On July 23, Arizona took 117 minutes to execute a convicted murderer named Joseph Rudolph Wood III. It was not the nation’s first execution to last that long. In September 2009, Romell Broom entered the Ohio death chamber and exited two hours later still breathing—the only inmate in U.S. history to survive a lethal injection. The executioners had scoured his arms, legs, hands, and ankles for veins in which to stick their needles. But they repeatedly missed the vessels with the IVs. After at least 18 failures, Ohio had no choice but to cancel the execution.
In Wood’s execution, the trouble began when the drugs began to flow. Arizona’s executioners first injected Wood with a combination of midazolam and hydromorphone, two drugs they had never used before in an execution. When the first dose failed to stop his heart, the executioners administered a second. And then a third. The execution team injected 15 doses in total before a doctor finally pronounced death. An Arizona Republic reporter witnessing the execution said Wood gasped more than 640 times and that he “gulped like a fish on land.”
Despite their different problems, the attempted execution of Broom and the execution of Wood are connected by more than just their lengths. Had executioners in Ohio been able to insert IVs into Broom’s veins, Wood’s execution might have gone much more smoothly. That’s because the Broom debacle led Ohio to write a “Plan B” for lethal injections, introducing into the death chamber for the first time the untested drugs Arizona would use years later to kill Wood. And emails I obtained from Ohio reveal some of the state's internal debates and concerns about these drugs—including fears that an inmate could “gasp” and “hyperventilate” as he died.
Doctors warned from the beginning that midazolam and hydromorphone could create “a distasteful and disgusting spectacle.” And yet the drugs spread from Ohio across the country, revealing the lengths states will go to in order to carry out death sentences despite constant IV trouble, drug shortages, and problematic executions.
Broom’s survival left Ohio in a panic. While states had been botching executions since they first began to use lethal injection in the 1980s, none had ever failed to kill a prisoner. Ohio’s first step after Broom was actually a concession to lethal-injection critics, who had long argued against the standard three-drug cocktail that chemically paralyzes the prisoners. In future executions, Ohio announced in November 2009, it would use as its main method of execution only a large dose of a single barbiturate—similar to the way veterinarians euthanize family pets.
But this adjustment failed to address the fundamental problem of Broom’s execution: The barbiturate still required IV access. So Ohio also added to its execution protocol a “Plan B,” in case executioners failed again to find a vein. It was the first time a state devised a new method of execution since they began switching to intravenous lethal injection in the 1970s.
Ohio spoke with several medical professionals, including Mark Dershwitz, a professor of anesthesiology at the University of Massachusetts Medical School. Dershwitz has provided expert opinions, either in writing or court testimony, about lethal injection for more than 20 states and the federal government. He has never testified on behalf of an inmate. Dershwitz had been providing feedback to Ohio about lethal injection on a continuing basis since 2003, when it first used him as an expert witness. Now, in 2009, he offered his medical opinion on several potential new execution drugs. Ohio was willing to consider all options, including inhaled drugs, which Dershwitz cautioned against. “The process bears some resemblance to witness reports I have read of execution in the gas chamber,” he wrote in an email to Gregory Trout, a lawyer at the Ohio Department of Rehabilitation and Correction.
Emails 1 and 2: Ohio and Dr. Mark Dershwitz discuss the option of an "inhaled" drug for executions.
Ohio was mainly interested in drugs that it could inject directly into the muscles, like an EpiPen or flu shot. As the conversation continued, Ohio focused on the sedative midazolam and the opioid hydromorphone. Neither drug is generally used in medicine to induce unconsciousness before surgery—midazolam is typically used to calm patients and hydromorphone is a powerful painkiller—but Dershwitz said that, in combination, large doses of the two drugs would knock the patient out and stop his breathing.
Ohio still had concerns. Among the state’s fears, Trout wrote, was that a combination of midazolam and hydromorphone might result in a prisoner “gasping for air in a hyperventilating fashion, with eyes still open.” Trout worried this “would create the appearance, at least, of suffering, which would upset witnesses and inspire litigation.”
