“Our Most Cruel Experiment Yet”

Chilling Testimony in a Tennessee Trial Exposes Lethal Injection as Court-Sanctioned Torture

Julie Hall smiled on the witness stand as she recalled a memory of her old client Joseph Wood. He had spent most of the last two decades living in solitary confinement, with his recreation confined to a cage, when the Arizona Department of Corrections began to loosen some restrictions over people on death row. A basketball court was built outside his unit on the sprawling desert prison complex in Florence, about an hour south of Phoenix. At 55, Wood was relatively healthy — “he loved going out and playing,” Hall said. A prison sergeant even played a round of basketball with Wood, which meant a lot. “He felt like he was being treated like he was human for the first time in a long time.”

Hall’s smile disappeared when she described the day Wood died. It was July 23, 2014. His execution was scheduled for 10 a.m. Hall arrived at the prison that morning at 6:45, then waited almost an hour to see him. When the Arizona Supreme Court granted a temporary stay of execution, Hall told him the good news. Wood was prepared to die, she told the court; ever since he committed the murders that sent him to death row, he had felt he did not deserve to live. Still, “he wanted someone to listen to us when we said that this was an experimental method of execution.”

Wood was the first to face a new form of lethal injection in Arizona that used a combination of the opioid hydromorphone and the sedative midazolam. The latter had raised controversy over its use in executions. Florida first tried it in 2013 to kill a man named William Happ “in what seemed like a labored process,” according to one media witness. Happ “remained conscious longer and made more body movements after losing consciousness” than people executed under the old formula, according to another report. The Florida Department of Correction, which refused to say how it chose the drug, dismissed the concerns — and soon other states were trying out midazolam. In January 2014, Ohio used it to execute Dennis McGuire. Witnesses described how he struggled and gasped, clenching his fists and striving to breath. A few months later, in April 2014, Oklahoma used midazolam to kill Clayton Lockett in one of the most notorious botched executions in recent memory.

But Arizona stuck to the plan. By noon that day, Wood’s stay of execution had been lifted. Prison staff provided Hall with a pencil and paper and led her to the witness chamber. No phones were allowed. Once inside, she was told, she would be forbidden from leaving the room. Hall watched as a pair of TV monitors were turned on above the closed curtains. “That’s where we could view the insertion of the IV lines,” she explained. Hall was surprised at the amount of blood she saw — some of it dripped onto the floor. With the IVs eventually placed, the monitors went dark. The curtains opened. Wood lay strapped to the gurney, thick straps over his arms and a white sheet covering his legs.

After 20 minutes and 134 gasps, she stopped counting.

At 1:52 p.m., a voice came over the loudspeaker. The lethal injection was about to start.

After five minutes, with the first dose of midazolam presumably administered, a man entered to conduct a consciousness check on Wood. The voice came back to announce he was sedated. But three minutes later, Hall said, “I saw a quiver in his cheek, which surprised me a little.” She didn’t know whether it was normal or not. It was two minutes after that when she saw Wood gasp for air. Then he did it again. And again.

“He just kept gasping,” Hall said. She began counting the gasps on her notepad. After 20 minutes and 134 gasps, she stopped counting. “I just didn’t know what the point was anymore.” Hall struggled to describe what it looked like. It reminded her of a fish that was dying after being pulled from the water — “that opening of the mouth; trying to get air and just not getting it.”

At 2:50 p.m., Dale Baich, supervising attorney of the Arizona Federal Public Defender’s Capital Habeas Unit, who was seated behind Hall, passed her a note. “Go now,” it said, instructing her to call their colleagues in Phoenix. Hall hurried out of the witness room and asked a guard if she could use his phone. He refused, then escorted her outside of the death house, through a maze of sally ports and checkpoints, and finally, out to the administration building. It took nine minutes. Only then was Hall able to make a call, to tell someone that “something was going very, very wrong and it looked like Mr. Wood was suffering.”

Hall was still on the phone when Wood was finally declared dead at 3:53 p.m. The next day, media witness Michael Kiefer published his own account of Wood’s struggle to breathe. Over the two-hour execution, he reported, Wood gasped more than 640 times.

Hall told her story in fits and starts, answering questions in a courtroom in Nashville, Tennessee. It was July 9, 2018, day one of Abu Ali Abdur’Rahman v. Tony Parker, a trial over Tennessee’s lethal injection protocol. Parker is the head of the Tennessee Department of Correction, or TDOC. The named plaintiff is one of 33 men facing execution under a new formula that includes midazolam. Three have been scheduled to die by the end of the year. One of them, Billy Ray Irick, is set for execution on August 9.

Hall was one of more than 20 witnesses called by the plaintiffs, including some dozen defense attorneys who had witnessed their clients’ executions. They dramatized what lawyers argued in their trial brief: that Tennessee’s new protocol violates the Eighth Amendment ban on cruel and unusual punishment. First issued in January, it called for the injection of three drugs: midazolam, followed by a paralytic called vecuronium bromide, and culminating with potassium chloride to stop the heart. With midazolam chosen to provide anesthesia, the attorneys argued it was not only possible but very likely their clients would suffer. What’s more, they said, the protocol prevents defense attorneys from having access to a phone during the execution, in violation of their clients’ constitutional rights.

The inescapable conclusion was that states have almost certainly been torturing people to death in their execution chambers.

The witnesses described executions in Alabama, Arizona, Arkansas, Ohio, Virginia, Florida, and Oklahoma. Many had never spoken publicly. Their accounts ranged from subtle but unusual movement on the gurney to gasping, lurching, and clenching of fists. They were bolstered by leading medical experts who explained the scientific reasons why midazolam was inadequate to provide anesthesia.

