administered them or with what intent.

Disagreement and the limits of the findings

The results became part of a larger dispute over whether drug concentrations in decomposed tissue could reliably indicate what patients had received before death. Minyard consulted toxicologist Steven Karch, who argued that drug levels measured in a cadaver might bear no reliable relationship to levels before death. After reviewing the evidence, Karch considered it impossible to determine causes of death in bodies that had remained in very hot conditions for ten days and recommended that the medical causes be left undetermined.

Other experts challenged that conclusion. Forensic pathologist Michael Baden argued that uncertainty about the precise dose did not erase the fact that morphine and Versed were present, and said there was no competing cause of death. Pathologist Cyril Wecht considered some detected levels exceptionally high. To test whether heat and humidity might substantially alter levels, Wecht compared samples from overdose deaths subjected to hot, humid conditions with refrigerated samples. His small study found no significant differences, but it was not definitive, independently vetted, or published.

Minyard had to weigh those conflicting expert opinions as well as the prospect of a trial in which they would confront one another. He feared that the case might fail because of reasonable doubt and considered the consequences for New Orleans’ recovery. The toxicology findings thus occupied a consequential middle ground: they gave investigators evidence that drugs were present in specific remains and produced a notable match with their suspicious-death list, while leaving questions of dose, timing, medical effect, causation, and criminal intent to be tested against other evidence.

Place in the disaster’s wider chronology

The findings also acquired meaning against the conditions in which the hospital crisis unfolded. Before Katrina, concerns about flood-prone electrical components and the possibility of losing power had been documented, but proposed improvements were constrained by cost. During the emergency, Memorial’s generators ran for two days despite being intended to support the hospital for at least three; maintenance workers had to restart an overheated unit by adding water to its radiator. These circumstances help explain the pressure and uncertainty surrounding patient care, but they do not determine what the toxicology results mean for any particular death.

The book places the proposed patient injections alongside a critical change in available rescue capacity: helicopters had arrived in force at the time the plan went ahead. That contrast cautions against treating scarcity as a fixed condition rather than one that could change as rescue operations developed. In this account, the laboratory report is neither a verdict nor a self-sufficient reconstruction. It is a set of concrete findings whose significance depends on careful comparison with medical records, witness accounts, the timing of rescue, and the limits of forensic interpretation.