Dershwitz did not think this was likely. “The combination of an opioid and a benzodiazepine (almost always midazolam) is commonly used for conscious sedation,” he wrote. “I have never seen the scenario described (in my 27-year career) although I will admit nothing is impossible.” In later court testimony, Dershwitz estimated an intramuscular injection of these two drugs would take nine to 14 minutes to kill a prisoner.
Email 3: Dershwitz answers Ohio's questions about midazolam and other drugs.
When Ohio adopted the new protocol, death-row inmates immediately challenged it in court. They brought in other doctors as expert witnesses, who were skeptical that the procedure would work as smoothly as Dershwitz described. “I do not believe that any method of execution ever proposed in the United States is as slow as this one,” Dr. Mark Heath, an anesthesiologist at Columbia University, told a U.S. district court judge in 2009. He called Plan B a “disaster,” warning that it could create “a terrible, arduous, tormenting execution that is also an ugly visual and shameful spectacle.” Before losing consciousness, the patient would be “subjected to the intoxicating effects of these drugs, which include hallucinations,” Heath said. He ended his testimony with a description of what he believed to be a worst-case scenario:
It could easily happen that the prisoner is beginning to lightly doze off and the vomit, the gastric contents, come up into his mouth and then are breathed in, and that would be agonizing and would cause the prisoner to struggle awake and find himself with a mouth full of vomit and lungs full of acidic vomit, lying on his back, unable to really cough it out effectively. It might be like water-boarding, but worse.
When Dershwitz testified in the case, he was dismissive of Heath’s warnings: “I submit here that Dr. Heath is uninformed on this topic,” he told the district court, which ultimately upheld the execution protocol. Nevertheless, Ohio was at least a little spooked by the potential problems Heath laid out. When the first execution under the new protocol took place in December 2009, state officials warned witnesses that the prisoner could convulse and throw up if they resorted to Plan B.
Ohio never used Plan B as an intramuscular injection. But in 2009, the sole domestic manufacturer of thiopental, the barbiturate Ohio wanted to use in Plan A, reported a shortage. After Ohio used up its supply of thiopental, it replaced the drug with pentobarbital, another barbiturate. But in 2011, pentobarbital’s Danish manufacturer refused to sell any more of the drug to U.S. states for executions.
By fall 2013, Ohio’s stockpile of pentobarbital had run dry, and it rewrote its protocol again. Now Plan A would be an intravenous injection of midazolam and hydromorphone, the same two drugs that Plan B administered intramuscularly. Ohio planned to use the new protocol for the first time in the execution of a convicted rapist and murderer named Dennis McGuire. In his appeals, McGuire’s legal team hired as an expert witness Dr. David Waisel, an anesthesiologist at the Boston Children’s Hospital, who predicted a problematic execution.
“It is substantially likely that McGuire will remain awake and actively conscious for up to five minutes, during which he will increasingly experience air hunger as the drugs suppress his ability to breathe,” Waisel wrote in a declaration. In this scenario, McGuire would experience "the same agonizing and horrifying feeling as suffocation," Waisel said. He also noted that neither midazolam nor hydromorphone is generally used in medicine for anesthesia (that is, inducing deep unconsciousness).
Once again, Dr. Dershwitz testified on behalf of Ohio. Because the drugs would now be applied intravenously, Dershwitz said they should work more quickly than they would intramuscularly. He also told the court, “I have never once acted as a consultant, which my definition means that I was helping them write a protocol or giving them advice about a protocol.” In the 2009 emails with Trout, Dershwitz reviewed a draft of Ohio’s proposed protocol and he corrected at least one error. Dershwitz says that this is consistent with his role as expert witness: "If I told him he had a typographical error in the size of a syringe, that is not helping them to write the protocol," he told me. "It's telling them, 'As part of the litigation, you need to fix it because the other side will see it too.' But the litigation was already underway, and an expert witness has the responsiblity to tell the side that they're helping where there are problems or holes in their case."
Email 4: Dershwitz reviews a draft of Ohio's protocol and corrects an error.