One pathologist presented evidence that had never been shown in court. He had reviewed 27 autopsy reports out of the 32 total executions carried out using midazolam. In most of the cases, he found signs of pulmonary edema — fluid in the lungs that indicated the men had been in respiratory distress. The inescapable conclusion was that states have almost certainly been torturing people to death in their execution chambers — and that Tennessee might be ready to do the same.

July 23, 2018; Nashville, TN, USA; Davidson County Chancellor Ellen Hobbs Lyle presides over a lawsuit brought on by death row inmates contesting the use of lethal injection as unconstitutional because, they say, it would lead to cruel and unusual punishment, at Chancery Court Part lll, Metro City Hall. Mandatory Credit: Larry McCormack/The Tennessean via USA TODAY NETWORK

Davidson County Chancellor Ellen Hobbs Lyle presides over a lawsuit contesting the use of lethal injection in Nashville, Tenn., on July 23, 2018.

Photo: Larry McCormack/The Tennessean via USA Today Network

After weeks of testimony, a ruling came quickly, on July 26. It sided with the state. In her order upholding Tennessee’s lethal injection protocol, Davidson County Chancellor Ellen Hobbs Lyle wrote that the plaintiffs had failed to prove their case, while acknowledging that the use of midazolam might leave them vulnerable to pain during their execution. The U.S. Supreme Court was “aware of the risk of midazolam,” she wrote, and upheld it anyway in Glossip v. Gross. Though “dreadful and grim, it is the law that while surgeries should be pain-free, there is no constitutional requirement for that with executions.”

For anyone who has followed the legal evolution of lethal injection, Lyle’s ruling was not a surprise. The decision ultimately turned not on midazolam, but on a different provision of Glossip. Under the ruling, the plaintiffs had to prove not only that Tennessee’s protocol was cruel and unusual, but that there was a viable alternative. In her dissent in Glossip, Supreme Court Justice Sonia Sotomayor decried this “surreal requirement,” one that puts attorneys in the perverse position of identifying methods that should be used to kill their clients. Though Lyle conceded that this law “seems odd,” the requirement was clear. “That proof has not been provided in this case.”

Decisions in chancery court have limited sway. Under Tennessee’s Declaratory Judgment Act, Lyle’s ruling amounts to a “declaration” — an opinion that can only be weaponized by bringing it to a different forum. Most lethal injection challenges are brought before federal courts that have the power to stop executions. Lyle did not. In bringing the lawsuit in chancery court, Federal Public Defender Kelley Henry hoped to win a ruling that could influence the state Supreme Court or governor to intervene.

Kelley Henry argues on behalf of plaintiffs during oral arguments on whether death row inmates can properly challenge the constitutionality of electrocution as a possible method of execution on Wednesday, May 6, 2015 in the Tennessee Supreme Court in Knoxville, Tenn. (Adam Lau/Knoxville News Sentinel via AP)

On May 6, 2015, Kelley Henry argues on behalf of plaintiffs on whether death row inmates can properly challenge the constitutionality of electrocution as a possible method of execution.

Photo: Adam Lau/Knoxville News Sentinel via AP

Yet the order belies the significance of the trial itself. As Henry said in her closing argument on July 24, it was the first time a three-drug protocol using midazolam had been the subject of a “real trial.” Until now, most hearings on midazolam were on whether to grant a preliminary injunction to stop a looming execution. Such hearings are rushed by their nature — witnesses often appear by Skype. This was not the case in Nashville. Though the trial moved quickly, the testimony was extensive and nuanced, providing a much fuller picture of the science behind the drugs used in lethal injection. Lyle was deliberate and measured — and cautious not to allow witnesses to testify beyond their expertise.

The questionable analysis of expert witnesses has had major consequences where lethal injection is concerned. At the preliminary injunction hearing that paved the way for Glossip, Alabama-based pharmacist Dr. Roswell Lee Evans peddled opinions divorced from scientific reality. Among his claims was that 500 milligrams of midazolam — the same dose as in the Tennessee protocol — would render someone unconscious to the point that they would not feel pain. Anesthesiologists adamantly disagreed. In an amicus brief to the Supreme Court, 16 professors of pharmacology cited the “overwhelming scientific consensus” that midazolam was incapable of inducing the “deep comalike unconsciousness” called for in lethal injection. On the eve of oral arguments in Glossip, the case was embroiled in controversy over the revelation that Evans had relied on sources like the website Drugs.com.

“Those states that have experimented with this drug have seen firsthand that it is a failure.”

There is “no debate around midazolam,” anesthesiologist Dr. David Lubarsky told the court in Nashville. Among such experts, Evans has no credibility. But among prosecutors intent on carrying out executions, Evans remains a useful and willing witness, “recognized by numerous state and federal courts,” as Deputy Attorney General Scott Sutherland told the court. If anyone lacked credibility, he suggested, it was the “highly biased” defense attorneys who watched their clients’ executions, he said, quoting a 6th Circuit ruling over Ohio’s lethal injection protocol. As a more authoritative source, Sutherland offered the official department of correction records from 19 executions carried out using midazolam in Arkansas, Florida, and Ohio. Many of them were described as problematic, but these records showed everything had gone fine, he said.

Yet states have already begun shifting away from midazolam. After a drawn-out legal battle following Wood’s death, Arizona agreed in 2016 to “never again use midazolam, or any other benzodiazepine” for lethal injection. “Those states that have experimented with this drug have seen firsthand that it is a failure,” Henry said. If Tennessee learns this lesson the hard way on August 9, it will not be for lack of warning. The trial in Nashville was a case study in Glossip’s twisted legacy — and a chilling look at the botched executions that continue amid little national controversy. If the death penalty is making a mockery of the Eighth Amendment, it is at least because not enough people are paying attention. In the words of the Supreme Court, “It is difficult to regard a practice as ‘objectively intolerable’ when it is in fact widely tolerated.”