Once again, the federal judge approved the execution, though not without some reservations. "There is absolutely no question that Ohio's current protocol presents an experiment in lethal injection processes," the judge wrote. McGuire’s execution on January 16, 2014 took 26 minutes overall. Witnesses said McGuire began to make snoring or snorting sounds five minutes into the execution, and that he opened his mouth repeatedly. “I watched his stomach heave, I watched him trying to sit up against the straps on the gurney, I watched him repeatedly clench his fist,” McGuire’s son said. “[It] appeared to me he was fighting for his life while suffocating.”
Neither Waisel nor Dershwitz attended McGuire's execution, but they read about it after. “I didn’t think it would last that long,” Waisel says. “I was wrong about that.” Dershwitz, for his part, says McGuire's execution was "exactly what I predicted to happen." He says McGuire appears to have lost consciousness a minute or two into the execution and that he then snored, which is common for patients under anesthesia. He also says that McGuire likely died more quickly than was reported, since he says Ohio has a built-in "administrative delay" in pronouncing death.
Still, the drugs had clearly not worked in the manner that Ohio expected—at the very least, they had created the "appearance of suffering" that Trout had feared. Nevertheless, other death-penalty states were quick to follow Ohio’s example. Like Ohio, they had run out of thiopental and pentobarbital and still wanted to carry out their scheduled executions. While McGuire’s execution may have been problematic, the method had at least already won the approval of the U.S. courts. So they adopted it as their own. Eleven days after McGuire died, Louisiana said it would use intravenous midazolam and hydromorphone to execute Christopher Sepulvado (though a U.S. judge has since stayed that execution). Kentucky and Oklahoma have also added a two-drug combination of midazolam and hydromorphone to their protocols. In March, Arizona did the same, and became the second state to use midazolam and hydromorphone when it executed Wood. (Dershwitz played no role in Arizona's defense of its protocol.) Still more states began using midazolam in combination with other drugs. In total, four of the 12 executions that have used midazolam in some way have appeared problematic.
On Wednesday, Montana announced in a state court filing that Dershwitz had withdrawn in June as an expert for the state in death-penalty litigation—a role in which he had served Montana for six years. In fact, Dershwitz "has terminated his role as an expert witness on behalf of all states and the federal government," the filing said.
In an email and phone interview, Dershwitz explained his decision. In 2010, not long after his conversations with Ohio about midazolam and hydromorphone, the American Board of Anestheseogy added to its professional standing policy a prohibition against the participation in lethal injection and threatened to revoke the certification of any member who did. By "lethal injection," the ABA meant not just actual executions but also the authoring of protocols. Dershwitz says he contacted the ABA at the time to ask whether doctors could still serve in court as "expert witnesses," and the ABA said they could. So Dershwitz started insisting explicitly in his contracts that he always be referred to as an "expert witness" and never as anything like "consultant" or "adviser."
After the McGuire execution, however, Ohio issued a press release that Dershwitz said made him lose his faith in states' ability to abide by their agreement. Ohio, in announcing that it would increase the dosages of midazolam and hydromorphone in future executions, said it had "discussed" the McGuire execution with Dershwitz. However, Dershwitz says that the correspondence was "entirely one way"—that Ohio officials told him what had happened and he never replied. "These lawyers in Ohio clearly used very misleading language to make their own side look better," he says.
"It became clear to me that, contract or no contract, the mistakes made by Ohio in the April press release could happen again regardless of my efforts to codify in writing my role as an expert witness," Dershwitz wrote in his email. "I cannot take that chance and thereby terminated my role as an expert witness on behalf of Ohio and the other four jurisdictions with which I had a contract."
Dershwitz's decision will make it more difficult for states to defend their execution protocols in court. "He has been the prime states' expert since this issue started gaining traction in the 2000s," says Deborah Denno, a professor and lethal-injection expert at Fordham Law School. "He has always been at the top of states' list, if not the only one."
Montana has asked the court for more time to prepare its case for trial: Dershwitz, it said, was "the sole witness for the State Defendants."
Ben Crair is a story editor at The New Republic.