Nashville’s Metropolitan Courthouse.

Photo: Radley Balko/The Intercept

Davidson County Chancery Court is located inside Nashville’s Metropolitan Courthouse, steps from where the Cumberland River cuts through town. The historic art deco building is anchored by Public Square Park, home to festivals, concerts, and the occasional protest. A few days a week over the course of the lethal injection trial, food trucks lined the southwest corner, where the Nashville Downtown Partnership hosted something called “Adult Recess”: whimsical lawn games like oversized checkers.

In her opening statement on the morning of July 9, Henry acknowledged the strange task at hand. “When we talk over and over and over again about ways in which to inject our clients with chemicals, it can become numbing,” she said. But the state has chosen a method of execution that is akin to being burned alive, she said. It’s easy to dismiss such descriptions — “Oh, those are just the words of those anti-death penalty people. That’s not real.” But, she went on, “the medical proof will show those words are not hyperbole.”

Henry was accompanied by eight other lawyers representing men on death row. Several of the cases date back decades, to an era when the execution chamber lay dormant in Tennessee. Henry arrived in Nashville in 2000, one month before the state carried out its first execution in 40 years. Six have been carried out since. In the meantime, like many death penalty states, Tennessee has changed its protocol repeatedly and haphazardly.

It wasn’t always that way. When the U.S. Supreme Court took up the precursor to Glossip — Baze v. Rees — many assumed that the questions over lethal injection would soon be settled. In its 2008 ruling, the court upheld a prevailing three-drug protocol that had been in use for decades. But the decision was followed by an unexpected sea change in lethal injection, which would throw the death penalty into chaos. The drug at the center of the ruling, sodium thiopental, became suddenly unavailable after its sole U.S. manufacturer stopped making it. As states sought new sources, many began altering their protocols, adopting new drugs based on what was obtainable and not necessarily what was most likely to work. After Glossip came down in 2015, midazolam became the new go-to replacement for sodium thiopental.

Close up of Thiopental Sodium in a tray, Thiopental Sodium is used as a barbiturate general anesthetic. (Photo by Universal Images Group via Getty Images)

Thiopental sodium in a tray.

Photo: Universal Images Group via Getty Images

Speaking before the court, Henry explained why this was such a serious mistake. In Baze, the Supreme Court acknowledged that the three-drug formula relied on an efficacious dose of sodium thiopental. Without it, the second and third drugs would cause extreme suffering, Chief Justice John Roberts wrote, making an execution “constitutionally intolerable.” Although the plaintiffs in Baze had argued there was too much room for error, the drug itself was at least designed to provide anesthesia. Midazolam was not. “Sodium thiopental is a barbiturate,” Henry explained. “It is a completely different class of drug than midazolam.”

Henry pushed back against the state’s argument that the true effects of large quantities of midazolam are unknown since there have been no “human experiments” to collect data. “Unfortunately, we do have human experiments,” she said. “We have 32 human experiments. Men who were executed using a protocol that involves midazolam.”

Sutherland began by invoking the gruesome crimes for which the plaintiffs had been convicted. “These facts provide context for this court as to why we are here,” he said.

With a low voice that was sometimes hard to hear, Sutherland wore a look of mild irritation — and the slightly casual air of a man who knows the law is on his side. He quoted Justice Samuel Alito’s reasoning in Glossip: “Capital punishment in this country is constitutional, and it follows, necessarily, that there must be a constitutional means of carrying it out.” The Constitution does not require a painless execution, Sutherland went on. It only prohibits the deliberate infliction of torture, such as disembowelment or being burned alive. What’s more, “in the history of its existence,” the court “has never invalidated a state’s chosen method of execution as cruel and unusual punishment.” As for midazolam, there was nothing new to discuss.

Sutherland seized on the main problem with the plaintiffs’ lawsuit. They argued in favor of a one-drug protocol using the barbiturate pentobarbital, a formula used by states like Texas. But they showed no proof that pentobarbital was available, he said. Instead, they argued that TDOC never made an effort to procure it. This was not true, Sutherland said, but regardless, “it’s not our burden to prove that it’s unavailable.” The plaintiffs had to prove that it was.

Sutherland echoed the late Antonin Scalia’s complaint during oral arguments in Glossip, blaming anti-death penalty activists for the drug shortage. There was truth to his claim — the human rights group Reprieve has waged a successful campaign over the past decade to convince drug companies to block the use of their products for execution. But the specter of overzealous anti-death penalty activists has also proven useful to states — a way to justify heightened secrecy around the procurement of execution drugs. Throughout the trial, the identities of the state’s supplier and drug procurer were kept secret.

For all the blame heaped on activists and capital defense attorneys, the trial would reveal the recklessness and repeated mistakes shown by the state in its relatively short history with lethal injection. Henry called it a “timeline of indifference.”

Ricky Bell, the warden at Riverbend Maximum Security Institution in Nashville, Tenn., gives a tour of the prison's execution chamber on Wednesday, Oct. 13, 1999. Both the electric chair and the lethal injection gurney are kept in the room until the time is near for a prisoner to be executed and the one not used is taken out. Robert Glen Coe, the prisoner most likely to be the first executed in Tennessee since 1960, has chosen lethal injection over the electric chair. (AP Photo/Mark Humphrey)

Ricky Bell, then-warden of Riverbend Maximum Security Institution in Nashville, Tenn., gives a tour of the prison’s execution chamber on Oct. 13, 1999.

Photo: Mark Humphrey/AP

Tennessee first adopted lethal injection in 1998. With the state’s execution machinery about to restart, lawmakers were concerned that the U.S. Supreme Court would strike down the electric chair on Eighth Amendment grounds. They proposed new legislation to get with the times, while making clear that they weren’t trying to make things easier for the condemned. “We should draw and quarter those suckers who commit these heinous crimes, but that ain’t constitutional,” then-Rep. Chris Newton, the bill’s House sponsor, said at the time.

To design a lethal injection protocol, TDOC put together a committee of prison officials to look at other states’ methods. Ricky Bell, then-warden of Riverbend Maximum Security Institution, where executions take place, went to Texas to see a lethal injection firsthand. The result was the same formula used across the country: a three-drug cocktail, beginning with a fast-acting barbiturate, sodium thiopental, to provide anesthesia. This was followed by pancuronium bromide, a paralytic drug that froze the muscles used for respiration. Lastly, potassium chloride would stop the heart.

“It gave me the creeps. It is a classic sign of an anesthetized patient being awake.”

The three-drug formula was never particularly scientific. Invented by an Oklahoma medical examiner in 1977, the method was simply replicated from state to state. Decisions on doses were left to prison officials; in Texas, which first carried out lethal injection, the gurney was chosen to make it look antiseptic, like a medical procedure.

The first person killed by lethal injection in Tennessee was Robert Glen Coe in 2000. Media coverage was heavy, yet there was relatively little detail describing his death. A witness for the Associated Press reported that he briefly “convulsed and coughed,” then lay still, but no one seemed concerned that anything had gone wrong. Under Tennessee law at the time, Coe’s defense attorneys were not allowed to attend.

This image taken from videotape recorded by prison officials shows Robert Glen Coe as he is wheeled to the execution chamber at Riverbend Maximum Security Institution in Nashville, Tenn. on  April 19, 2000. The state blacked out areas that show people other than Coe. The state released a heavily edited version of the video this week after several news organizations sued for access under the state's public records laws. (AP Photo)

This image taken from a video recorded by prison officials shows Robert Glen Coe as he is wheeled to the execution chamber at Riverbend Maximum Security Institution in Nashville, Tenn., on April 19, 2000.

Photo: AP

It was not long before questions began to be raised, however. In 2002, Abu Ali Abdur’Rahman challenged the state’s execution protocol in chancery court. At the heart of the lawsuit was the second drug in the protocol, the paralytic pancuronium bromide, marketed under the name Pavulon. At an evidentiary hearing in 2003, one month before Abdur’Rahman’s scheduled execution, his lawyers called Dr. Mark Heath, a professor of anesthesiology at Columbia University. Heath said he had begun studying lethal injection after the execution of Timothy McVeigh in 2001. The Oklahoma City bomber died with his eyes open, according to witnesses, some of whom described a tear welling up in his eye. “It gave me the creeps,” Heath would later tell reporter Bruce Shapiro. “It is a classic sign of an anesthetized patient being awake.”

On the stand in 2003, Heath explained that if the first drug in the protocol, sodium thiopental, was not adequately administered, the pancuronium bromide would cause suffocation while creating a “chemical mask,” concealing any evidence of the excruciating burning pain that would result from the injection of the third drug, potassium chloride. Lawyers called a woman named Carol Weihrer, who described her terror during eye surgery in 1998, when she woke up while under the effect of pancuronium bromide and was paralyzed, unable to alert her doctors.

Presiding over the 2003 hearing was Ellen Hobbs Lyle, the same judge who handed down the ruling last month. On June 1, 2003, Hobbs sided with the state, concluding that lawyers for the condemned had failed to prove that Tennessee’s protocol was unconstitutional. But she was critical of the lack of research behind the protocol — and particularly pointed in criticizing Pavulon, “a drug outlawed in Tennessee for euthanasia of pets.” It served no purpose except to give “a false impression of serenity to viewers, making punishment by death more palatable,” she wrote. And if the anesthetic failed to work, she warned, the paralytic would hide the “excruciatingly painful ordeal of death by lethal injection.”

Gov. Phil Bredesen talks about his proposed budget on Monday, May 8, 2006 in Nashville, Tenn. (AP Photo/Mark Humphrey)

Tennessee Gov. Phil Bredesen talks about his proposed budget on May 8, 2006, in Nashville, Tenn.

Photo: Mark Humphrey/AP

By the time Tennessee carried out its next execution, killing Philip Workman in 2007, evidence had come to light to confirm what Heath had feared. The Lancet, a prestigious medical journal, had published findings from a review of 49 executions across four states, showing evidence that many of the condemned had not been adequately anesthetized. Toxicology reports showed that concentrations of sodium thiopental in the blood of 43 of the men “were lower than that required for surgery,” while 21 had “concentrations consistent with awareness.” One of the co-authors of the study, Dr. David Lubarsky, later reviewed the autopsy report for Coe; in a front-page story in May 2006, The Tennessean summed up his conclusion: Coe was “probably awake and suffering silently.”

Later that year, a Florida man named Ángel Nieves Díaz died in a harrowing execution after IV lines sent drugs into his tissue rather than his veins. The following February, then-Tennessee Gov. Phil Bredesen halted four scheduled executions. A botched execution was “a governor’s nightmare,” he said. “How do you know that’s not going to happen here?”

Bredesen announced a 90-day moratorium and the formation of a five-member commission to revise Tennessee’s lethal injection protocol. The current execution manual was a “cut-and-paste job,” he said. An AP report revealed how its “minute-by-minute guidelines for lethal injection” included rules clearly written to apply to electrocutions, for example, the directive that staffers shave the head of the condemned, as well as the need to keep a fire extinguisher on hand.

The commission, made up of TDOC employees, proved to be little more than political theater, however. In April 2007, TDOC released its new protocol. Although it now included instructions on doses, it preserved the three-drug formula, including the paralytic. At 1 a.m. on May 9, Workman died by lethal injection at Riverbend. “As a media witness at last night’s execution, I can say it’s true Workman showed no obvious signs of pain,” Nashville Scene reporter Sarah Kelly wrote. “But even if he was in agony, he wouldn’t have been able to move.” A few months later, Tennessee carried out a third execution — that of Daryl Holton. Given the choice between the gurney and the electric chair, he chose electrocution.

In September 2007, while the Supreme Court prepared to consider the writ of certiorari in Baze v. Rees, a U.S. District Court held a hearing on Tennessee’s revised protocol. The four-day proceeding was “filled with absurdities,” according to the Nashville Scene. Testimony revealed that TDOC Commissioner George Little had actually rejected the advice of Bredesen’s commission, which had recommended a single dose of a barbiturate to replace the three-drug formula. On September 20, 2007, U.S. District Judge Aleta Trauger declared the protocol unconstitutional.

Just five days later, the U.S. Supreme Court granted certiorari in Baze v. Rees. It would be the first time the court would consider the three-drug method used for executions across the country. In April 2008, the justices upheld the protocol, 7 to 2. In light of Baze, Trauger vacated her previous decision.

It was not long afterward that states stopped being able to find sodium thiopental. Some sought new dubious sources, while others tinkered with their protocols. But there was a silver lining to the chaos. As they considered new methods, states began drifting away from the traditional three-drug formula, eliminating the paralytic from many protocols. In 2013, after a supply of sodium thiopental was seized by the Drug Enforcement Administration, Tennessee did what the commission had suggested years before, adopting a one-drug protocol using pentobarbital. But after Glossip in 2015, states started going back to a three-drug protocol, mimicking the old formula with something that was actually far more dangerous. Like many states, Tennessee is relying on an anonymous, unregulated compounding pharmacy for its supply of midazolam. As far as the paralytic, 15 years after Lyle first warned about its dangers, Tennessee has gone full circle.

On Friday, July 13, Riverbend Warden Tony Mays took the stand. After nearly a week of testimony about botched executions, it was a chance for the state to show that the situation would be in good hands. It didn’t. Mays, named warden in 2017, seemed alarmingly ill-informed, unfamiliar with various details of the protocol and unable to answer what, if anything, he would do if problems arose.

Some of this was no fault of his own. With no warning to the plaintiffs, the state had issued a revised protocol on July 5, just days before the trial was about to start. Henry called it a “cynical ploy to gain litigation advantage.” But it also created confusion for Mays, who was tasked with training his subordinates to carry out executions.

6-21-16 Parker_Tony_Department of Corrections_Offical Portrait

Tennessee Department of Correction Commissioner Tony Parker.

Photo: Dawn Majors/TDOC

The next witness was the official who ultimately has to answer for such concerns: TDOC Commissioner Tony Parker. A lifelong employee of the state’s prison system, Parker was named commissioner in 2016. The position made him responsible for selecting the state’s lethal injection protocol. Presumably, Henry asked, this meant that he would want to know what other courts and officials had said about previous versions of Tennessee’s lethal injection, correct? “I think I’d consider all the comments if I felt they were relevant,” Parker said.

Henry walked Parker through some of the state’s lethal injection milestones. Had he considered Lyle’s concerns over Pavulon back in 2003? Parker said he did not know what Pavulon was. Henry showed him portions of the ruling, which established Pavulon as the paralytic agent. Parker said he had never seen the decision before. “I guess it would be more important to me if we were using Pavulon,” he said. “But we’re not.” Would it be relevant if Pavulon was indistinguishable from the paralytic used in the protocol he chose? Henry asked as patiently as possible. “I would have to talk to a medical professional,” Parker said. “I don’t know any of that.” Henry asked Parker if he knew the difference between sodium thiopental and midazolam. No, he said.

Parker’s testimony was galling. But it was not entirely surprising. Parker had largely delegated his duties to his general counsel, Deputy Commissioner Debra Inglis, a veteran TDOC employee who was intimately acquainted with the state’s lethal injection history. She was a member of Bredesen’s 2007 commission that overhauled the execution manual. “Did Miss Inglis share with you that the committee’s recommendation at the time was to abandon the three-drug protocol in favor of a one-drug protocol?” Henry asked Parker. “We never had that particular conversation,” he said.

The email showed that the state had been warned by the source about the dangers of midazolam in the fall of 2017.

Inglis was the last state official to take the stand. Over more than two hours of testimony, she reiterated that Parker was the man in charge of selecting a lethal injection protocol. As for the move to adopt midazolam, she said, “it was his sole decision.”

Questioning Inglis, criminal defense attorney Kathleen Morris asked about an email that had raised controversy earlier this year. Written by an anonymous entity referred to as Source B, who had been tasked with procuring drugs for executions by TDOC, it was revealed by the Nashville Scene through an open records request. The email showed that the state had been warned by the source about the dangers of midazolam in the fall of 2017. “Here is my concern with Midazolam,” Source B wrote. “Being a benzodiazepine, it does not elicit strong analgesic effects. The subjects may be able to feel pain from the administration of the second and third drugs. Potassium chloride, especially.” Not necessarily a “huge concern,” the author added, but it could “open the door to some scrutiny on your end.”

Morris asked Inglis if she had discussed the email with Parker. He was made aware of it, she said. But they did not discuss it.

If such warnings were not enough to convince Tennessee to reconsider its search for midazolam in 2017, the drug had attracted plenty of scrutiny that year. In Arkansas, Gov. Asa Hutchinson had announced a plan to carry out eight executions over the course of 11 days that April. The reason for the rush: The state’s supply of midazolam was scheduled to expire at the end of the month.

Like Tennessee, Arkansas had not carried out an execution in years. It had never used midazolam. The planned execution spree threw attorneys into disarray. Only four executions ultimately went forward. Although there were some signs of trouble in the first three, no one reported any dramatic scenes. But in the final execution, on April 27, Kenneth Williams died a disturbing death. Media witnesses at the prison described how Williams had lurched and gasped. A spokesperson for the governor immediately dismissed the movement as “involuntary.”

He heard moaning, then “choking and coughing and heaving.” The sounds were loud enough to hear through the wall.

Among the witnesses that night was Eric Patrick Motylinski, a Pennsylvania attorney appointed to represent Williams. Taking the stand in Nashville, Motylinski described what he saw.

At 10:52 p.m., after giving his last statement, Williams began speaking in tongues, continuing after an announcement had been made that the lethal injection was about to begin. “His words kind of became slow and halting and they eventually stopped,” Motylinski said. But then “I saw his chest kind of pumping and I could see his head kind of moving back and forth.” He also appeared to be clenching his jaw. At 10:55 p.m., Motylinski heard moaning, then “choking and coughing and heaving.” The audio to the witness chamber had been turned off, but the sounds were loud enough to hear through the wall. Williams began to convulse, Motylinski said. “He was rising up from the gurney repeatedly, rhythmically, and finally kind of hitting up against the straps.” At 10:57 p.m., Motylinski decided to leave the witness chamber. The prison had agreed to allow him access to a phone. When he re-entered, Williams was lying still. He was declared dead at 11:05.

Motylinski’s testimony was disturbing. But it was also instructive. Despite being unable to stop the apparent suffering of his client, he had been able to take action on his behalf relatively quickly. This was precisely what Tennessee denied attorneys in its protocol, Henry argued. In a deposition, Parker had suggested it might be possible to accommodate lawyers’ requests for a phone to be available if needed. But Inglis did not support this. An attorney could go get their phone from their car.

Even when a state ostensibly allows for such safeguards, they are no guarantee. When Ohio defense attorney Carol Wright witnessed the 2017 execution of her client Gary Otte, who violently struggled on the gurney, she tried to leave the room only to be stopped by prison staff. “They blocked the door,” she testified. “They said, ‘Sit down.’” After she saw tears coming down Otte’s face, she went to the door again. “And I said, ‘Dear Lord.’” Only then was she given permission to leave and call a fellow federal defender, who phoned a federal court. Wright “is reporting that there were signs that Mr. Otte was conscious, crying, clenching of the hands, heaving at the stomach,” the colleague told a federal judge, according to a transcript. But the judge declined to intervene. In a ruling five days later, he concluded that the description was not enough to show that Otte “was experiencing unconstitutionally severe pain.”

Like all things involving prison, the witnesses’ testimony revealed how arbitrary the different rules can be from state to state. In Alabama, lawyers were not even allowed to have writing materials during executions. Defense attorney Spencer Hahn described how he kept mental track of the duration of Ronald Bert Smith Jr.’s heaving and coughing during his 2017 execution by recording the start time in his mind and counting the minutes and seconds on his hands.

The point, of course, is to control the narrative. When Motylinski was shown an internal affairs report from Williams’s execution in Arkansas, he said the notes “substantially minimize what I saw.” There was no mention of the coughing or clenching of his jaw. It did not capture the violence of his convulsions, the way he hit against the straps. The state’s official report was “sanitized almost to the point of being unrecognizable.”

Cross-examining Motylinski, Sutherland asked only one question: Was that the first execution he had ever witnessed? Yes, Motylinski said. But other defense attorneys were asked a range of questions that were meant to show bias. Wasn’t it true, Sutherland asked Dale Baich, that he once received an award from the abolitionist group Death Penalty Focus? Wasn’t it true, Assistant Attorney General Rob Mitchell asked Julie Hall after her testimony about Wood’s execution, that she exclusively represented clients in death penalty cases?

Cross-examining Assistant Federal Defender Leslie Smith, Assistant Attorney General Charlotte Davis asked how long she had represented her client before seeing him struggle on the gurney in Alabama. Fourteen or 15 years, Smith answered. Was she close to him? “Yes,” Smith said. “He was my client and I cared about him.”

July 23, 2018; Nashville, TN, USA; Federal Public Defender Kelley Henry speaks during a lawsuit brought on by death row inmates contesting the use of lethal injection as unconstitutional because, they say, it would lead to cruel and unusual punishment, at Chancery Court Part lll, Metro City Hall. Mandatory Credit: Larry McCormack/The Tennessean via USA TODAY NETWORK

On July 23, 2018, Federal Public Defender Kelley Henry, center, speaks during a lawsuit brought by death row inmates contesting the use of lethal injection.

Photo: Larry McCormack/The Tennessean via USA Today Network

On July 16, the plaintiffs called Dr. Mark Edgar, a diagnostic pathologist at Emory University. Edgar had reviewed all the available autopsies of people executed using midazolam across the country. Conducted by the local medical examiners in the counties where the men were executed, the documents ranged in their level of detail. Some states, like Ohio, don’t conduct autopsies after executions at all.

But Edgar found some details that jumped out. “I was struck by the abnormality in the lungs,” Edgar said. “All of the lungs were heavy with fluid.” An average lung, he explained, would weigh about 350 to 400 grams. But the autopsies he studied showed lungs more than double that weight. What’s more, most of the lungs showed signs of pulmonary edema — “evidenced by bubbles, froth and foam both in the lung tissue and in the larger airways.”

Edgar had created a chart to compile his findings. He underlined the parts that indicated proof of pulmonary edema. In total, 23 of the 27 autopsy reports. Among them were the autopsies of Joseph Wood and Kenneth Williams. Wood’s right lung weighed 980 grams, his left weighed 945. There were “marked amounts of blood and frothy fluid” indicating “acute pulmonary edema,” Edgar said.

“As it gets even worse, they may have a sense of terror, panic, drowning, asphyxiation.”

Assistant Federal Public Defender Amy Harwell asked Edgar to explain the symptoms of pulmonary edema. “When it begins, the patients are short of breath. They feel like they can’t catch their breath and they breathe a little bit faster,” he said. “As it gets worse, they may have a sense of air hunger and be gasping for air. As it gets even worse, they may have a sense of terror, panic, drowning, asphyxiation. It’s a medical emergency and a stage of extreme discomfort.” In a hospital setting, he explained, a patient would be given diuretics to remove fluid from the lungs. “Because they’re in such a state of panic,” they would also be given morphine, he added.

Edgar’s testimony was supported by a renowned pharmacologist, Dr. David Greenblatt, the longtime head of the department of pharmacology and therapeutics at the Tufts University School of Medicine. Greenblatt was not only an expert in midazolam, he was responsible for some of the earliest studies of the drug, which were used by the Food and Drug Administration in its approval for clinical practice. Greenblatt explained why pulmonary edema might result from a large dose of midazolam. To be injected, midazolam has to be mixed with hydrochloric acid, which makes it water soluble. Once injected, it would go first to the heart and then to the lungs, whose capillary circulation is dependent on “a very thin and delicate membrane” that is “very sensitive to acid.” A dose of 500 milligrams of midazolam would be destructive to the membrane, he explained, making it leaky. “So the lungs acquire fluid … and that makes air exchange difficult if not impossible.”

The testimony from Edgar and Greenblatt added a disturbing new dimension to the more widely known features of midazolam. Dr. David Lubarsky, the respected co-author of the 2005 Lancet study, said it was fine for sedating a patient — he called it “a martini in a syringe.” But midazolam has no analgesic effects on its own. When it is used in minor surgical procedures, like colonoscopies, it is paired with an opioid. What’s more, it is limited by what is often described as a ceiling effect, a widely accepted property among benzodiazepines. Even at extremely high doses, their effects eventually plateau. The 500 milligrams of midazolam called for by Tennessee will not make a difference. Dr. Craig Stevens, a neuropharmacologist, compared it to taking a bottle of aspirin to treat an amputated leg. Lubarsky said it was “like throwing a glass of water into the ocean.”

Midazolam is therefore useless in the face of “noxious stimuli,” the experts explained, especially something as severe as the injection of drugs like vecuronium bromide or potassium chloride. The former is “formulated in an acidic solution,” Greenblatt testified, which made it “painful going in.” If one is conscious as it takes effect, “basically you’re suffocating. You want to breathe but you can’t, because you can’t use your muscles.” The potassium chloride was also “extremely painful when injected,” Greenblatt said. When patients receive it as part of a medical procedure, “you have to dilute it tremendously and also give it very slowly.”

The expert testimony illuminated why the petitioners in Glossip called lethal injection using midazolam “the chemical equivalent of being burned alive,” as Sotomayor wrote in her dissent. As the Supreme Court has rejected further challenges to lethal injection, she has continued to speak out. “What cruel irony,” she wrote last year, “that the method that appears most humane may turn out to be our most cruel experiment yet.”

July 23, 2018; Nashville, TN, USA; Assistant Attorney General Rob Mitchell andDeputy Attorney General Scott Sutherland listen as Federal Public Defender Kelley Henry speaks during a lawsuit brought on by death row inmates contesting the use of lethal injection as unconstitutional because, they say, it would lead to cruel and unusual punishment, at Chancery Court Part lll, Metro City Hall. Mandatory Credit: Larry McCormack/The Tennessean via USA TODAY NETWORK

Assistant Attorney General Rob Mitchell, left, and Deputy Attorney General Scott Sutherland listen as Federal Public Defender Kelley Henry speaks in court on July 23, 2018.

Photo: Larry McCormack/The Tennessean via USA Today Network

Cross-examining the experts, Sutherland asked if they knew of any potential sources of pentobarbital for lethal injection. Not surprisingly, they said no. Surely they had professional contacts who might, Sutherland said. But the point was not to impeach their expertise. It was to show that the plaintiffs had not sought alternative sources to midazolam as required by Glossip.

The questioning became tense at times. Sutherland proved particularly irritating to Greenblatt. Much of his cross-examination was devoted to a tedious review of official records and timelines from executions in Florida, Arkansas, and Ohio, portions of which Sutherland asked him to read aloud. In Florida, beginning with William Happ, the first man executed with midazolam — whose death was described by witnesses as “labored” and prolonged — the records repeatedly showed “no unusual occurrences or problems,” Greenblatt read again and again.

On July 18, Roswell Lee Evans, the expert who helped pave the way for Glossip, took the stand for the state. With white hair and a white beard, he struck an affable tone. “My occupation at the moment is that I’m retired,” Evans said. “Previously I was a dean and professor at Auburn University School of Pharmacy in Auburn, Alabama.” Sutherland went through his curriculum vitae. He got his pharmacist’s license in Georgia in 1971, then a Ph.D. at the University of Tennessee College of Pharmacy in Memphis, where he became interested in psychiatry. In 1975, he went to the Western Missouri Mental Health Center, “probably the third center in the country to develop psychiatric pharmacy as a specialty.” There, he dabbled in research on benzodiazepines, and helped treat patients with schizophrenia and anxiety disorders. In 1994, Evans moved to Auburn, where his duties were largely administrative.

Had he ever done clinical research on midazolam? No, he said.

Sutherland asked the court to qualify Evans as an expert, “based upon his scientific, technical, specialized knowledge through his career.” Henry objected. Lyle allowed her to question Evans before continuing. “Dr. Evans, you are not a pharmacologist, correct?” Henry asked. Correct, Evans said. He had no experience prescribing midazolam in any anesthetic context, did he? No, Evans said. “In fact, you’re not a medical doctor, are you?” He was not. Had he ever done clinical research on midazolam? No, he said. Sutherland pushed back. “Dr. Evans has been recognized in state and federal courts in the field of pharmacology and I think his experience speaks for itself.”

Over Henry’s objections, Lyle allowed Evans to testify in pharmacology. But she monitored his testimony. “I have no clinical experience with midazolam other than personal,” Evans conceded as his direct testimony resumed, but said that he had studied the drug and lectured on it. “Tell me what you mean by ‘studied it,’” Lyle interjected. “Reviewed the literature on it,” he said. Lyle pressed further, but Evans remained vague. “The studying … was largely in preparation for educational materials, lectures and so forth,” Evans said. Most recently, it was “in relation to hearings such as this.”

Henry continued to object over the course of Evans’s direct testimony. Other times, Lyle intervened. Evans was eventually able to provide some opinions, but they were somewhat hard to follow. He drew a distinction between the therapeutic use of midazolam and “the toxic use.” Although Greenblatt had testified extensively about studies showing that overdoses of benzodiazepine were not fatal on their own, Evans provided some examples. There was a 63-year-old man who had “received 10 milligrams of midazolam and he expired as a result,” he testified. In her dissent in Glossip, Sotomayor had lambasted the apparent logic at hand: that “because midazolam caused some deaths, it would necessarily cause complete unconsciousness and then death at especially high doses. … One might as well say that because some people occasionally die from eating one peanut, one hundred peanuts would necessarily induce a coma and death in anyone.”

Nevertheless, Evans was ultimately able to render the opinion most sought after by the state. “To a reasonable degree of pharmacological certainty,” Sutherland asked, “could 500 milligrams to 1,000 milligrams of midazolam render someone unconscious and insensate during the lethal injection?” It could, Evans said.

Cross-examining Evans, Henry reviewed his track record. Until he first assessed midazolam for the state of Florida in 2014, she asked, “you had not researched what would be necessary to achieve a lethal dose of drug?” No, he said. She pointed out instances in which he gave erroneous opinions, for example, claiming that midazolam was approved by the FDA for use as a general anesthesia. In one Alabama case, Evans had conceded at a deposition that he had previously misinterpreted a study he used to question midazolam’s ceiling effect — yet he had included that same flawed interpretation in his declaration to the state of Tennessee. “I reassessed the article and I stand by my initial comments,” Evans explained.

Henry went back to his early Florida cases. In one, Evans had said that midazolam had no pain-relieving properties at all, only to testify in a later case that it might help with lower back pain. Explaining the contradiction in a deposition, Evans said he’d had “a chance to do a little more digging.” Now, Evans appeared to be changing his mind again. When Henry asked whether midazolam had any pain-relieving properties, he said no. But then he elaborated. “There is some pretty hypothetical information” to show that “it may — may — have analgesic effects,” he began to say before Lyle stopped him. She wanted to know what further digging he had done in the previous case. “Your honor, I don’t remember,” he said. “But it was in a journal.”

On July 30, Henry and her colleagues asked the Tennessee Supreme Court to halt the August 9 execution of Billy Ray Irick. The plaintiffs planned to appeal and “it would not be appropriate to move forward with an execution while the issue of the constitutionality of Tennessee’s lethal injection protocol is still pending in the courts,” Henry said in a statement. Even while ruling against them, Lyle had acknowledged the risks posed by midazolam, she added.

Henry reiterated an argument she had tried to make at the end of the trial. If the state could not carry out an execution using the alternative they had put forward — a single dose of pentobarbital — the plaintiffs moved to amend their complaint to consider an “alternative to the alternative”: a two-drug cocktail that removed the vecuronium bromide altogether. This option is “clearly available and readily implemented,” which would satisfy the Glossip requirement. And it would remove one of the well-established risks: that their clients would be paralyzed, suffocating, and suffering as the lethal drugs took hold.

It seemed simple enough. Parker himself has suggested it would be possible. Indeed, Lyle had been among the first in the country to criticize the paralytic back in 2003. “If the state is sincere in its belief that midazolam will work the way that they say it will work,” said Bradley MacLean, counsel for Abu Ali Abdur’Rahman, “there is no reason why the state should oppose this.”

But it did. Sutherland called it a “desperate” move, while Lyle explained that the law prevented her from granting the motion. As for her prescient opinion 15 years ago, she wrote in her ruling, it came before Baze and Glossip. The Supreme Court had found a legitimate purpose for the paralytic: hastening death, while dignifying the process for witnesses and the condemned alike. Her previous decision was “of minimal use.”

In his closing statement, Sutherland decried the repeated challenges to Tennessee’s execution protocol over the years. “Nothing has ever been good enough,” he said. Whether he meant to or not, he also got to the heart of the problem — with the paralytic, with lethal injection, with using the tools of medicine in order to kill. The plaintiffs wanted a dignified death, he said, but why should their deaths be peaceful? The deaths of murder victims weren’t peaceful, he said, turning to stare at the audience in the courtroom. “Death is not pretty, your honor.”

Join The Conversation